![]() ![]() |
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
196.3.50.254
In Reply to: Re: Hummmmmmmmmmmmm. posted by Dan Banquer on November 3, 2006 at 05:27:55:
You are the one who is trying to patch a non-linear circuit with a crutch that above about 1Khz simply stops working. If YOU understood linearity you would also realize that the transfer function of a triode tube is closer to linear than either a bipolar or mosfet transistor.However; it is clear that 1)global feedback simply doesn't work well up to 20Khz. This can be seen readily in damping factor vs. frequency and THD vs. Frequency plots. Most high feedback amps show a decreasing damping factor above 1KHz and an increasing THD starting about the same frequency. 2) It appears to increase the ratio of high order harmonics (this can be shown mathematically and in practice) and signal correlated "noise" floor. Granted it is not the only culprit in this as noise and power supply hum can add IM products and muck up the sound as well as underbiasing the transistors/tubes. 3) Feedback can make the compatibility with speakers worse due to back EMF from the speaker being reinjected into the amp and reamplified. This was shown by Otalla et al in the late 70s early 80s.
So from my listening experience and based on technical analyses I have seen and measurements I conclude that in general feedback does more harm to sonics than the good it does "linearizing" the circuit. This argument of "its more linear so it must be better" is so shallow and so easy to expose when one looks behind the THD curtain. Give it a rest, we are not oscilloscopes who don't give a shit about harmonics. We are more sensitive in a lot of ways than the oscilloscope and especially to the harmonic content of distortion.
You and I both know that it is not the absolute amount of THD that is important but the harmonic and IM content of the distortion under DYNAMIC conditions. So, absolute linearity is not necessarily a requirement or the issue because an amp can be less linear in an absolute sense (say 2% THD vs. 0.1% THD) than another and yet have less AUDIBLE distortion (the 2% amp could be all 2nd and 3rd harmonics while the other has harmonics right up to 20Khz for a 1Khz sine wave). Audible linearity is the issue, Dan so stop with this linearity according to the oscilloscope is everything. I know you don't even believe that yourself or you wouldn't bang the drum that your amp makes "only" 2nd order harmonic distortion (yeah sure it does).
![]()
Follow Ups:
Damping factor over 400 hz is irrelevant, it's only a factor for controlling woofers.Slew rate is deliberately limited to 13v/usec to avoid rf damage to speakers.
Distortion is virtually non existant, inaudible under any conceiveable conditions of use. And BTW, if this isn't enough power to play your speakers at any volume you'd need without clipping, there are two more larger versions offering considerably more power.
This amplifier is unconditionally stable with ANY load and has a three year no fault warrantee. So reliable, many of the world's professionals count on it to satisfy their customers. Hardly the garage built typical audiophile junk we see selling for as much or far more and certainly not one of those feeble peewee amps so popular among the one driver horn lovers. And yes it uses lots of feedback. Performance of this high caliber is not possible without it.
![]()
There's a guy selling two of them on ebay starting at only $700! Fortunately for you, no one has placed any bids for the first five days, so you've got a great chance of making your dream come true!
Distortion is virtually non existant, inaudible under any conceiveable conditions of use.You're still parroting that THD is relevant? You continue to forget that some of us have actually owned Crown amps. I got over mine when I was 18 and learned better. You're what - 65? Maybe if you're lucky you'll eventually get it in your lifetime.
So reliable, many of the world's professionals count on it to satisfy their customers.
Exactly. It is the audio equivalent of a Ford F-350 pickup with a V-10 Triton engine. Powerful. Durable. Capable of heavy loads. Just the kind of high performance vehicle that can really carve the corners at Nurburgring!
And yes it uses lots of feedback. Performance of this high caliber is not possible without it.
And yes it uses lots of leaf springs. Performance of this high caliber is not possible without it.
You are so dense soundmind. My point about damping factor has nothing to do with woofer control!! Besides the frequency response variations with a damping factor of greater than 10 are small enough that many other factors are much more important. Look at the audioholics article on it if you don't believe me.My point about damping factor is that it indirectly shows that the feedback is losing its "corrective" power when the damping factor drops (one of the main features of negative feedback being a lower output impedance). The subsequent rise in HF distortion also points this out.
As to whether I would like the amp...I doubt it because I don't like the sound of many many amps so statistically it is at a disadvantage.
![]()
Here is a link to some measurements done on a tube amplifier with a damping factor very close to 10. The deviations from a flat frequency response into the standardized dummy speaker load are't huge but are certainly large enough to be audible.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Like my Sound Labs. Or Dunlavys.
The load is used to give an idea of the magnitude of the FR variations of an amplifier into a typical speaker load.An amp with a low output impedance will have virtually the same frequency response into that roller coaster load as into a resistor. And that includes most good solid state amps. So I don't see that you have an argument here in favor of tubes.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
and then found the best sounding amp that meets my priorities for the $$ to drive it. Speakers with linear impedance curves don't restrict the choices.rw
Well, one needs an amplifier that can drive one's speakers.Still, I already have a fairly capable amplifier that will drive most speakers very nicely. When I audition speakers, I use SS amps as a rule and I can be fairly confident that they will sound much the same as they would with mine. I've no wish to get speakers I really don't like and try to improve their sound with different tube amps. Life is too short. I prefer to find speakers that sound good with accurate amps.
On occasion, I have heard some speakers using both SS and tube amps and so far, I have preferred the sound with the solid state amp. I heard my PSB Stratus Minis with a tube amp which did nothing to improve them, though they sounded OK. Said tube amp also sounded pretty coloured with the Dali Helicon 400, which sounds quite nice with a good solid state amp with a low output impedance.
I heard a Rogue integrated with some Usher X-719 speakers and didn't like the result. I preferred the sound with a SS amp, though I still wasn't all that impressed with the speakers.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Why do you find the +/- 1 db variation caused by a roller coaster speaker impedance load (boy look at that S2 ride!) significant yet (apparently) find the +8 / -13 db response of your speakers insignificant?
![]()
Look closely at the VTL's errors in such a load: a 1 db rise centered around 70 hz (S2 trough 2.5 db), another 1 db rise centered around 1.5 khz (S2 trough 5 db), a 1.25 db trough centered around 4.5 khz (S2 rise 2 db) and a gradual downward slope starting around 10khz (oh well can't fix the spectacular S2 peaks/troughs here). With the exception of the top octaves, it seems to me the errors are quite complementary.
Let's clear a couple of misstatements up right now. First of all, it is not the alleged roller coaster impedance curve that causes the audible frequency response variations, but it is in great part the of the interaction of the output impedance of the amp with the impedance curve of the speaker.Second, I have never maintained the FR variations in speakers are insignificant, and I have no idea why you would think so. It is certainly something you have made up.
Third, as to the variations caused by the interaction of the impedances of the amp/cable and speakers, I make two basic points about that. First of all, that the FR variations of amps with a high output impedance into many speakers loads is likely to be audible, sometimes quite audible, whereas the FR of low output impedance amps varies very little into most speaker loads as compared to driving a resistor. Of course, as you pointed out, an amp with a high output impedance driving a speaker with a pretty flat impedance vs. frequency curve would change its FR much--but then, it would sound pretty much like an amp with a low output impedance!! That's hardly much of an argument for favoring tubes . . . but if you prefer tube amps, that's fine with me. I'm not trying to convert you, but appear to find my preference for solid state threatening.
A second point is that I generally audition speakers with good solid state amps (which mostly have a low output impedance). This removes a variable which is likely to be significant from the audition.
If I thought some FR shaping would improve my speakers, I would get an EQ (1/3 octave or perhaps parametric) rather than trying to try out different tube amps in the hope that something would improve. But, then I make every effort to get speakers I really like and I have succeeded in doing so over the past decades.
I'll have to suggest the same thing to you as I do to RGA: if speaker measurements don't mean anything to you, then don't bother with them.
Now, I can only tell you what I know, and there are people who know a lot more than I do about this. If you really want to find some answers to your questions about what is significant from technical people, why don't you look at Dr. Toole's White Papers? I think this is the most relevant one:
http://www.harman.com/wp/pdf/Loudspeakers&RoomsPt2.pdf
JA, who took the measurements of the Paradigm Signature S2 for Stereophile, commented that "the balance overall is impressively flat," and it is. In JA's measurements, there is no such thing as a "+8 / -13 db response" in his curves above the bass. The only thing about +8 I can see is tweeter ringing above the audible range at maybe 27 kHz (many speakers have this, even some 'stats, I think), and on axis there is a very narrow interference dip (I should guess)around 12 kHz, but not off axis. JA pointed this out, BTW, but perhaps you missed it.
Now, though I find JA's dispersion measurements to be very revealing, I think the NRC measurements are more accurate, and even JA likes to compare his results to them when a manufacturer such as PSB provides them. And the NRC measurements as shown on the Soundstage site, linked below, are in many respects extraordinarily good for a speaker, such as the Listening Window curve in Chart 2. While you might prefer the response of the S2 with a tube amp, then again you might not. In any case, if I want a more downward sloping response with some recordings, there is the Tilt control on my Quad preamp. Much handier than changing amps!
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
What about the interaction of high feedback amps (like those SS amps you love because of their low output impedance) with highly reactive speakers (yep those ones with a roller coaster impedance curve)?Don't know what I am talking about?? It was described in the late 1970s and 80s by Otalla regarding the back EMF of the speaker CREATING distortion in the amplifier because the feedback loop provides a pathway for the back EMF to be reamplified. They suggest that it could be a big reason for the variability in the sound of amps with various speakers and of course in no way affects the frequency response. Low damping factor amps (usually low or no feedback) are immune to this problem and simply dissipate the back EMF as heat at the output transistor or tube.
Its not all about FR. I have done experiments where I made equalization curves for different amps to the same target. The amps still sounding drastically different...not just slightly different but the difference between finding the music interesting and boring (talking strictly about acoustic music here).
"If I thought some FR shaping would improve my speakers, I would get an EQ (1/3 octave or perhaps parametric) rather than trying to try out different tube amps in the hope that something would improve. But, then I make every effort to get speakers I really like and I have succeeded in doing so over the past decades."
Different amps are NOT about tailoring FR because even most tube amps will have little effect on the FR response. Its about tailoring the DISTORTION response, specifically going to an amp that has less audible distortion vs. one with more audible distortion.
"Now, though I find JA's dispersion measurements to be very revealing, I think the NRC measurements are more accurate, and even JA likes to compare his results to them when a manufacturer such as PSB provides them. And the NRC measurements as shown on the Soundstage site, linked below, are in many respects extraordinarily good for a speaker, such as the Listening Window curve in Chart 2. While you might prefer the response of the S2 with a tube amp, then again you might not. In any case, if I want a more downward sloping response with some recordings, there is the Tilt control on my Quad preamp. Much handier than changing amps!"
The tilt control doesn't change the distortion character of the amp and so the same audible problems remain. Likewise, a FR measurement says nothing about driver and cabinet colorations, response decay measurements (ie. waterfall plot), distortion, thermal compression (which affects dynamics and tonal balance at different output levels), diffraction effects etc. etc. all of which contribute to the final speaker sound. Driver and cabinet noise can be riding as little as 10db below the main signal in the worst case. Thermal compression can cause all dynamic peaks to be truncated for one driver over another (low sensitivity drivers heat up quite quickly), resulting in a tonal balance shift on peaks and/or sustained high levels. Distortion at crossover points is common and if that point is in the 1-4khz range possibly quite audible. High Q resonances in the cabinet and port store energy that smears later signals. All of these dynamic effects have just as much impact on the sound as FR.
This is why the Wilson X1 Grand Slaam sounds terrific with a great amp even though it has a far from perfect frequency response. Dynamically, the speaker is very well engineered. Freedom from cabinet colorations at all levels, lower distortion than most amps, no thermal compression at sane listening levels, low order crossovers so low Q circuits and thus no ringing, low Q tuning on the bass system etc. etc. Just about the ONLY thing it does wrong is FR and we both know that this is easy to fix these days.
In JA's opinion, my speakers are not particularly difficult to drive. Are you seriously trying to maintain a Quad 606 can't handle them?morricab
"Its not all about FR. I have done experiments where I made equalization curves for different amps to the same target. The amps still sounding drastically different...not just slightly different but the difference between finding the music interesting and boring (talking strictly about acoustic music here)."How closely did you equalize the FR of the different amps? What controls did you use? What were the blind test results?
I could ask the same questions about your comments on distortion. Do you have audibility data?
I've never heard the Wilson X-1 Grand Slamm. Collom's displays are quite different from JA's, I must say. The FR results in Stereophile don't look all that bad, though Colloms' curves are more smoothed than JA's, and the dispersion isn't bad, either.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"In JA's opinion, my speakers are not particularly difficult to drive. Are you seriously trying to maintain a Quad 606 can't handle them?"Name? What are these wonderful mystery machines? The Quad 606? Whatever floats your boat.
"How closely did you equalize the FR of the different amps?"
The limits of the RTA are around 0.2db and I try to keep each band within 0.5db of each other. This is measured in room with pink noise with a microphone at the listening position. This is about as close as one can reasonably do without a TACT.
"What controls did you use?"
Not sure exactly what you mean here? A blank? I can bypass the equalization with the push of one button. There is no blank in this situation because SOME amp has to be driving the speakers or there is no sound at all. What controls do YOU use? Do you even equalize your system? How many different amps have you tested with NO equalization?
"What were the blind test results?"
Tests were not conducted blind. Results are based on long term listening impressions. I have done single blind tests with preamps and the results were pretty conclusive that differences were easily audible...I would expect the same with amps but it is far less practical with amps.
"I could ask the same questions about your comments on distortion. Do you have audibility data?"
I made a test cd of solo violin with distortion added as per the software that Keith howard made available after his stereophile article. Most of the added distortion was audible compared to the undistorted original but certain types were clearly less pleasant than others. The original undistorted was the preferred track.
There have been published studies on the audibility of distortion. I don't have them immediately avaiable but if you search I am sure you will find some.
The X1 is not a terrible measuring speaker but it is far from the flattest. I have seen $500 speakers with a flatter FR and that was my main point that this flat measuring $500 speaker sounds far worse than an X1 even though its FR is far better.
A basic technical control is level matching. I would think using a volt meter at the speaker terminals would be a better way of matching levels than your method. The audio clubs long ago discovered that when decent amps were EQ's to within .1 dB, no one managed to prove under blind conditions that they could tell the difference as long as the amps were in the operating range. When you can do that, then maybe you'll get some results that others might take seriously enough to want to replicate.morricab
"Name? What are these wonderful mystery machines? The Quad 606? Whatever floats your boat."Hummmphhh! You have chosen to comment on my speakers without knowing what they are. Under the circumstances, I see no reason to tell you. There's nothing mysterious about them. E-stat knows and you could look them up just as easily as he could. I suggest Inmate Systems.
Quad 606. Well, I like it, and it can drive most speakers quite well, as can lots of other amps.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"The audio clubs long ago discovered that when decent amps were EQ's to within .1 dB, no one managed to prove under blind conditions that they could tell the difference as long as the amps were in the operating range"Proof? Show me don't tell me.
"You have chosen to comment on my speakers without knowing what they are"
What? Calling them wonderful mystery machines? Not much of a comment really. I can't access inmate systems from work (it is blocked). Maybe I will look when I am home.
This is the comment you made, and I already linked the post:morricab
"I shudder to think how bad your speakers must be if you can't hear the difference between a good tube amp and a crown amp."
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Well maybe if you told me your speakers I would no longer think it was your speakers that can't resolve the differences...just your ears.
![]()
Of course, that part of the sentence is incorrect, totally made up, not anything I have said, and I addressed that in the post below it. Of course I have come expect that sort of thing from tubeguy, E-stat, sometimes Clark, and now perhaps I'll have to expect it from you.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
I see now why you are obssessed with NRC measurements and FR response because you own paradigms. I was also kind of that way when I owned them because afterall it is sort of the their USP. You have to believe their marketing hype to buy the speakers of course. Oh and they don't sound too bad...but not great either.What I realized along the way in this journey is that flat FR is one of the easiest things to fix and not even that important to getting a live sounding system. Transparency, coherenecy (including the time domain), low level resolution, dynamics, freedom from compression, low coloration, and low distortion are all just as important and more importantly, not correctable by DSP.
You are apparently still in that stage of the audio hobby where you think the speaker is the only or at least by far the most important part of a system. This is most likely due to the fact that you have only mediocre components in your system. Switching with other mediocre components brings no real improvement. Of course one needs a sufficiently resolving speaker to begin with. You can do much better than the Paradigm S2 (not new maybe but there are plenty of better used speakers to be found). You can do seriously better (the whole system that is) for not that much money.
What I have found is that once you have a speaker that is truly transparent, low in coloration, high in coherency (means time aligned or single driver usually), low in distortion, and highly resolving of low level information (something paradigms do not excel at doing) then the speaker no long becomes the limiting factor in your system. Gear differences become so obvious that you wonder why others even argue about it.
![]()
The NRC measurements are probably the most accurate available to me and they work pretty well for me. They are a useful screening tool, and I see absolutely no reason to consider speakers that don't measure reasonably well. Such speakers don't sound good to me. Measurements are a tool and I fail to see anything obsessive about that.As I've sometimes said before, the kind of measurements available to me can tell me what speakers not to get but they can't tell me which ones to buy.
Let's look at the factors you propose:
-Transparency: a piece of optical imagery, not really very clear in its application to audio. Some mean sonic invisibility, in which respect the S2 is excellent. Otherwise, this must require an even FR.
-coherency (including the time domain): God knows what you mean by coherency. Time and phase coherence do not seem to be that important in speakers save as they affect the FR. In any case, the S2 is subjectively very quick, and I have owned Quad ESL-63's, but they don't really fit our room acoustics very well. The S2's sound better here.
-low level resolution: I wonder whether people who use such jargon have any idea what they are talking about.
-dynamics: this usually refers to punch in the bass. Anyway, I listen well within the dynamic limits of my speakers.
-freedom from compression: Soundstage has begun to measure this and even an inexpensive PSB speaker seems to do well in this respect.
-low coloration: And you were denigrating the importance of even frequency response? GMAB But of course, the dispersion counts, too, not merely the on axis response.
-and low distortion: well, Soundstage has measured this for quite a while and the S2 does pretty well.
-Resolution is something you also mention: it's also an optical image whose applicability to audio is again not too clear. I suppose it means one can hear what's on the recording.
"Gear differences become so obvious that you wonder why others even argue about it."
Gee, I want to listen to music, not gear differences.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Well the fact that you have difficulty understanding some generally accepted audio language makes this discussion rather difficult, or are you merely being obtuse? I assumed that at least some of these definitions are obvious and at least a few have been explained and discussed enough that you know what I mean.However; I will address the points and give you what I am thinking about when I use these terms.
1) I never said even frequency response wasn't important, just less important than you give it credit because of its relative ease of correction. With modern DSP, FR is the least of a speaker designer's worries.
2) Transparency "a piece of optical imagery, not really very clear in its application to audio"
You are the only one who would question this. Yes it is an optical analogy but one that is quite clear (no pun intended) to most people because the auditory equivalent is the feeling that there is nothing between you and the original sound. It mainfests itself in that on recordings with natural acoustic one can hear the ambience in the recording clearly, for example. In this sense it is related to low level resolution but low level resolution also includes clearly perceiving subtle dynamic shifts at low levels, not simply did you hear the sound or not. Does the character of the instrument retain the correct timbre and expression at that low level or is it indistinct?
3) Coherency: The sense that the sound is coming from a unified source and not multiple sources. Nearly every box speaker betrays the drivers it uses to make sound. This is due to crossover anamolies, differing driver materials, sharp cutoff, cone resonances etc. Time coherence is another matter and why I put it in parenthesis. It has been my experience that speakers which are time coherent (for example full-range electrostats or a speaker like the Thiel CS3.6) are also more coherent in the first sense of the word. An example of one of the least coherent speakers I know is the B&W 802N. This speaker uses a different material for each driver and high order slopes. Hearing the transition between drivers is painfully obvious as a change in the coloration of the timbre of instruments. It has very little to do with FR.
4) Low level resolution: see point 2 above. I will repeat, it is not just if you can hear a soft sound or not (although with some speakers and systems this is an issue) but if that soft sound retains the complete characteristics of the thing making the sound, including dynamics and timbre. MOST speakers in the world have problems at the low level sounds. Nearly all conventional medium to low sensitivity speakers fail miserably.
Example: Some friends of mine have Apogee Scintillas. These are very high resolution speakers; however they are a bit old and one friend has had them rebuilt and the other has a stock pair. They did a speaker cable test (quite important with a 1 ohm speaker) comparing the Speltz anticables to some DIY silver cables. They did this test on the rebuilt Scintillas that have the same DIY silver cable as internal wiring. They had a track of a new woman singer from Norway done in what seems to be a home studio. On one track they noticed very quiet in the background a dog barking(presumably outside the studio). They had not noticed this with the anticables but it was clearly audible with the silver cables. When they went back to the anticables and listened carefully for the dog THEN they heard it. However; it was much less distinct and barely noticable as a dog. Back to the silver cables and the dog was distinct with timbre of its voice and probably even would be recognizable to the owner. Oh they were using only 1/2 meter long cables and this was still obvious.
The other friend took the recording home on his stock Scintillas, which are wired with 20 year old Monster cable and he has 3 meter speaker cables. He couldn't hear the dog at all!!! It was simply not there no matter how hard he tried to hear it and even knowing exactly where it came in. Clearly information was being lost and now he is going to have his speaker cables shortened and rewire his speakers. This is an obvious example of what I mean by low level resolution.
5) Dynamics. This has nothing to do specifically with bass. Does a trumpet's blast rely on bass response? What about a Cymbal crash? Dynamics has to do with a speaker's or system's ability to go from one level to another level with lifelike speed and in the case of big amplitude shifts do so at ALL frequencies equally and without compression. Not just bass. Speakers that are restricted in dynamics in some frequency band will exhibit a certain "character" that is not obvious from a FR measurement (same is true for amps and sources). If a speaker sounds "dark" even though the FR is nearly flat then it suggests that the high frequencies don't have the same dynamic character as the mids or bass. If it sounds bright and the response is relatively flat then it is possible that the bass driver has begun compression whereas the mid and tweeter have not.
"Anyway, I listen well within the dynamic limits of my speakers"
Do you? Probably if you listen only to compressed music then you are right. If you listen to relatively uncompressed jazz or classical then I doubt it.
Let's be clear: Dynamic range is not just about how loud something can play. It is also about how SOFT a speaker can play and retain the correct character of the instruments (see my thoughts on low level resolution). The key word here is RANGE. From soft to loud not loud to louder. If you listen to an orchestra at REALISTIC levels then how good is your system through the soft passages and does it make it through the loud ones uncompressed? That is what I mean by dynamics and dynamic range.
6) Freedom from compression: Soundstage's measurements are steady state and not necessarily indicative of dynamic conditions.
I have seen studies that show 86 db/watt drivers that begin to show the effects of thermal compression as low as 90 db. Normally, doubling the power gets you 3db more output from the speaker. What they found was that above 90 db or so they were getting only about 2 db per doubling of power and then at higher levels only about 1 db. Eventually the driver will cease to get louder and you will probably melt the voice coil. VCs can heat up very quickly on big sudden peaks and are relatively slow to cool down thus affecting the signal that follows the big transient. This is called hysteresis and its reality is that big dynamic peaks DO compress for most drivers and at lower volume levels for lower sensitivity drivers. Tweeters are often much higher sensitivity than woofers in normal speakers. Many speakers get edgy sounding when pushed. Is it distortion?? Maybe. Cone breakup? maybe or maybe its also a momentary imbalance between the outputs of the drivers due to thermal compression.7) Low coloration: This has very little to do with FR so I am not sure why you bring it up. Coloration is the sound the speaker makes by itself when the drivers are in motion and is correlated with but not a part of the actual signal that was put into the speaker and what the drivers are putting out. This manifests itself as: Harmonic distortion, driver flexing colorations, cabinet energy storage and release, crossover anamolies, resonances etc. This altogether can be thought of as "self-noise" or the noise the speaker makes itself that is not part of the original signal. It rides typically 10-30 db BELOW the main FR and rarely, with exception of strong resonances, makes a noticeable impact on the FR. However; it is most definitely audible and probably one of the main reasons that two speakers, even perfectly corrected for FR can sound very different (along with dynamic behavior).
Often, the lower the speaker's "self-noise" the better that speaker will retrieve low level information while retaining the proper character of the sound. Transparency and natural soundstage will also improve (assuming the recording is made in a natural space). Most speakers are so "noisy" that most people don't even realize that all that garbage is riding along until most of it is gone. This relative lack of "self-noise" (all speakers have some) is one of the reasons (along with freedom from compression and dynamics) why big Wilson speakers (like the MAXX and X2) can sound quite lifelike despite their relatively poor FR.
8) "Resolution is something you also mention: it's also an optical image whose applicability to audio is again not too clear. I suppose it means one can hear what's on the recording."
See points 2 and 4 its not just what's on the recording but if the correct character of that thing is properly preserved. When I say high resolution I am also including the other points like dynamics, coloration, and transparency. Clearly there is some overlap in the terms and some terms are used in the definition of others. We use optical analogies because they are ones that are easier to understand. It is not easy to describe hearing phenomena.
""Gear differences become so obvious that you wonder why others even argue about it."Gee, I want to listen to music, not gear differences. "
This is a disingenous statement and a cop out because obviously in this discussion we are talking about how the gear affects our perception of music. A high resolution speaker will give you insight into which gear and cables are giving you the full information and which ones have audible problems. However; if the speaker is mucking things up too much then the ability is flawed. You can still hear differences but it becomes more difficult to tell different from right or wrong.
The right speaker and right gear will give you more of what is on the recording and presuming the recording is a good one then more music.
![]()
I've heard much of the jargon in the audio subculture. Sorry, I just don't think most of it makes sense.I may or may not be the only one who questions the use of optical imagery in audio. For example, we have an idea what transparent means in relation to glass. It means you can see through it. Now, in audio, you explicitly relate it to feelings: "the feeling that there is nothing between you and the original sound." Nothing objective there.
Ah yes! The B & W 802N, a speaker I have heard and which certainly is not invisible. I didn't like it, either. I suspect the uneven off axis dispersion below the crossover between the midrange and the tweeter is one fault. When one talks of FR with speakers one must really include dispersion as well.
As is clear from my Inmate Systems entry, I listen mostly to classical music.
I'm so far not into DSP.
I don't know about barking dogs. Don't usually hear them at concerts. Sounds like a noise floor issue to me but then we really don't know what is going on. And you never will as long as you continute with your 'black box' approach of simply swapping pieces. I should also point out that, even as stated from your own perspective, you have not traced the problem to speakers. Do you mean that the barking dog can't be heard on other speakers?
I am simply astonished by your assertion the FR has little to do with coloration! You put it down to distortion. In any case, the distortion of my speakers is quite low at normal listening levels.
Dynamics--well Soundstage is beginning to measure compression, and did so with this inexpensive but nice measuring speaker:
http://www.soundstagemagazine.com/measurements/speakers/psb_image_t45/
Low level resolution seems to have a lot to do with Fletcher-Munson effects, also noise levels.
There are a number of things I can think of that effect whether a speaker sounds harsh at high levels. Fletcher-Munson effects can do that, too, among other things. So we haven't gotten totally away from FR.
"A high resolution speaker will give you insight into which gear and cables are giving you the full information and which ones have audible problems. However; if the speaker is mucking things up too much then the ability is flawed."
Well, there we are into optical imagery again. I'm afraid that a 1 ohm speaker will help create problems with a lot of amplifiers which can't drive it. I can't see that that makes it particularly revealing otherwise, or that it has much applicability to speakers designed more sensibly. If I wanted to drive Apogee Scintillas, I would get much bigger amp amp with lots of current capability. But my amp seems quite adequate to drive most speakers, including the Quad ESL-63 and my Paradigm Signatures S2's. I fail to see, for example, that the speakers are mucking things up just because less powerful amps can drive them.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"Nothing objective there"
Not so. It is an observation based on what one is hearing. This can be quite objective and reproducible. Not all complex reactions to what one observes can be easily described in language. This is why we use visual analogies becuase language is far better at explaining visual observations than auditory ones. This is probably because human vision is a more highly developed sense and thus more directly related to our survival."When one talks of FR with speakers one must really include dispersion as well."
I am always referring to room response...never simply on-axis response.
"I don't know about barking dogs. Don't usually hear them at concerts"
You miss the point. I doubt it is noise floor in the conventional sense. I have heard his system and there is no hiss or hum.
"And you never will as long as you continute with your 'black box' approach of simply swapping pieces"
I have no "black box" approach just by swapping pieces. I think about the why and then think about which pieces are most likely to affect that. This is also what my friends have done with regard to the wire. See the problem is that you invalidate observation and experiment as invalid tools for making a correct sounding hifi system and I tell you that it is essential to having a correct sounding system. You observe a phenomenon, change a thing, then observe again. Eventually, insight can be had without having to change all the time. This is experience from experiment and observation...something you seem hell bent on denying yourself. My friends and I are careful observers and not afraid to find the root cause.
"Do you mean that the barking dog can't be heard on other speakers?"
That is one possible implication. It is also possible that the other friend's system is lacking resolution elsewhere. If he swaps the wiring and finds the dog barking is still missing then obviously the problem was not simply the wiring, although replacing 20 year old copper wire that is probably not oxygen free and therefore corroded is not a bad idea.
"I am simply astonished by your assertion the FR has little to do with coloration! You put it down to distortion. In any case, the distortion of my speakers is quite low at normal listening levels."
That is not what I said. It does of course affect timbre of instruments but again it is easily correctable so who cares. I don't put it down to distortion in the sense you are thinking of distortion, ie. Harmonic distortion. By this measure, most modern speakers are quite low in level and harmonic order. I am talking about NON-harmonic distortions, cone self-noise, cabinet energy storage, port resonances, thermal compression, dynamic compression etc. etc. Paradigm doesn't specify these things and therefore you have no idea if your speakers are low in this kind of distortion or not. You seem very obtuse about these very important aspects of speaker design.
"There are a number of things I can think of that effect whether a speaker sounds harsh at high levels. Fletcher-Munson effects can do that, too, among other things. So we haven't gotten totally away from FR"
No one is suggesting that FR is not an issue. In fact it is probably momentary, ie. during peaks, FR imbalances that make things sound bright or dark. This issue though isn't the FR but what causes the FR to become imbalanced? This goes back then to compression and the like. FR is the symptom in this case not the cause. The speaker might measure flat at a constant level but heat up the woofer VC and how it behaves differently from the tweeter, which has a different thermal compression curve, not to mention a different amount of energy being put into it.
"A high resolution speaker will give you insight into which gear and cables are giving you the full information and which ones have audible problems. However; if the speaker is mucking things up too much then the ability is flawed."
Well, there we are into optical imagery again. "
How so? I see no optical imagery in that quote. Resolution is a general term not necessarily related to vision.
"I'm afraid that a 1 ohm speaker will help create problems with a lot of amplifiers which can't drive it."
Maybe so but we are talking about changing cables with the same amplifiers that ARE capable of driving it.
"I can't see that that makes it particularly revealing otherwise, or that it has much applicability to speakers designed more sensibly."
What about ultra low mass, low coloration, low distortion drivers with essentially NO thermal compression (only dynamic compression at high levels)? This means the speaker retains the same character over very wide dynamic range (particularly as you go down in level). This is what the Scintilla brings to the table along with a pretty smooth FR. The character is maintained even at very low drive levels...something most dynamic speakers fail miserably at (including paradigm). Normal suspension cone drivers have to much resistance to motion at low input levels (overcoming surrounds and spider mechanical resistances necessarily limits what input signal gets a motion out).
"I fail to see, for example, that the speakers are mucking things up just because less powerful amps can drive them."
You fail to see this because no one said this. I am generally an advocate for lower powered amps anyway. I see the fact that the Scintilla needs a high current amp as its one true weakness. Its why I like other, easier to drive Apogees better overall.
A speaker mucking things is not directly correlated to its ease of drive. Lowthers muck things up pretty badly in a lot of ways: poor FR, no bass, highly colored "paper" sound. However; they are highly efficient (100db watt), which means that they move a lot with very little input and thus do superb low level reproduction and within their limits are highly dynamic. They have the opposite problems of most conventional speakers that get the FR right and maybe have less coloration but are really undynamic sounding because they don't play soft well nor do they play loudly well.
You just told me that the “auditory equivalent” of transparency “is the feeling that there is nothing between you and the original sound.” There is nothing objective in that definition and it says really nothing about the equipment or its performance but about feelings.I also fail to see that I have denied anything in your observations. Hearing a barking dog seems off hand a fairly unambiguous thing. I do wonder about your explanations, though. Here is one of the things you said about your methodology: “I think about the why and then think about which pieces are most likely to affect that.” You still don’t know technically why the barking dog is heard in one case and not another. What on earth is that wire doing? You don’t really know and that’s why I call it a black box methodology, although perhaps “plug and play” would be more appropriate. Plug and play is suitable for some types of problems, of course, like troubleshooting by a process of elimination.
"Dynamic" seems to mean different things to different people. There's one fellow here who touts his speakers, which have a hump in the upper bass to lower midrange area, who then regards some other speakers which are capable of pretty high output in that area as undynamic. Parenthetically, I think a driver is much more likely to heat up on steady signals than short transients--in the old Audio magazine, they tested some speakers up to over 10,000 watts in the tweeter range--which they surely could not stand continuously.
As for resolution, well, I’ve read high end reviews for quite some years and I am aware of some of the kinds of things that are said. Just because an amplifier isn’t suitable for driving a very difficult load such as that of the Apogee Scintilla doesn’t mean it’s deficient for driving lots of other speakers. But I’ve seen reviewers try to extrapolate from the performance into a very difficult speaker load.
But let’s stick to the barking dog. I wonder just how far down that barking dog is? It would be interesting to measure for that could raise some interesting possibilities. I could conceive of scenarios having nothing to do with cables. You compared two expensive cables—how about some 12 or 14 gauge speaker cable from the hardware store? Would the barking dog still be audible?
As I understand it, you propose that dynamic drivers have enough resistance to moving that they are likely to muck up low level signals. Now, that at least is a testable hypothesis and perhaps you should test it. It should be measurable. If you want anecdotes, I can turn the Quad preamp down to its lowest level and I still get a lot of detail and stereo spread--actually with any of the main speakers I have had: the old Kef 104s, the Quad ESL-63s, the PSB Stratus Minis, and the Paradigm Signature S2s—that old Quad 606 is very clean at low levels. I don’t have the CD with the barking dog in the background so I can’t test this (I might not even be interested in the music).
I do wonder how a speaker can maintain the same audible character at extremely low levels without something happening to counter Fletcher-Munson effects. The Quad ESL-63, for example, isn’t linear in the bass at higher levels, that is, it starts limiting itself in the bass at more than moderate levels. So we get more relative bass at lower levels. The Kef 104 or 104aB is quite robust in the bass and laid back in the upper midrange, and it also played very nicely at very low levels.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"You don’t really know and that’s why I call it a black box methodology, although perhaps “plug and play” would be more appropriate. Plug and play is suitable for some types of problems, of course, like troubleshooting by a process of elimination.
"Well in this case that is probably what is needed because 1) In order to develop a model of how electrical signals propagate down a wire one needs an appropriate model. I guess the work of Malcolm Hawksford is a good place to start (although Jneutron might not agree). 2) I need detailed information on the metallurgy and design of the wires. This information I largely do not have.
So I can't really tell you mathematically what is going on with the wire, can I? I COULD maybe measure the signal loss at various frequencies through the wire, IF I had such an analyzer. If I was a serious cable manufacturer I would have such a thing. As I am not, I don't have one. Neither do most audiophiles. So since I can't really measure the cable I guess that is out of the question also.
I plug and play and use the analyzer between my ears. Its very sensitive and has the ability to be quite specific but it has problems with reliability and reproducibility, which is why it requires long term exposure to be certain of real effects vs. imaginary effects.
However; hearing a dog, that presummably was supposed to be on the recording and maybe wasn't heard by the engineer making the recording, and then not hearing the dog or that it flies below normal detection is something that is repeatable. It is also something clearly related to the passing of low amplitude signals.
As the only part of the chain that was changed, it seems obvious then that one can lay the blame at the cable. Plug and play or I prefer trial and error, is a perfectly valid way to test such a thing. Using a human detection system (ie. ears/brains) is a valid way to determine the audibility of such a thing and/or its relative realism.
""Dynamic" seems to mean different things to different people"
That's the problem, dynamic means one thing. If people use it incorrectly that's another issue. I expect you to use it correctly.
"Parenthetically, I think a driver is much more likely to heat up on steady signals than short transients"
I will try to find the sources but I don't think this is correct. A light bulb sure heats up and glows pretty darn quickly don't you think? A big surge of current will heat up a coil instantly but it will cool down only as function of the airflow and the coil materials.
"As for resolution, well, I’ve read high end reviews for quite some years and I am aware of some of the kinds of things that are said. Just because an amplifier isn’t suitable for driving a very difficult load such as that of the Apogee Scintilla doesn’t mean it’s deficient for driving lots of other speakers. But I’ve seen reviewers try to extrapolate from the performance into a very difficult speaker load."The Scintilla is a very special case because no other speaker is as brutal on an amp. Most of the amps that work without dying are not very good sounding as revealed by the brutal (in the resolution sense) Scintilla. So in effect it is a doubly brutal speaker. Amps that would sound good on it won't drive it and the amps that can drive it generally don't sound so good. We have now found some exceptions and a tube amp is in the works (really a 1 ohm tube amp!!).
"But let’s stick to the barking dog. I wonder just how far down that barking dog is?"Good question. I will be getting a copy of the cd soon. I hope I hear the dog on my system!!
"I could conceive of scenarios having nothing to do with cables. You compared two expensive cables—how about some 12 or 14 gauge speaker cable from the hardware store? Would the barking dog still be audible?"
Under the circumstances of this test I don't know how it could have been anything else. They were the only variables changed in the test (the first test where they heard the sound in both but greatly diminshed in one vs. the other cable). The other system could have multiple explanations (like cd player, preamp, amp, speaker's internal wiring, ambient noise). Sure they could add cheap 12 gauge wire to the test next time, why not. Maybe the dog is MORE audible! I for one am not a proponent that cables need to be expensive to be good (I personally don't have expensive cables).
"I do wonder how a speaker can maintain the same audible character at extremely low levels without something happening to counter Fletcher-Munson effects."
It doesn't matter because your brain already knows what things SHOULD sound like at low levels. Fletcher-Munson works with real sounds as well. The change heard with speakers at low levels is ON TOP of the usual changes in how you hear with level. Loudness curves on stereos in the past were to combat the problem with the stereo systems behavior not your hearing mechanisms behavior. So a stereo system that maintains the correct response, FR wise and dynamically will sound natural because the instruments sounded more or less that way live and that is what you are used to hearing from real sounds.
"The Quad ESL-63, for example, isn’t linear in the bass at higher levels, that is, it starts limiting itself in the bass at more than moderate levels. So we get more relative bass at lower levels"
This is a dynamic range limitation of the Quad-63 and means that the speaker is probably more correctly balanced from low to middle volumes. This is true for a lot of electrostatic speakers. However; some have a much higher range upward while maintaining the downward range (Soundlabs and Acoustats come to mind here) because of their nearly massless drivers.
"The Kef 104 or 104aB is quite robust in the bass and laid back in the upper midrange"
Then it likely had a dip in the FR through the presence region, which once corrected with equalization, it would no longer sound so laid back. OR it has some dynamic constriction at the top of the coaxial midranges passband or some funny breakup modes? How good is its balance at low levels compared to a top electrostatic speaker? That is the question.
![]()
The old Kef 104 was pretty flat in the listening window but with the 8" woofer crossed over to a 3/4 tweeter at 3 kHz the off axis response suffered between about 1000 and 2500 Hz, which would affect the total power response. You can't really fix that with an equalizer, though it can help things to a degree, as it won't improve the dispersion. It was quite a good speaker in the right set up, but not really room friendly. Compared with the Quad ESL-63, it really didn't sound bad at all, but not as detailed.The ear is less sensitive to both low frequencies and high frequencies. This has nothing to do with the deficiencies of stereo systems. It's easy to look up on the net.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"The ear is less sensitive to both low frequencies and high frequencies. This has nothing to do with the deficiencies of stereo systems. It's easy to look up on the net."THat's my point! In real life with real live instruments your hear sensitivity also drops at high and low frequencies with decreasing level. When we hear the highs and lows dropping out with a stereo at low levels its because there is EXCESS loss over and above our normal lack of sensitivity in these ranges. THis loss is coming from the stereo itself. If it were to maintain the same FR and dynamic envelope at 50db as it has at 80 db then the softness would sound natural as when an instrument gets softer from 80 to 50db. Since most stereos don't sound natural at low levels this is a clear indicator of excess loss from the stereo not because of the change in hearing sensitivity. Loudness contours were designed to overcome the limitations of the gear not the listener. Your brain knows what sounds natural at low levels and so the hearing curve is irrelevant for live sounds. If the response of the stereo system is constant regardless of level then the sounds on the recordings will soften in a natural sounding way but most systems are not constant thus the loss of realism at low levels (and high levels).
![]()
Well, just from a layperson's point of view, I think it's a little more complicated than that. First of all, we have your idea of "natural" sound, which has not been verified. Second, in real life, sounds farther off are often softer but have more reflected sound. Third, turning the volume down often makes the stereo image seem to move back some. Fourth, like it or not, at lower volumes, the ear is less sensitive to low and high frequencies.Many posters here besides my self have suggested that speakers which sound good at low volumes tend to approximate the Fletcher-Munson curves (or someone else's equal loudness curve).
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"First of all, we have your idea of "natural" sound, which has not been verified"True and the same can be said for when you say your system sounds fine as well. However; are you a regular goer of live classical and/or other music that is unamplified? Are you a musician or have you lived with someone who is a professional musician? Do you now or have you ever been involved in the making of recordings in natural spaces (ie. a concert hall)? If you answer yes to these then I perhaps we are coming from the same background.
"Second, in real life, sounds farther off are often softer but have more reflected sound."
Yes this is correct; however, we are discussing the systems effects on sounds that were recorded from farther away or that are soft but close. You don't want the system to impose the same effect as a live sound from far away. It is like a double negative, its a no no!
"Third, turning the volume down often makes the stereo image seem to move back some."
Why should this be? Is this simply your observation in your system or do you have some physical explanation? If the system is properly preserving the harmonic content of the instruments when you decrease the volume then it shouldn't happen. With a real instrument, like a sax or trumpet, the sound recedes more inside the instrument at lower volumes than when it is played full blast. This has a lot to do with how the harmonic content of the instrument changes with the volume level its played at.
If I understand you correctly you are saying that a loud trumpet (as it is on the recording) played back softly (as in the volume control turned down) will sound farther away than it would if the volume control was set to give a realistic SPL level. IMO, this is a clear sign that your system has a different set of harmonic components that it adds at the low level vs. the high level or the speakers FR balance is different at low spl level vs. high spl level. Either way this would seem to indicate a problem with that system. I don't disagree that this receding of the sound with decreasing level could happen just that it should not happen.
"Fourth, like it or not, at lower volumes, the ear is less sensitive to low and high frequencies."
No one is disputing this. However; the microphone doesn't suffer from this problem, which means that soft sounds are captured with close to the original harmonic content and freqeuency balance. The reproduction will then sound natural (ie. you will hear the same content on the recording that you would have heard live if you were there) IF the system is not further altering the harmonic content and FR with level. If it is, namely losing bass and HF response at low levels, then the sound will lose life at low levels and no longer sound realistic. As long as these are preserved at low levels it will sound like a natural softening not the usual lifelessness most systems exhibit.
For example, whenever a reviewer says a speaker needs some juice to "wake up" this is a clear sign to me that this speaker cannot perform naturally with low level signals, this means either when played at low volumes or when there are soft sounds (like ambient information) in the louder matrix of the music.
"Many posters here besides my self have suggested that speakers which sound good at low volumes tend to approximate the Fletcher-Munson curves (or someone else's equal loudness curve)."
Maybe they are purposely designed this way to compensate for the losses that typical speakers show at low levels (the losses are always there just more noticeable at low levels where the percentage of error is larger). However; now this speaker will not sound correct at higher volumes because the bass and treble are tipped up. Manipulating frequency response is not really the answer to what is really a loss of dynamics. The only way I can think to do this in the frequency domain is with DSP where the DSP monitors the current to drivers and then applies a sliding scale equalization curve dependent on the signal level. It can be fast enough but I don't know anyone making a level dependent equalizer. Maybe a bright idea though.
It is the same problem with Class AB amps. They have zero crossing distortion that makes itself most known at low levels and is a major reason why these amps don't sound good at low levels because the distortion to signal ratio is much greater than at full power. The zero distoriton is always the same in absolute level terms.
![]()
Of course, my idea of natural sound has not been verified, either.Musician? Well, I've been a choral singer most of my life and have been capable of baritone solos for some decades. I've done some recordings. Just heard the dress rehearsal for Messiah last night (I have laryngitis, so couldn't participate). Maybe I'll be able to do some of the performances . . .
Yep. Speaker with a depressed midrange will sound that way played louder too, but that can be quite pleasant on a lot of material. But it doesn't prevent them from sounding great at low levels.
Walking away from a live performance is not the same as turning the volume down.
You'd better look at those equal loudness curves again. Sensitivity in the bass drops quite a bit as the level goes down.
When you get around to measuring the behavior of speakers at low levels I'll give your theories more credence. Perception of the distance of the sound source seems to be a controverted area with natural sounds and the sound image from a stereo system lacks many of the clues we can have with natural sounds.
Oh, I just looked on the Stereophile site and there is an interesting article by Keith Howard on voice coil heating and changes in impedance. He thinks it's likely to be a negligible factor in normal listening.
http://www.stereophile.com/reference/1106hot/
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
http://www.rythmikaudio.com/servo_tech.htm
http://www.jblpro.com/pages/general_faq.htm#What is "power compression"?
http://www.nearfieldacoustics.com/whiteppr.pdf (read the section on dynamics where they state that even small compression effects in the bass can have big effects on perceived loudness).
http://www.klippel.de/download/bin\AN12 - Amplitude Compression.pdf"Don’t Make a Bad First “Compression”
All these SPL ratings are 1w/1m ratings. “But what type of performance can I expect if I apply more than 1w of input electrical power to my subwoofer?”Well, it’s quite simple: you suffer from compression. There are two types of compression to be concerned with:
1. Thermal/Power
2. BLAs electrical power is applied to a voice coil, the voice coil heats up. This causes an increase in the Resistance of the voice coil, also known as Re. This is referred to as thermal or power compression.
As the voice coil moves out of the gap, motor strength (also known as BL) begins to drop. The further the voice coil moves out of the gap, the lower BL drops. As BL decreases, output drops as well. A 30% decrease in BL will cause a 3dB decrease in acoustical output.
Now which is worse? I guess you could say that depends. Thermal/Power compression is typically a slower process. Because music is very dynamic, it does take a few seconds for the voice coil to increase in heat, and therefore, a few seconds for the voice coil’s resistance to increase. Once this voice coil is hot, it takes 10-30 seconds to cool back down. As you can see, once the voice coil has become hot, the effects of Thermal/Power compression are fairly constant.
On the other hand, BL compression is a "quicker" issue. A voice coil can move right out of the gap and back into the gap in a single second. As excursion increases, it only gets worse. It is quite obvious that BL compression is the more serious issue of the two.
Don’t Make a Bad First “Compression”
All these SPL ratings are 1w/1m ratings. “But what type of performance can I expect if I apply more than 1w of input electrical power to my subwoofer?”Well, it’s quite simple: you suffer from compression. There are two types of compression to be concerned with:
1. Thermal/Power
2. BLAs electrical power is applied to a voice coil, the voice coil heats up. This causes an increase in the Resistance of the voice coil, also known as Re. This is referred to as thermal or power compression.
As the voice coil moves out of the gap, motor strength (also known as BL) begins to drop. The further the voice coil moves out of the gap, the lower BL drops. As BL decreases, output drops as well. A 30% decrease in BL will cause a 3dB decrease in acoustical output.
Now which is worse? I guess you could say that depends. Thermal/Power compression is typically a slower process. Because music is very dynamic, it does take a few seconds for the voice coil to increase in heat, and therefore, a few seconds for the voice coil’s resistance to increase. Once this voice coil is hot, it takes 10-30 seconds to cool back down. As you can see, once the voice coil has become hot, the effects of Thermal/Power compression are fairly constant.
On the other hand, BL compression is a "quicker" issue. A voice coil can move right out of the gap and back into the gap in a single second. As excursion increases, it only gets worse. It is quite obvious that BL compression is the more serious issue of the two."
A well designed voice coil system shouldn't move out of the gap unless very long excursions are happening so thermal is the bigger problem at realistic volume levels.
Almost every speaker maker in audio is concerned with the effects and not many mention FR at all. Could it be because compression affects realism in sound more than FR abberations, perhaps?
"Yep. Speaker with a depressed midrange will sound that way played louder too, but that can be quite pleasant on a lot of material. But it doesn't prevent them from sounding great at low levels."I disagree, a depressed midrange most certainly affects things from sounding correct at any level. Speech intelligibility for one thing suffers. I have heard this with a number of speakers that have a dip in the presence region. Also, when the level gets lower that 3-5 db or so dip may now drop the response in that region below the level where the speaker performs correctly thus further reducing the clarity of sounds in that frequency range.
"You'd better look at those equal loudness curves again. Sensitivity in the bass drops quite a bit as the level goes down"Not sure what point you are trying to make here. It would be helpful if you completed the full thought rather than give a one liner and expect me to interpret what you mean. I try to be as explicit as possible so as not to be misunderstood. I frankly have now idea what you want to make of this obvious factoid. It seems like a non-sequitor.
"Oh, I just looked on the Stereophile site and there is an interesting article by Keith Howard on voice coil heating and changes in impedance. He thinks it's likely to be a negligible factor in normal listening."
I read this but I question his methodology a bit. Here is another test done with small hifi and pro monitors (not very different from the kind of speaker Howard used) and they show quite demonstrable differences in the performance due to thermal compression.
http://www.soundonsound.com/sos/jul02/articles/monitors2.asp
If you do, the least of your worries is thermal compression.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
No the average levels are that loud but the peaks could be because I often listen to wide dynamic range recordings (like 20+ db over the average level). So if it takes only a few ms to heat the coil up (one large peak) and 10-30 sec to cool it down then it is clear that thermal compression can be a problem for the music following the peak(s). The point is that the response of the system is not constant but instead constantly changing subtly. This can affect imaging stability also (for instance if the woofer in the left speaker has sublty different dynamic behavior than the right speaker woofer).
![]()
First of all, it is not the alleged roller coaster impedance curve that causes the audible frequency response variations, but it is in great part the of the interaction of the output impedance of the amp with the impedance curve of the speaker.Same thing. Speakers with linear curves don't exhibit the +/- 1 db variations.
First of all, that the FR variations of amps with a high output impedance into many speakers loads is likely to be audible, sometimes While you might prefer the response of the S2 with a tube amp, then again you might not. quite audible, whereas the FR of low output impedance amps varies very little into most speaker loads as compared to driving a resistor.
Indeed. 1 db variations are audible. 7 db variations are grossly audible.
JA, who took the measurements of the Paradigm Signature S2 for Stereophile, commented that "the balance overall is impressively flat," and it is.
We must be reading different reviews.
"The graph is impressively flat from 80Hz to 20kHz, though with slight excesses of upper-bass and mid-treble energy apparent. The former goes some way toward compensating for the S2's lack of mid- and low-bass output, while the latter is not unexpected, given my feelings about the speaker's slightly forward treble balance."
In JA's measurements, there is no such thing as a "+8 / -13 db response" in his curves above the bass.
My mistake. I was looking at the crossover plot. The summed response is only +7 / -5.
While you might prefer the response of the S2 with a tube amp, then again you might not.
Well the errors remain complementary given the extreme impedance curve. I'm not much of a box speaker fan anyway.
Where is the summed response +7 -5??? There ain't no +7 anywhere near the audible range. I can see why you could misread the divisions above 10K, but I assure that high frequency tweeter resonance is way up about 27K or so, and not even you can hear it. JA didn't mention it, possibly presuming that by now readers would know how to read his graphs, but to go by your example, evidently a dangerous assumption. Anyway, here's an earlier review where JA did describe it. See his remarks above Fig. 2:http://stereophile.com/standloudspeakers/272/index6.html
So much for the +7 dB!!
Now, the -5 dB. This is a very narrow interference dip on axis around 12.8 kHz. Being very narrow and quite high, it's not very audible. As well, the dip does not occur in the off axis curves, again reducing its audibility. If you read the review, you will see that JA commented on that as well, but apparently it slid past you. Now, the research of Dr. Floyd Toole has shown that off axis dispersion affects the sound of speakers a great deal in the real world, a conclusion accepted by JA, BTW, among others. But you totally ignore that.
You also have utterly failed to see one of my points, which is that the broad modifications introduced by using using the tube amp specified would probably be quite audible--in fact, certainly more audible than that ultrasonic ringing which even you can't hear, and also more audible the very narrow interference dip around 12.8 kHz on axis. So much for the significance of the -5 dB dip, whose audibility you exaggerate.
If the impedance curves are flat, then a good tube amp and a good SS state amp should have virtually identically FR's into the speaker load. So I see no argument for preferring tubes.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Throughout the range of 100 hz to 10khz there is more than twice the variation in the speaker response than the +/- 1 db variation introduced by the tube amp. We'll totally ignore the fact that the errors are complementary.
![]()
Where do I start? Lessee here.
+3 db at 150 hz
-2 db at 300 hz
-4 db at 800 hz
-2.5 db at 1.5 khz (entire octave from 1khz to 2khz depressed)
-2 db at 6 khz
+3 db at 10khzrw
Well, we've talked you down to an extremely narrow -5 dB dip at 800 Hz (which you have thus far ignored for some reason) and a +3 dB peak at 10 kHz . I should point out that JA's measurements of the upper bass tend to introduce a hump anyway, and that the NRC anechoic measurements give a somewhat different picture.Point one, we all know that amplifiers have smoother frequency responses than speakers. You seem to be attempting to make some point or other by repeating this, but you are far from clear as to what it is.
Oh, and by your standards, the response of the amp you showed is about +1, -1.5 dB into the simulated speaker load--I personally would ignore that narrow little dip, but since you don't, I won't. I point out that this will be audible.
Second, if you look at the NRC measurements, it is not at all clear that the variations in the FR of the tube amp into the simulated speaker load is complementary to the the on axis response for the Paradigm S2. In any case, if you prefer tubes, that's fine with me. I don't and for some reason that seems to bother you.
In any case, you have not made a very useful or significant comparions between the FR of the tube amp and that of the S2 speaker. An amplifier's frequency response is expressible with a single amplitude vs. frequency plot whereas a speaker radiates sound in all directions, and the FR in various directions is different. If the microphone moves even slightly in relation to the speaker, the FR will be different--something which is very clear in the NRC measurements and is also ascertainable from the dispersion plots in Stereophile.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Compare that with the latest measurements from the Wilson Audio Sophia 2 and you will see that these deviations from the amp are small compared to the speaker. Are there other speakers with much better frequency response measurements? Yes there are. Are they that good in room? Not usually. Now, if you use DSP correction you are not only correcting the speaker's response but the whole system's response, rendering damping factor completely irrelevant for frequency response.I am more interested in the CHANGE in damping factor not its value.
Read this article for some more background:
http://www.audioholics.com/techtips/audioprinciples/amplifiers/dampingfactor.php
If I were to use DSP, the Crown amp would seem to be a good one. Why on earth would I want a tube amp?
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Because I am sure the crown wouldn't sound worth a damn, that's why. And if its response is so perfect, why would it matter if it were DSPed or not? You should go for it anyway. Maybe when you get the mud out of your ears and realize that the speaker is not the end all be all of sound reproduction, you will finally "get" it. I shudder to think how bad your speakers must be if you can't hear the difference between a good tube amp and a crown amp.
![]()
My speakers are Class A Restricted LF in Stereophile's Recommended Components, if you need testimony other than mine as to how good they are.morricab:
"Because I am sure the crown wouldn't sound worth a damn, that's why."How would you know whether it wouldn't sound good?
Besides, I thought we were talking about the hypothesis of DSP, in which case one would want an amp with some real guts--which tends to be very expensive with tube amps. As you point out, the FR differences would be compensated for with the DSP, assuming the amp has the guts for it.
DSP would correct the room response at the listening position, so I don't see what your point is about the speakers. However, the speakers, the room acoustics and set up, signal processing, and the program material are the biggest determinants of the sound.
How you get the idea that I would think that one can't tell the difference between tube amps and SS amps (strictly, amps with a high output impedance vs. ones with a low output impedance), I can't say. That has, after all, been proven in some cases using double blind tests--and the results of one such test was published in Stereophile, with measurements that made it quite clear why the two amps sounded a bit different on some material.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
Not for AR1,1W,3,3a,LST,10pi,11, or 9 especially if you like pipe organ music. This unit probably costs many thousands. Excellent by 1956 standards, not very good by 2006 standards. It's not on my Christmas gift wish list either. But if I saw one in the salvation army store for $50 or by the curbside.....
![]()
My point was to counter Morricab's assertion. A damping factor of 10 will likely result in audible variations in the FR with many speakers.Actually, the list price at the time of the review was $7000 USD! I could think of a lot of amps that are cheaper that would do better. But I think it would drive most speakers, though it's not a very good value for the performance.
If I were in the market for a powerful amplifier, the Crown would be a good bet.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
![]()
"My point about damping factor is that it indirectly shows that the feedback is losing its "corrective" power when the damping factor drops (one of the main features of negative feedback being a lower output impedance). The subsequent rise in HF distortion also points this out."Your statement that negative feedback doesn't work at high frequencies is just plain WRONG and for you to repeat it over and over again demonstrates you are just plain IGNORANT.
Do you know what a phase locked loop is? Every radio, television set, computer, cd player, dvd player telephone switch, even your hand held cell phone and many other electronc devices made within the last 30 years has them, sometimes many of them. They would be impossible without them. It not only depends on negative amplitude feedback but on negative phase feedback to work and it works beautifully up well into rf frequencies, by now probably well into microwave frequencies. You can rail against the principles of negative feedback all you want, it is not going away, it is a bedrock of modern electronics engineering and control systems. The prior art is JUNK.
BTW, distortion in that Crown amplifier can rise as high at 20 khz to a whopping 0.05%. in terms of sound that's less than three parts in ten thousand. Even you can't hear that, no matter what its harmonic distribution.
![]()
"BTW, distortion in that Crown amplifier can rise as high at 20 khz to a whopping 0.05%. in terms of sound that's less than three parts in ten thousand. Even you can't hear that, no matter what its harmonic distribution. "Prove it.
Do you have measurements like this one below that show what you claim to be true? Didn't think so.http://www.soundstagenetwork.com/measurements/amplifiers/belles_350a_reference/
This sums up nicely what I am talking about. If we work on the assumption that 1) Negative feedback lowers output impedance (ie. increases damping factor) and 2) Negative feedback lowers THD
Looking at the measurements it is clear that the damping factor begins to drop at around 200Hz. Likewise, the THD vs. Frequency is also increasing around this frequency. If we take the two points above then it is obvious that the negative feedback is no longer effectively lowering the output impedance and is also no longer effectively keeping the THD as low. It is also reasonable then to expect that the nature of that distortion is likely changing as well.
Far better to start with a more linear design than to "correct" to spec.
![]()
I would like to remind everyone here that Soundlab apparently has two amplifier types in operation. One is the Crown--- Big, rugged, and dutiful, like a Mac truck, and the JC-1 mono power amps rated at 800W ea. into 4 ohms. Why?
I also challenge anyone to show where this amp is modified to sound different from the signal going into it. Come on Soundmind, put your 'proof' where your mouth is.
![]()
"BTW, distortion in that Crown amplifier can rise as high at 20 khz to a whopping 0.05%. in terms of sound that's less than three parts in ten thousand. Even you can't hear that, no matter what its harmonic distribution. "
Not that it really matters, because the harmonics are all well above the range of hearing, but maybe we should ask Morricab just what the distortion is of his non linear tube amps at 20 kHz? Maybe we can expect an answer such as "the harmonic distortion of non linear tube gear is so much more musical than linear solid state gear". The late Bud Fried used to ask me all the time; "what do they mean when they say more musical"? I had a number of answers for him, but none are suitable for print.
d.b.
![]()
Dan, you will never see me write that such and such distortion is more musical than any other kind of distortion. I don't believe it and i have said so. I do believe that some distortion is more AUDIBLE than others. THerefore, the amp that is doing the LEAST amount of AUDIBLE distortion will be the better sounding amp.I am not a "subjectivist" nor am I a hard core "objectivist". I believe that he measurments and the listening are both important (the listening guiding the measurements not the other way around though).
OTL amps can for example be flat in response to several hundred kilohertz with very low distortion at high frequencies. Remember that radio broadcasting and television were (and in some cases still are) transmitted by tubes, which can deliver full power into the MHz range...something transistors struggle with.
![]()
I give up: you are twenty or more years behind the times and refuse to catch up to the present. Wake up Rumplestiltskin: it's 2006.
BTW: will you mind if I quote some of yourposts for my upcoming article?
d.b.
![]()
Sorry, couldn't resist :-).
![]()
Yes I mind. That means you do NOT have permission so don't use any of my quotes here.
![]()
nt
![]()
No, just don't want quotes to be taken out of context or to be your pin cushion with no recourse. So don't quote me in your article.
![]()
This post is made possible by the generous support of people like you and our sponsors: