|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.249.104.209
Good morning audio junkies,
I thought you all might enjoy this article which appeared on my yahoo homepage today:
http://tech.yahoo.com/blogs/null/33549
Not necessarily a revelation but at least confirmation of what we already know.
es347
Follow Ups:
....everyone learns that URLs go in the box marked URL on the bottom of the page.
But then there'll be that Stomp and Shout...
nt
If I compare the average classical CD I bought in the early days of CDs to those I buy now, I don't hear any loudness difference. However, my impression is that on average the more recent classical CDs tend to sound quite a bit better than the early-days ones do.
Joe
Compare recent classical remasterings (particularly the ones reissued by Universal as "complete sets") with their 80s counterparts.
the more recent remasters not only sound hotter, but when you rip them onto a computer and compare the waveforms with the original releases, you can clearly see the peaks chopped off on the remasters.
Of course, there are still very good new classical releases that appear to be only minimally compressed, but distressingly the loudness war has made it's presence here as well as for pop and jazz.
Of course, subtle and moderate use of compression actually can create an impression of a better sounding recording. That could explain why modern recordings tend to sound better.
For my own recordings, I tend to compress around 6dB prior to burning onto CD. This seems to benefit most systems (very few systems are truly able to reproduce the full dynamic range of an uncompressed recording without distortion).
I tend not to by "cheapo-label" like Universal re-issues. I was referring mainly to new recordings now being released on CD by what I consider to be the top labels. And remember I said I was not talking about all recordings, but on average.
Joe
*** I was referring mainly to new recordings now being released on CD by what I consider to be the top labels. ***
*** remember I said I was not talking about all recordings, but on average. ***
Think about this, "new recordings released by top labels" would hardly constitute the "average", by definition.
I would assert that "on average", loudness levels for classical recordings show an increasing trend. You can prove this yourself by taking a sample of CDs made in the 80s, measure average loudness, and do the same for recent CDs. They tend to differ by about 6dB.
The classical music CDs that I am buying as new recordings today are on average superior in sound quality to the CDs I was buying in the early days of CD production. That is what I was saying. At least, that is how I hear it. I have not noticed on average any level difference in the newly-recorded classical CDs that I am buying today compared to the ones I bought newly-recorded in the early days of CDs. I have not attempted to measure the average sound levels on any of these discs, however, and 6 db on average could easily escape listening notice when one is comparing two entirely different recordings made 25 years apart in different halls by different orchestras on different labels. The improvement in quality over that time is not hard for me to miss though, and that was what I was talking about. I was a very early adopter and have classical CDs going back to the first releases. The original article was claiming sound quality was going down with time on CDs because they were being made louder. I just tried to express a mild demure, trying to say that I think the sound quality on classical CDs has been on average going up with time. I shouldn't have mixed in the level business.
Joe
of those remasters that are "hotter"?
of life long before the advent of stereo as one earlier poster postulates.
As for your point about systems not being capable of reproducing full dynamic range without distortion, that's primarily because most self-acclaimed audiophiles do not realize how important it is to have at least 18 dB headroom -- which will require far more power and transient response than is capable from the garden variety of audio components. (The opportunity to hear very efficient speakers - like some horn systems - driven with > 100W amps is a real eye/ear opener.)
The statement "dynamic compression is as old as stereo sound" is not true, at least not the sort of dynamic compression that is now applied on recordings.
And your extrapolation that "compression has been a fact of life long before the advent of stereo" is even further from the truth.
Techniques such as peak limiting has been used since the early days of recording (since tubes and magnetic tape have peak limiting characteristics when stressed) but this is a different kettle of fish from the techniques applied today, because digital recording have hard limits for maximum amplitude, unlike analog.
*** how important it is to have at least 18 dB headroom ***
it's actually not headroom, just simple maths. Uncompressed music has dynamic range that can exceed 130dB, a good recording probably captures about 120dB. By contrast 99% of systems would struggle to reproduce more than 110dB of usable dynamic range.
So there's typically at least 10dB (or about an order of magnitude difference in recording levels, since the dB scale is logarithmic) between what can be captured and what's realistically playable.
If you don't believe me, send a low frequency signal to your speakers, and measure the voltage using a voltmeter. It will peak and start clipping long before you get anywhere near 110dB. For example, on a amp rated at 100w with reasonable headroom, don't be surprised if it starts clipping as low as 40w. This is because the power rating of amps can be a misleading measurement (the "power" of an amp is measured at 1% THD when the amp is well and truly overdriven and exhibiting gross clipping - typical THD is more like 0.01%). In fact to get an amp that truly delivers 100w without clipping you may need something with a rating of 300-400w.
Someone else mentioned tubes clip differently than solid state. This is true, and the reason why many prefer tubes. The behaviour of a 20w tube amp amplifying a loud signal can result in a far cleaner signal (more "rounded" waveform) than a 100w solid state. But it's still clipping - in fact more so - it's just that the results are far less harsher to the human ear (due to the peak limiting characteristics of tubes when overdriven - something that early recordings often took advantage of - hence the "warm" sound).
without regard to what you want to call it, or what methodology engineers have been using, compression in one form or another HAS indeed been in use long before the advent of Stereo. No, this is not the same form of compression that's in use today, but the goals are the same. And, while I don't doubt the accuracy of your stated numbers for this specification or that, "headroom" is, indeed, an important factor in portraying what's on the recording - regardless of format.
"Uncompressed music has dynamic range that can exceed 130dB"
Help me understand this; In live concerts dealing with unamplified instruments, based on my experience, the highest dynamic range can be experienced with symphony orchestras. The sound pressure level in a dead quiet orchestra hall is about 40dB, a live symphony orchestra can safely generate something close to 110db as experienced from the front row, the difference comes to about 70dB, give and take a few.
Some of the better classical recordings that claim no-compression at all at best can approach about 60dB, and yes very few systems can reproduce this with a realistic volume.
The only recording, so far, that comes close to this is Mahler Symphony No.2 conducted by Ivan Fischer on Channel Classics label (SACD), have you heard this recording? I would be very interested in hearing your comments on the sound of this recording.
Thanks,
Vahe
*** The sound pressure level in a dead quiet orchestra hall is about 40dB ***
Yes, but this does not mean our ears can't hear below 40dB SPL.
SPL measures ambient noise level, but our ears can easily hear below that.
To prove it, measure your living room using an SPL meter. Most rooms measure around 40-60 SPL.
A whisper is only around 20-30 db SPL. Yet one can clearly hear a whisper, sometimes even halfway across a room with 50dB SPL.
Back in 1997, Conifer Records of London, England released an Extended Dynamic Range recording of Daniele Gatti and the Orchestra dell' Accademia Nazionale di Santa Cecilia performing Respighi's "Feste Romane", "Fontani di Roma" and "Pini di Roma".
The enclosed booklet says:
"This is an EDR (Extended Dynamic Range) recording. This disc was made with the latest 20 bit technology to capture the full dynamic range of the performance. No compression was used, and the result is as faithful a reproduction as possible of the natural range of the orchestra. This gives the listener the opportunity to experience the huge breadth and dynamic variety of the recording."
The dynamic range between the pianissimo and sforzando passages, the ultrawide soundstage and the definition and detail of instruments are remarkable on this recording. I don't know what the actual dynamic range is on this RBCD, but, as you mentioned, you really have to ride the volume control to listen at a reasonable level on most systems.
Barney
DSD, if you can exceed it's massive headroom, clips softly, like tube amps ... no hideous tearing noises.
Dr. S.
Yes, DSD behaves differently than PCM when amplitude exceeds full scale. But the noise increases exponentially, together with a corresponding loss of detail. And you end up with a bitstream that is illegal to master onto an SA-CD, so you need to reduce levels, which means processing the signal.
You just aren't manacled by the absolutely no clipping mandate of PCM.
The Good Doctor
I seem to recall he said it was better to record under -3 to -6dB below full scale, otherwise DSD loses definition at the top end.
whether or not DSD behaves better at higher than 0dB FS is kind of a moot point, because the signal is illegal by Scarlett Book standards, and if you try to play it back on a DAC, the signal may clip anyway (because the Vref+ and Vref- is still set at fullscale, so the DAC is physically incapable of generating higher than fullscale voltages).
0dBfs = -6dB 'SACD' or another way; The nominal maximum level for an SACD is 6dB below PCM's absolute maximum of 0dBfs.
The maximum legal limit of an SACD's amplitude is +3.1 dB SACD or -2.9dBfs PCM. The mentioned 'Scarlet Book' also sets limits for ultra-sonic noise and DC offset.
What I actually said is that I hear problems with pushing levels up to the +3.1 maximum level and prefer to keep my maximum to 0dB SACD, with occasional peaks going up to about +1.5 dB SACD.
Really fast transients (essentially inaudible) are not really the problem; it's when the peak duration is around 200 samples or longer that it starts to become noticeable.
There is nothing gentle or soft about signals that exceed the +3.1 level; it's illegal because it is gross over-modulation of a system that behaves best with 50% modulation (-6dB), but will tolerate 76% modulation (-3.1dB) - just don't spend any 'musically significant time' at that latter level...
Now, if you'll excuse me, I have to go and finish making someone's CD really $%&$%ing loud.
:)
Graemme
NT
If a "100 watt" amplifier "clips" (minimally defined as 1% THD at 1kHZ) at 40 watts "continous" power, isn't it, in fact, a 40 watt amplifier and not a 100 watt (continous power) amplifier? Would such an amp be mislabeled under most accepted modern (last 25 years) measurement "standards", at least in the United States?
Robert C. Lang
*** Would such an amp be mislabeled under most accepted modern (last 25 years) measurement "standards", at least in the United States? ***
No, because the measurement "standard" (actually, it's not a standard, just a convention) specifies power at 1% THD, so the rating of 100watts is correct. It's just that the level at which the amp begins to clip may be far lower.
You might ask: why set the bar so high (1%)? Why not set it at 0.01%? Well, in the early days, many tube amps had very high distortion even at the best of times (0.7% is not uncommon). So if we set the bar at 0.01%, where most solid state amps operate under "normal" conditions, some tube amps never get that low, which means their power rating is 0 watts :-(
A better representation of power amp performance is a power vs distortion graph, but even then this can be misleading. The best way to measuring when clipping occurs is to do it at the speaker terminals when fed a real signal from the amp. People who have done so (and there's been a few on the asylum) have been surprised at how often their systems clipped under "normal" listening conditions.
... interesting to see this behavior in these little tiny SE amps, even with very efficient speakers. I think people would be surprised.
Doc S
.
The Achilles hill to almost all speakers, large horns are among the exceptions, is the lack of adequate dynamic range to approach "realism" at loud levels. Likewise, it is widely known that a chief cause of driver failure, especially high frequency drivers, is too little power, not too much power, when playing the speaker near, at, or above its limits.
I have found that many SACD classical recordings actually have noticeably more quiet passages that in years past but also have louder loud passages----increased dynamic range and more need for greater amplifier power.
Robert C. Lang
Headroom in a system, just like in a recording, is essential to high fidelity. It takes enormous power, even with relatively efficient speakers to produce real dynamics. A single bass drum transient can absolutely (if briefly) exaust a 500 watt amplifier feeding a moderately efficient loudspeaker
I think people clip their systems much more often than they think, and because tubes tend to soft clip, part of the growing support for tubes in the digital age is their clipping behavior.
Accurately dynamic recordings tend to annoy people, because their systems don't perform well at low volumes OR at high volumes. The volume averaging of compression is a way of addressing these limitations.
Doctor S.
Your comments on tube equipment and their soft clip behavior is right on the mark.
I always wondered why audiophiles prefer tube gear over SS, despite much higher S/N and constant maintenance.
There are also lots of comments on various SS equipment and the listening fatigue that they generate, again the preference is for tubes, from what I can tell the tube gear actually alters the sound by shaving the hard and rough edges presenting a lot more pleasant sound, to me that is tampering with the original audio signal anyway you look at it.
It seems that hi end audio is no longer concerned with preservation of the original signal, but "improving" it to be a bit more pleasant.
Vahe
Since nothing can reproduce the sound of live music, the goal among many of us who prefer the sound of tube audio systems is to mimic what we perceive to be closest to the original sound. The artifacts that result from measuring any list of characteristics against the original sound suggest that a more appealing sound can be had with tubes.
I know mine are. Sometimes I want the unadulterated truth, and other times as much sugar as I can find ... I want tone controls ... I want SS AND tubes, vinyl, AND digital ... every position in the KS.
The Good Doctor
Most super expensive hi end speakers these days come with tiny woofers, it is not at all uncommon to see 7" or 8" woofers in speakers costing $20K or more, and regardless of what space age material is used in their cones there is a limit as to how much air they can move.
The bottom line is if you play one of the current non-compressed classical orchestral recordings with one of these speakers, the speakers will automatically compress the sound for you when they face one of the high transient dynamic passages.
If you hear these recordings reproduced with a high power, high efficiency speaker with minimal compression you will be amazed as to what is possible even with the good old RBCD.
There's a thread below discussing the demise of hirez in the pop world but the survival of the classical niche so far. I think part of this is due to the willingness of some people to spend the TIME to listen. Small and independent labels are producing excellent quality classical CDs and some of us are still buying the discs rather than downloading compressed crap.
I don't know how long this will last and I also don't know if we are at saturation. How many new Brahms 4ths can you listen to in a busy day. Can you persuade the kids to listen patiently?
The "window" for multichannel recordings may be already closing and I obviously spent too much for a limited stack of software. But I'm not sorry.
A few days ago I looked at the new classical releases on SACD on SA-CD.net. I was surprised by the steady stream of what I would call non-mainstream pieces of music being recorded. Yes, there are many Brahms 4th (maybe there should be- it's a great work and I haven't heard one yet on SACD that I really like), but there is also a lot of music coming out that I have not seen available in decades, if ever. Though I didn't do a quantitative study, the pace seems to be picking up. How long it will last is another question. But I'm getting to the point where I have so many SACDs that I'm considering investing in a really good SACD palyer rather than use the cheapie that I have; I didn't think new releases would go on even this long.
Joe
What you are seeing is a new trend, the major record labels are not very active or interested in SACD, its the smaller labels that are issuing the "obscure" stuff. In a way this is the best thing that has happened to the classical recordings, I just hope to see a bit more chamber music released on SACD.
Vahe
Yes this phenomenon is more likely to occur with pop/rock music which has never been that high of fidelity anyway.
manual gain riding was very common in the heydey of analog and you would be surprised how many alterations were made at the cutting heads, just ask an old veteran, like Stan Ricker. Analog could be a very harsh mistress when it came to dynamic range.
Doctor S.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: