 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.169.192.194
As a "data point" and discussion for those who have been singing the praises of hi-res audio downloads, I did a test recently.Recently got the very versatile E-Mu 0404 USB (AKM AK4396 DAC) to play around with on my computer (quad core 2.8GHz, 8G RAM, low DCP latency, Win7, through USB of course). With some recent high definition downloads / DVD-A source:
Rebecca Pidgeon - The Raven: "Spanish Harlem" (24/88 Bob Katz 15th Anniversary Ed)
Carol Kidd - Dreamsville: "When I Dream (2008)" (24/96 Linn Studio Master)
Laurence Juber - Guitar Noir: "Guitar Noir" (24/96 AIX DVD-A rip)Took these FLAC/WAV files, down sampled in Adobe Audition to 16/44 (no dither, no noise shaping) then resampled back up to 24/96. Verified that frequencies all truncated to 22kHz. Then listened to them with Foobar 2000 ABX comparator using the E-Mu ASIO output plugin. This allows me to A-B on-the-fly and do some "blind" ABX'ing.
Listened with headphones: Audio Technica ATH-M50, Etymotics ER-4B.
With this setup, I figure I've removed all variables except for sample rate change - same mastering, same DAC running at same sample rate.
Results: Essentially NO DIFFERENCE between the native 24/96(88) and 16/44. Blind ABX results NO SIGNIFICANT DIFFERENCE. When I do the rapid A-B switch in the middle of a song, I thought there MAY have been slightly more smoothness/openness in the high-def version but this could just be placebo and the improvement was MAYBE 5%.
At 38 years old, very few loud concert experiences, I don't think I have 'tin ears' (hey my wife thinks I have better ability to pick out music in noisy environments so I guess it's at least as good as some females :-).
My conclusions:
1. Either my equipment sucks or these samples suck and there's alot more but I need to fork up more $$$$.
2. Or high-def cannot be well appreciated with headphones.
3. Or the upsampling back from 16/44 --> 24/96 somehow reconstitutes the sound.
4. Or, there's really not much difference.
5. At this point I'd probably spend a few more dollars to buy a high-def download (maybe at most $5-10 more if it's something I like) when given the option but not expect significantly more revelation in the sound.I've listened to good SACD as well and like them but there's no way to do tests like this. I didn't bother with 24/192 material since I figured most improvement should come from this first step up 44 --> 96. Anyone else done such tests for themselves?
Edits: 02/13/10 02/13/10 02/13/10Follow Ups:
The poor sonics of the 0404 card are not helping you.
If you can swing the $$$ go for the Lynxstudio card, if not, the audiophile 192 from M-Audio is a better sounding card for a little more $$. I have three DAWs set up, one with the 192 and two with the Lynxstudio TWO card.
If you have the original high rez tracks, why not downconvert to 16/44.1, save as second track and then you can switch back and forth between tracks, very quickly with editing software.
I have many recording session masters at all 4 common digital word depths/sampling rates (16/44.1, 24/96, 24/176.4, 24/192). Its interesting to note, that an original raw 16/44.1 session master does not sound as good as the exact same track recorded on a different recorder at 24/176.4 that has been down converted to 16/44.1.
24/96 does edge out 16/44.1, but it still sounds like digital audio. The higher res formats begin to sound more ligh high speed analog tape session masters.
best.
If either yes or no, why bother with this test? You are the customer.
I agree, that's why I bother to post this piece.
IMO, I have *rarely* felt I've heard a difference between hi-res or 16/44. My main system at home "only" costs about $10,000 but even when I go to show rooms with > $30K systems, I stick in CD's of music I've burned or listen to their SACD/DVD-A, I usually cannot tell the difference, and even if I do, it seems to be mastering differences. The only exception was one time about 5 years ago I repeatedly could hear more detail on the cymbals in Telarc's 1812 SACD vs. RedBook layer through my friend's Sony SCD-1 and Merlin speakers. (Is there a good 24/96 1812 I can test out?)
When I read folks here talking about the seemingly miraculous differences, I think to myself either I'm crazy or my ears must be bad. Yet family and friends must have bad ears too because no matter who I have listen with me (unless they know about hi-rez disk vs. regular), they usually don't comment on anything different or can't seem to tell a difference either (many are musicians and 'audiophiles').
It just seems to me that the folks who claim to hear differences attributable to higher sampling rate / bit depth are few and far between, yet tout this experiential difference as if gospel.
he just doesn't have the ability to hear the difference. Many people don't have trained ears or hearing sensitive to discern the differences. Just like many people don't have a palate that is trained/sensitive enough to discern the differences between vintages of wine.
You might be right...
However, I'd like to know if you've tried for yourself; cuz if you have and there is a significant difference, I'd sure like to better my abilities with some tips.
it's the e-mu.
If you listen to the tonal quality of most music you may not hear the difference between 44.1 and higher sampling rates. Depending on the filters used the high frequencies can be dulled, made edgy, or preserved at the expense of the soundstaging. If you aren't listening to the right thing at the right time it may be very hard to notice a difference. Even if you are focused in on a real difference you may still miss it, because the mind plays tricks. For example, if a particular sound can be heard in one format but not the other, when switching back and forth the mind may fill in the missing pieces, thereby making both versions sound the same.
Hearing the difference between 16 bit and 24 bit resolution will be very difficult on music that has little dynamic range or lots of background noise. Similarly, hearing the difference will be difficult if your system isn't resolving in bit depth and has low noise. Also, if your room isn't quiet any subtle low level differences may be masked. If you listen to acoustic music recorded in a reverberant room (e.g. a well recorded symphony orchestra) you may be able to hear anomalies in the reverberation tails as the music dies out.
If you were to use an editor extensively and downsample many hi-res recordings at a variety of settings you might (eventually) become clued into the subtle differences involved. If you buy a RBCD that has been reduced to 44/16 by a skilled mastering engineer most people will not notice a large difference for most musical genres.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Thanks for the wise words Tony.
Will definitely need to log more hours listening to build up skills to appreciate the nuances... Certainly not a bad thing :-)
I have limited experience at home, but in the Reference Sampler provided by Weiss with the Minerva, on some the symphonic pieces I really hear more information and locational information. When I have checked high definition sites, I find little that tempts me. But I now have the capability to download hi-def. I must say, however, that I am on a DSL access, so downloads might take all night.
Overall, I do not find hi-def. equal to the improvement I hear between a computer server and an optical transport.
Any way to verify if those reference samplers used the same mastering on the hi-res vs. non-hi-res? That's I think one of the main problems. I find I can't trust that I'm listening to the same source / mix say between the SACD vs. Red Book layer when trying to make comparisons.
There have already been many suspicions about the Red Book layers being purposefully degraded in order to sell us on the merits of hi-res.
As far as I am concerned the only way to assure a consistent test of formats is to do the mastering yourself, i.e. start with a single hi-res version and down sample it to a lower res format. That way you know the source of all differences you may hear. This also allows you to experiment with the various settings of SRCs, i.e. when you place the 10 pounds of hi-res music into the 5 pound RBCD bag you get to select which portions get discarded.
If you compare professionally mastered versions then you will not be able to separate out decisions made by the mastering engineer and characteristics of his equipment from limitations of the format. This will be true even if both versions were done by the same engineer using the same equipment. As one example, if the engineer has a "hot" monitor with excessive high frequency response he will be biased toward smooth anti-alias filters with gradual roll off and away from heavy noise shaping, whereas he may favor other trade offs when using monitors with rolled off high frequency response.
It is also a good idea to place little faith in anything you hear on a strange system. It's best to do comparisons using equipment and a room that you know well, in which you have hundreds (or thousands) of hours of listening experience using a wide range of recordings. Finally, for various reasons (technical or otherwise) there may be gain differences across versions. If you compare two versions of the same recording at levels that differ by more than about 0.1 dB there is a good chance that psycho-acoustic factors will distort the results.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Yes, obviously for you personally, as listened on your system, there's little to no difference - and that's the one and ONLY takeaway from your test. And, as long as we do not extrapolate it to any other listener and any other system, it is perfectly valid.
As others suggested, you'd probably want to try something better than EMU-0404 USB, to give yourself better chance to hear the differences. Also, you did not mention any optimizations done to the computer (HW and SW) - if there indeed are none, your results are even less surprising. Look at the threads dedicated to cMP and cPlay to get an idea what has to be done.
Last but not the least - please, drop ABX comparator in Foobar. It's sound-degrading piece of garbage, that should not, and cannot, be used for any tests or comparisons. Just un-install that plugin, and never bother with it again.
In terms of computer optimizations, I do have alot of RAM (8GB), fast processor (Core2Quad 2.8GHz). I've tuned DCP latency down with turning off C1e and EIST in the BIOS, also shut down unneeded drivers. When I listen with Win XP, the DCP latency is down to 5us with no spikes though Win 7 seems to have higher latency overall but DCP Latency Checker stays in the green zone (can't remember how many us).I have tried cPlay. Honestly, again I can't say I hear much difference if any between this and ASIO output on the E-Mu with Foobar. Will try ReadyBoost with some high speed UDMA CF cards as some folks have suggested; I guess this may decrease noise/interference from the 7200rpm drives but can't see how this is any different from using cPlay.
As for Foobar (I'm using newest 1.0 release) ABX. Is there a thread discussing why it sucks or degrades sound quality!? I see it's sending the audio via ASIO so how is it modifying anything? Because I ABX'ed with WAV files at the same sampling frequency & bit depth (24/96 or 24/88 whether native or through 16/44 downsampling step), ABX doesn't even need to switch any of these parameters toggling between the samples (no loud clicking or pause to resync).
Ultimately, I hear the criticism of the E-Mu 0404 USB and sure, maybe this is the limiting factor... Wish someone like John Atkinson would run these budget interfaces through his test battery just to have some comparison data though (of course, I'm sure many will disagree whether the results are useful or valid).
Edits: 02/15/10
Try your media comparison with a really good converter, like a Prism. The differences are easily discerned. Taken a step further, 16/44.1 sounds really unsatisfying compared to 24/192.
Happy listening.
Regards,
JerryS
So... You're telling me I need to fork over $4500 for an Orpheus or $9000 for the DA-2 so I can tell the difference?
Really wonder how many folks out there downloading hi-res material have this level of DAC capability!
Maybe I'll see if I can rent one of these units at the local pro-audio store when I have time over the summer :-).
I did not mention the DA-2, but I think the Orpheus would be quite sufficient to be enable most folks to clearly hear the differences between any digital format if:
1. Their hearing acuity is up to the challenge; and
2. The rest of their audio system has sufficient resolution.
Pro-audio stores often rent gear. It may give you another data point to consider.
A good data point.As a single test it proves nothing to us (although it may be 100% conclusive for you personally for your current setup), but a large body of such tests might enable us to make some sound conclusions about 16/44 vs. 24/96. I personally have also not heard significant, recognizable differences between 16/44 and 24/96 on my system in blind tests (nor has my wife, whose ears I trust more than mine).
This single data point is infinitely more valuable than a subjective listening test, however (e.g. "Guys I downloaded the new version of UltimatePlaybackProgram and it sounds so much better than version 9.7.4.2.1!!! Check it out!"), which not only hold no value whatsoever for us but also has probably caused further confusion for the test subject himself.
Edits: 02/14/10
"infinitely more valuable than a subjective listening test"
That's odd since it was one. But I too regard it as valuable.
Rick
I believe that, in this context at least, blind tests such as those conducted by the OP are not considered to be subjective (vs. a test where somebody hooked up some new piece of hardware, listened to some familiar audio, and posted a detailed description the changes they were hearing relative to their memory of their system's sound before the new device was added).
"blind tests such as those conducted by the OP are not considered to be subjective"
I guess it's down to considered by whom. They are unquestionably subjective in my book, but do provide a quantified estimate of the detectability of a particular (but possibly uncontrolled) stimulus by a particular population.
Glad you responded, I've got a question for you: Have you ever heard a difference in sound quality between various SW 'players' on a computer? I mean of course unexpected differences.
Thanks, Rick
Yes, I've heard differences (sorry for the late reply...I realize this might be a bit buried). The biggest differences I remember were between iTunes and Foobar2000 around 2007...but I chalked this up to iTunes doing something to the audio data (some kind of DSP "enhancement"). Beyond that, I've never heard a significant difference as long as the two programs in question were outputting the data in the same manner (e.g. both using ASIO, KernalStreaming, etc) and not adding any kind of effects to it.
I don't doubt what you heard at all. Unfortunately multiplying the sample rate or bit depth just doesnt give you muliples of resolution. You get that with pictures but with audio the realtime reconstruction is flawed to a greater or lesser degree depending on the hardware. I don't buy the RFI argument that's often put across, I think it's just down to timing. SSD hd's improve the sound, so does adding more RAM. As you're running W7 try adding a bunch of USB sticks as Readyboost memory, I've been trying this out lately and this also seems to improve SQ.
Myself, I think all these tweaks tighten up the timing of data packet delivery but I've no proof of that. I also expected that with the PC dialled in better I'd hear more difference between low and high res material. Actually they both improve and still remain quite difficult to tell apart sometimes even through high end gear. Disappointing in some ways but when 44.1 can sound so good not that disappointing really :)
With basic buffering (something any programmer should be capable of and should be implemented in all audio playback software) there should be no difference between storage devices due to their transfer speed.
This is easy to verify if you are skeptical...copy the same audio file to a RAM Disk (the fastest possible storage device) and to your slowest storage device (usually a hard drive) and compare the sound of the two (ideally, using a blind test...otherwise I am fairly certain you will proclaim the RAM Disk as sounding better thanks to my description of it being the fastest possible storage device).
If something's stored in a Ramdisk that doesn't mean it doesn't then interact with other hardware on it's way out of the PC. As I said I'm interested in timing issues not outright speed. Speed isn't an issue when you're transferring cd quality material for example, sequential reads are at 200+MB/sec on my SSD HD. I've set up a Ramdisk at times, also for example set up foobar to buffer whole files. You get subtle differences same as if you source the data from wired network, USB, wireless, HD etc etc. I'm talking about the interaction of RAM and cached/paged memory. Random read and write performance is much slower in a PC. Have you tried adding Readyboost devices to Vista or W7 to buffer paging, can you comment on what you hear?
But I got some 24/96 music from HDTracks.com, converted to Apple Lossless, and dropped into iTunes. Thought they sounded great through my new DAC, which is hooked up to the iMac via TOSLINK. Come to find out that the iMac is only outputting 16/44 audio by default, and if the digital cable is unplugged, it always reverts to that. Whoops. So this time I ensure that iMac is indeed putting out 24/96, but I don't hear a magical transformation. Maybe a vague sense that the higher-res music has a better sense of presence. Worse, if I leave the iMac set to 24/96 all the time, it seems to me that my regular 16/44 tracks actually sound a bit worse. Somehow more muddled. Hmm.
16/44 probably sounded worse because of the asynchronous (non-multiple) upsampling to 24/96 on the Mac. Does the Mac have 24/88 option? It should sound pretty good upsampling to that.
This is why it is handy to have a DAC that shows you what sample rate it is using (e.g. if you are playing a 44k file and your DAC says it is using 96K, something unexpected, and probably harmful, is happening to your audio data).In general, you don't want ANY resampling to be happening during audio playback. Yes, there are *rare* exceptions, usually due to poorly implemented hardware that is doing its own resampling, badly, later in the chain, so a better resampler can *sometimes* help earlier in the data chain.
Edits: 02/14/10
Home theater receivers are often useful to check out what the digital signal actually is.
Archi, you are on the right track with your curiosity and discipline. Just report discoveries as you go along.
While I do not agree with your results now, I had a similar impression long ago. I suspect that your curiosity will lead to further discovery if you keep looking. Now, real hi-rez for me started with SACDs (DSD)...and normal speakers, where the experiential dimension is different than with headphones. PC audio came later and grew on me as resolution improved.
For now, I'll single out the 0404 USB, which I've owned for years and use all the time (now mostly for the microphone input and tests). Some of the cheap USB DAC 24/192 capable stuff out there these days could widen your perspective. Even at 16/44, some of them start revealing a better sound than my 0404 USB. Search and you will find good things. Enjoy!
Which USB 24/192 DAC would you recommend JB? I might try the same methodology on my main system with the Transporter as DAC but much more difficult since it's not right in front of the computer and the rapid A-B switch would not be possible without Foobar ABX running in front of me.
I agree with you and others as well that soundstaging is compromised on the headphones and the experience is different.
I am using a Musiland 02 but there are a few others under $150 that a many people in Head-fi and here are also liking. Look around -- I have a terrible memory for names -- and you'll find them.
The Musiland 02 beats my 0404 USB both in digital and analog output modes. It also has twin headphone outputs (hi/low impedances). The $70 Musiland 01 US (no analog out) is being modded to good effect; there is a thread somewhere here, also.
There are more...they just keep coming out better and cheaper.
Thanks for the suggestion JB. $150 is certainly cheap enough for anyone to get in the game.
Do you personally notice a significant difference if you downsample hi-rez material on this unit?
dear archimago,
in general one should indeed praise you for not accepting that the emperor has no clothes on.....-but that´s only at the first sight.
the 24/96 format is not subjectively giving you higher fidelity per se, but it´s giving you higher resolution of your program material.
simplified the bit resolution is deepening the signal to noise ratio and the higher sample rate is giving you finer grain (or slices).
that in it self does not give you straight persievable better listening experience, but it widens the potential.
very few domestic digital systems can really demonstrate that obviously.
as "audioengr" is pointing out further down this thread, it´s about jitter and that´s a quite advanced subject that has a certain resemblance to power supply´s in amplifiers :
some technicians say "irrelevant" and others say "most important".
another analogy would be how important the frequency reproduction above 20kHz is.
some will say "you can´t hear that", others say it´s where the quality lies.....
now, jitter is those very fine voltage "irregularities" on the computers mainboard, -the switching power supply, -and the processors multi-tasking efforts amongst other disturbances.
this "jitter" affects, most audible, the high frequency spectrum and is therefore most obvious on systems where TIM and related distortion is kept at minimum without band limitations. this is for example an area where good equipment differs from the excellent.
I´m not implying that your gear is not of sufficient quality, nor that you cannot listen critical, but..... - the line between very good and sublime is sometimes very small in everyday life. very often one needs to push situations into an extreme to discover the advantage.
on large, complex orchestral music with choir the 24/96 format DOES reveal it´s qualities when the equipment is up to scratch, for a lot other situations 24/96 may be regarded as a bit of a PR fashion suffix.
kind regards
Thanks play,
Recommend me some music to try in 24/96 :-)
With some recent high definition downloads / DVD-A source:
Rebecca Pidgeon - The Raven: "Spanish Harlem" (24/88 Bob Katz 15th Anniversary Ed)
Carol Kidd - Dreamsville: "When I Dream (2008)" (24/96 Linn Studio Master)
Laurence Juber - Guitar Noir: "Guitar Noir" (24/96 AIX DVD-A rip)
These tracks are all minimal acoustic ballads or solos. I personally doubt anyone would notice differences between 16/44 and high-rez with any of these tracks (if RFI was somehow taken out of the equation). I think one would need either a larger acoustic jazz ensemble or symphony orchestra in order to notice such difference.
Like with Tony below, can you recommend a good piece I can look for and give a try? Always on the look out for great pieces to try out and expand my musical appreciation!
There may be differences in sound staging despite the minimal complexity of these recordings, but this is unlikely to come out when listening on headphones, since all sound stages are unnatural with phones. In addition, rapid switching back and forth will miss any of the fatigue aspects of a format, as fatigue picked up on one format will migrate to the other.
Once you get into the spirit of the ABX test you get into heavy analytic mode and focus on making decisions, thereby using a different part of the mind than that which enjoys music. So it is entirely possible for two components to sound the same in ABX tests and yet yield different musical enjoyment. The same can happen with formats, but I don't think this will be a big deal with simple material. Quite a different situation with a Mahler symphony, however!
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
So it is entirely possible for two components to sound the same in ABX tests and yet yield different musical enjoyment.Funny, in my perception it is exactly the reverse.
When AB(X)ing I’m focused on the sound. Comparing e.g WASAPI with DS I hear the differences. E.g. the piano forte is more detailed in WASAPI.
The moment I start listening to music these differences are totally irrelevant.
The beauty of Schubert’s string quintet in C - D.956 (most beautiful score ever written) is so strong that DS, WASAPI, KS, name it, with all their slight differences doesn’t matter at all.
So it is completely possible to hear differences when AB(X)ing but these differences are totally irrelevant when confronted with shear beauty.
The Well Tempered Computer
Edits: 02/14/10
" . . . The beauty of Schubert’s string quintet in C - D.956 (most beautiful score ever written) . . . "
Most beautiful score is clearly "Fantasia on a Theme By Thomas Tallis" by Vaughan Williams....
My post disappeared so I'll try again.
Listening tests are a problem for me. I get bored and lose concentration if the music bores me. However, if I like the music, I get involved and stop listening for sound differences. I'm very cautious about my ability to hear differences reliably.
> The beauty of Schubert’s string quintet in C - D.956 (most beautiful
> score ever written)
I have hundreds of favorite classical works and a similar number of favorite pop songs, Broadway shows and jazz tracks. The one I'm listening to has an impact unmatched by the memory of the work I listened to yesterday. For me, there is a new winner in the most beautiful contest every day.
Bill
Some people don't care about sound quality. My wife is a pianist and she never paid much attention to recorded sound quality, because, as she put it, she listens to what the musicians were thinking rather than the sound reaching her ears. However, some recorded sound is edgy and can create tension, headaches, etc. If your recording of the Schubert has these characteristics then you might not be so happy after listening to the complete piece. You might not catch these problems in A-B listening.
I would go with the 1952 Casals, Tortelier version of the Schubert. Wet string and tin cans would suffice with this music. :-)
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Hi Tony,
Really like your thoughtful responses over the years I've visited here.
Since I'm not a classical afficianado, any recommendation on a good/complex Mahler piece I should try out?
"Since I'm not a classical afficianado, any recommendation on a good/complex Mahler piece I should try out?"
I have a huge bunch of Mahler symphonies, multiple recordings of all of them, on tape, LP, RBCD and hi-res. I haven't done sample rate conversion tests recently, but the Darlington Mahler 6 on Acousense (from Linn Records) has lots of high frequency energy from the brass (up to 48 kHz, as the tracking was done at 96 kHz and then analog mixed and digitized at 192 kHz). At the opposite extreme, you can go with the Timerkanov Mahler 5 on Water Lily which you can get at HDtracks.com in an 88/24 PCM transfer from the SACD made by Bruce B. This is highly dynamic and minimally miked. It sounds great on my system if played at full concert volume, but needs about 8 dB more gain than the average recording to achieve this (and accordingly more headroom in the amp and speakers). If played at lower gain it fails to come alive, and so has gotten bad reviews.
You might try any of the Mahler symphonies done by Michael Tilson Thomas and the San Francisco Symphony. If you are not a Mahler fan, start with symphony number 1 or 4 if you are interested in getting into the music.
I'm sure some of the other inmates will have other suggestions.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Excellent recommendations Tony! Will look into getting them!
"1. Either my equipment sucks or these samples suck and there's alot more but I need to fork up more $$$$."You don't need to fork up more $$$$..... There is so much crappy software and artifact-laden music tracks out there, who knows what people hear what they hear........
"2. Or high-def cannot be well appreciated with headphones."
I don't think headphones have anything to do with it.......
"3. Or the upsampling back from 16/44 --> 24/96 somehow reconstitutes the sound."
I've never heard asynchronous sample-rate conversion that I thought was better than mediocre...... It's usually awful.........
"4. Or, there's really not much difference."
There are too many points of degradation taking place. The RFI on PC-based systems negates the inherent advantage of 16/44- Less number crunching per unit time means less RFI. But it must be conventional playback.
Personally I think RFI is the fundamental flaw for running any listening test on a PC.
"5. At this point I'd probably spend a few more dollars to buy a high-def download (maybe at most $5-10 more if it's something I like) when given the option but not expect significantly more revelation in the sound."
I think the two issues are RFI and excessive sample rate conversions...... The only way to properly evaluate 16/44 vs. high-rez is to convert A/D from an analog source directly to the two rates directly (two separate A/D conversions), and avoid sample rate conversions altogether.
Edits: 02/13/10
Regarding RFI, according to this review:
http://ixbtlabs.com/articles2/proaudio/emu-0404-usb.html
The 0404 using their measurements got noise levels at -111dB with 24-bit material! Wouldn't you expect much worse results?
Given that this is an external interface, other than the USB cable and maybe noise through the power outlet, the device sits probably about 5 feet away from the computer on my desk which should reduce RFI much better than an internal sound card.
Again, wish there was a way to better quantify the 0404's jitter results.
I did a similar test except it was 88.2/24 vs 44.1/16, so the sampling was not asynchronous.
I heard no appreciable difference even on complex classical.
I personally now run 44.1/24 in foobar2000, the main reason being that I sometimes like to use EQ or "freeverb" plugin (highly recommended!), and don't want to worry about truncation.
Interested in which classical piece you would recommend.
Good thought on the 24-bit output for plugin processing.
Thanks, appreciate the data point.
Rick
When I do the rapid A-B switch in the middle of a song, I thought there MAY have been slightly more smoothness/openness in the high-def version but this could just be placebo and the improvement was MAYBE 5%
Once I came across a posting about ABX testing
A guy working for Fraunhofer said that if they ABX codecs for transparency they use samples of 2 sec. max. According to him our auditory memory can’t cope reliable with a longer time span.
It might very well be that your rapid switching is a very good strategy
BTW: I appreciate that you took the time to put your testing method in writing.
An ABX is a good thing but it protects one only against one’s own bias.
A good write up of the experimental conditions is about as crucial too.
The Well Tempered Computer
Thank for the comment.BTW: I appreciate your web site!
Edits: 02/14/10
Thanks for the kind words about my website
A bit more about 16/24 in the link below
The Well Tempered Computer
"Took these FLAC/WAV files, down sampled in Adobe Audition to 16/44 (no dither, no noise shaping) then resampled back up to 24/96. Verified that frequencies all truncated to 22kHz."
Did you down sample all the tracks to 16/44 and them up sample them 24/96? If you did, then everything should sound the same because you threw out all the extra data in the 24/96 and turned them into 16/44. Upsampling them to 24/96 is not going to get back the data that was thrown away.
No, I only down-sampled to 16/44 to "simulate" what 16/44 (CD) should sound like then up-sampled back to 24/96 so I can use exactly the same file length & data sample rate to send back out to the DAC and take out the variable of possible DAC output differences between 24/96 and 16/44. This also lets be quickly A-B almost instantaneously since the DAC doesn't have to switch sample rates.
File A = native 24/96
File B = what's left of 24/96 after castrating down to 16/44 then back to 24/96 :-)
Thanks for the clarification
Methodologically, you've done virtually everything wrong, so it's not surprising you don't hear much, if any, difference.
"Everything wrong"? Absolutely no advice or recommendations?
Dude, are you in elementary school?
It is unhelpful and unkind to make such a blanket dismissal without any explanation. OTH, you my be right so it would be helpful to all if you were to spell out the mistakes you see and say how it could be done correctly.
...he just didn't hear a difference....Period.
It cracks me up when people try to debate what someone else hears. Its his house, his system, his room, and HIS ears, if His Brains says it didn't hear a difference then it didn't hear a difference.
Doesn't mean anyone else will or won't.
Music is the Bridge between Heaven and Earth - 音楽は天国と地球のかけ橋
I hope you are aware of the many times I have posted exactly that in this forum, that is, what you hear is what you hear and a report of this is a fact of observation. However, for it to be useful to others the circumstances and methodology must seem conducive to making the relevant discrimination or lack thereof. If it couldn't be heard with his method or if that method is likely to create overriding artifacts, then his perception while still true is meaningless.
It seems reasonable to assume to poster wished for his "data point" to have significance and be useful to others. He clearly indicated He felt the methodology mattered.
Useful or not, I assume we are all humans and physiologically we resemble each other more than we differ; including in terms of hearing ability. It's easy to say that 24/96 is better because the numbers are better, we rationalize they MUST be better, or advertising tells us it's better.
I would love for someone who believes 24/96 to be so much superior to tell me WHICH piece of MUSIC have they heard which when converted to 16/44 clearly is diminished so I can try for myself...
I assume we are all humans and physiologically we resemble each other more than we differ; including in terms of hearing ability.Physically the same....though perception is totally different. It is Perception that is the end product of hearing as the brain interprets the neurological messages from the ear. Hearing is in fact 100% perception, no human being can hear absolute sound without perception. We all perceive sound slightly different. And though many claim to be golden eared Audiophiles, their super hearing is still due to perception. To make an accurate test between person A and person B you will first need to train both of them to perceive sound the same. Train both people to perceive all the octaves at various amplitudes and wavelengths. Test them both to assure that both of their hearing capabilities are identical. Then play different tracks for them at different sampling rates, in the same room, in the same listening position and so forth. Now you will have an experiment with only one independent variable [the music]. As it stands you have to many independent variables, people's hearing [physical capabilities and perceptions], rooms, gear, etc. Having more than one independent variable in any experiment with always give different results.
Music is the Bridge between Heaven and Earth - 音楽は天国と地球のかけ橋
Edits: 02/14/10
Philosophically I agree with you. What you seems to be saying essentially is that art cannot be "proven" one way or another as so much relies ultimately on the qualia of experience.
Having said this, if what I claim is true (24/96 is perceptually no better than 16/44) for the vast majority of people (say... 99.9%) with technically capable equipment because neural physiology reaches threshold of perception already, I'd say that's important to know as a matter of scientific inquiry.
The stock EMU 0404 jitter is pretty high. I have one. Could be masking the difference easily.
Would the jitter be any better with the coax or optical inputs?
Has anyone measured the jitter rate?
Probably better with the Coax output, but you must use a good cable as well.
The reason that I know the jitter is high is that when I use it with a Pace-Car reclocker the difference is night and day. The difference is strictly jitter, nothing else.
Measuring jitter is another matter. You would need a dynamic spectral response, with and without signal, and the signal must not be just a single tone. This results in a series of plots. Even this thorough jitter analysis is only good for comparisons between clocks and devices. It tells you little about how it will sound. Only the ear can tell this.
Steve N.
That is correct. the jitter is pretty high, as well as the signal to noise ratio. Hard to believe its claim for 24/96, with SN ratios where they are.
16/44 format is for listening. Any higher format is intended for manipulating the audio. In listening tests, there will be no audible difference.
So why higher formats? Even though there is no audible difference, there is very good reason to record in higher format. If you will mix, eq, or manipulate audio in digital format, every change you make will degrade quality.
If you apply same effect in both formats, and then export to 16/44, they will sound different depending how much manipulation is done. Recording just one track (song) without any changes will have no audible difference in any format above 16/44.
Higher formats are intended for recording/mixing studios only. Once all is done, the exported files should be reduced to 16/44 since this will have no audible difference any more.
Using any higher format for listening is waste of space without any benefits.
My conclusions at this point agrees with this conclusion. Though there was still that nagging "openness" I thought I heard in the quick A-B switch :-).
Are there any pros/audio engineers/mixers out there who disagree?
But I believe that audiophiles are in for a world of disappointment if their only wish for the recording industry is to start shipping music in 24/96.
What their wish should be is this: Someone who knows how to make good sounding recordings and use compression, panning, digital eq and other effects minimally or at least conservatively should show the rest of the industry how its done.
Recording and mixing engineering is why most bad recordings are bad. Not sample rate.
Audiophiles don't seem to get this and are still spending tens of thousands trying to get good sound out of bad 16/44.1 recordings with hardware upsamplers. Guys here are trying to do the same thing with software upsamplers.
If you get a bad recordings, do throw it out. As an audiophile who collects GOOD recordings of GOOD music and not just good music I am much much happier and less neurotic (and have more money in the bank) than guys who think fixing a bad recording can happen with "system synergy" or "jitter control" at home. Once a bad recording is pressed it's pressed.
Toss it and move on.
Cheers,
Presto
I never use Foobar ABX because I believe it changes the sound.
Firstly, when I try it, the volume level in ABX is higher than normal for both tracks, and both tracks in ABX sound MUCH worse than each one played normally without ABX.
Interesting thought.
Anyone else verify that foobar2000 ABX not degrades sound? I've tested it a couple times and didn't think it modified the volume so long as I use ASIO to directly send data to the audio card.
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
