The Absolute Sound (magazine) take on many aspects of computer assisted music reprodu
- Thread starter rbbert
- Start date
A quick glance shows most of the ones I have seen will import/export FLAC (along with bazillions of others).
Of course. It's just another example of how clueless the authors are, despite their "years of experience" in the audio field.
Now, in the AV Guide Forums, Dr. Chuck Zeilig posts that the assertion that identical computer audio files may sound different has been described "many years ago" and "others have also reproduced these results". More, "since publishing our work, we have heard from many people (some highly authoritative) who have heard files which measure the same sounding different". No "authorities" are named. Interesting terminology "measure the same".
Yesterday I received Part 4 of the TAS article. One important aspect is the authors warning of the strong dependance of their tests on the mains quality and power chords - something I have also found with the ARC DAC8.
Unfortunately, I don't know what this means. As has been previously posted, and is a scientific truism, if methods are flawed then all results are suspect (i.e., no valid conclusions can be drawn from this study/test). And their methods are clearly flawed.
Rbbert,
In what aspects do you consider that their methods are clearly flawed? I already had concluded that no systematic conclusions can be lead from the study , but it would be helpful if you could make a summary on the flaws of the methods they used? Most people attack the study because the results seem unacceptable, but this is not a method per se. (If you did it in a previous post please just refer the post number).
If one gets some results that are obviously false (in this case, the most obvious one is that WAV > FLAC > WAV leads to different sounding files; more on that in a minute), then either methods or analysis (or both) are flawed. Read back through this thread and you will find posts by amir and Bruce B. (at least) noting potential problems with the audio systems and data chain as well as specifically criticizing single-blind methodology (as opposed to double-blind). With respect to the WAV > FLAC > WAV issue, I myself as well as other Internet posters have run simple batch scripts to do that conversion 1000 times, and then carefully compared the initial and final files. Not only does the data check as identical (by several different methodologies), but no one has been able to hear a difference using Foobar's ABX tester or even with single-blind testing in other settings. And unfortunately that is not the only non-reproducible result, but in fact you only need one to invalidate a study.
As I've said before, in a scientific study where there are non-credible and/or non-reproducible results (as in this case), it is imperative that the investigators re-examine their methodology. That has not happened in this case, Dr. Zeilig's persistent response is "we stand by our results".
Micro
Please do not take this as an ad-hominem.. This is not truly about you. i do find it hard to find a way to defend these guys methodology and even les their "conclusions". Not much ... This TAS piece is flawed and was meant to shock and to sound sensational.. Look how much electrons it has displaced here in the WBF...
You know the "study" is flawed as well as most people here in the WBF ... Out for now
Please do not take this as an ad-hominem.. This is not truly about you. i do find it hard to find a way to defend these guys methodology and even les their "conclusions". Not much ... This TAS piece is flawed and was meant to shock and to sound sensational.. Look how much electrons it has displaced here in the WBF...
You know the "study" is flawed as well as most people here in the WBF ... Out for now
Thanks. As far as I can see you are mainly using yours and others findings to question the results and questioning the reproducibility. Although I accept the later as an objection (even the authors warn us of it), your argument stating the results are non credible is not "a flaw of the method" per se.
BTW, the criteria of reproducibility is important in science, but unhappily almost impossible to fulfill in audio. Should WBF ask permission to Harman to use their state of the art listening room to repeat the Sean Olive studies?
I think that people should not consider this study as a recipe book, but it has the great advantage of showing us that building a music server is a serious task.
Frantz,
I know the study has a strong weak point IMHO - there are so many reasons why digital can sound different that no systematic study can be made with non standard PC systems with so many non controlled variables. But, again IMHO, it is interesting to debate also the reasons of others.
...
BTW, the criteria of reproducibility is important in science, but unhappily almost impossible to fulfill in audio. Should WBF ask permission to Harman to use their state of the art listening room to repeat the Sean Olive studies?
...
Surely you don't really mean this?? Subjective or objective, reproducibility is key! What you are saying is analogous to saying that (for example) the same cables, or the same CD, sound audibly different each time you listen to your stereo?
There are any number of serious investigators working at improving computer audio, among them members of this very forum, and this is my biggest objection to the whole series of TAS articles, because they mystify the whole field. If the music files themselves aren't stable, then there is no point in even trying.
Surely I mean it. Reproducibility in science means that you must be able to reproduce the results in other systems, not just repeat it your system 1000 times.
Digital has a physical existence. If you change format of your files and reverse it, the data will be spread in the hard-disk in a different way in your disk and another person system. Also unless you both optimize the system in the same way reproducibility will be affected. The key point is that when you play it is sounds different, not when you store it. As I am now aware, a lot depends on the hardware implementation.
As far as I could understand , people are commenting the sound of the file immediately played "live" in their systems..
Stability is this thread is not not changing the 0s and 1s IMHO. It is more complicated than that. Although I respect forum members opinions, I do not expect them to find the universal recipe to assemble the perfect music server , but a serious debate on implementations.
BTW1, I have been playing with an asynchronous USB DAC , the ARC DAC8, highly praised at Computer Audiophile . I bought JMRC and DBampower licenses and assemble three music servers in three very different computers at home - two laptops and a tower PC. And they sound different playing the same FLACs ...
BTW2, I can understand that results obtained by professionals, who use optimized and stable systems, SDD large disks, with excellent grounding and mains supply systems, are much more reproducible than those obtained by audiophiles at home, but as 99% of existing systems, mine included, belong to this last group, I am mainly concerned with it.
BTW3 Gary already suggested a music server implementation. The TAS paper also has some advice. Could the WBF experts also suggest one ?
Also a key point. If they are currently working on it, we can assume that there are some problems to be solved and the system is still not perfect, as some people state.
Micro
I am yet to see you being critical to anything coming from any high end friendly corner. You seem so bent on finding them right as to use all and any argumentation fallacies to support their point of views or conclusions.
It is hard to understand where you really stand on this issue. On one point you state that your various digital music systems sound different ... I know you can suppose why as there are countless of variables that would make them sound different if indeed they do ... OTOH it isn't hard to admit that the only constant in all those experiences are the files .. come one Man! You know they remain the same throughout the playback.. yet ... it is what you seem to imply on your reference about how the data is spread on the HDD.
Your knowledge of things High-End Audio is real and deep .. And I suspect that you are an engineer or trained in some technology field ... Admit for once that this TAS thing, so frivolously called "study" doesn't deserve to be defended so vehemently ...
Similar threads
