[lan]

Quote:

Originally Posted by Jon L Just for clarification, the 3 versions under test were: 1) Store-bought redbook CD playing in CDR drive via iTunes. 2) Same CD, ripped into .wav via iTunes ripper, playing via iTunes. 3) Same CD, ripped into Apple Lossless via iTunes ripper, playing via iTunes.

I would like you to try a few things.

Try EAC -> WAV -> iTunes VS iTunes -> WAV -> iTunes. EAC should be a better ripper and should be able to get the same extraction everytime.

Next Try EAC -> WAV -> iTunes VS that same file in foobar2000. Just making sure both are closely identical. Your Lynx soundcard does bypass kmixer in WaveOut / DirectSound right? That's a must if you're using itunes on PC.

Take the EAC -> WAV file and burn it as data CD NOT as redbook to CDR. Compare the same file on CD to the one on harddrive. I'm wondering if your CD drive is pumping noise back into your system if you use it.

Compare CPU usage in itunes on ALAC vs. WAV. I'd like to know how much more CPU power ALAC takes.

--lan

 

[Wodgy]

Quote:

Originally Posted by sgrossklass Ideally one should be comparing two files now: one converted from PCM (WAV or whatnot) to ALAC and back to PCM and the original file If the raw PCM data is not identical in both cases (ideally, both files should be identical), the format is apparently not lossless. If it is, one can only speculate - perhaps ALAC playback supports ReplayGain or can modify volume in some other way, which would mean you may not get exactly the same output as with PCM playback.

This is a very good observation. By default iTunes may enable its own form of ReplayGain (I can't remember what it's called, but it's not "ReplayGain"). There's information on the Apple Support website about configuring iTunes to be bit-perfect when used with the Airport Express. Those same settings should be used in iTunes even when you're playing back to a regular soundcard, otherwise what you're hearing isn't bit-perfect. Since .wav has no ReplayGain-style information by default, iTunes may be doing extra processing on ALAC and not WAV, and this could account for the difference.

As a file format, ALAC is perfectly lossless. (This is also true for WMA lossless, despite what some people sometimes claim.) So there shouldn't be any difference if things are configured properly.

 

[Patrickhat2001]

I'm willing to entertain the notion that the effect of placebo is overrated and that their may really be a difference in the way such files sound on THERE systems. Why? Because people are't stupid (especially those who hang out here) and (although there isn't any scientific evidence to support it) I believe that they really can notice whether files sound identical or not without imagining things. This isn't that hard people. If the differences are always the same and you still notice them after many days it's pretty likely that you're not imagining things. Instead what's probably going on is something unexpected is causing the differences. Being that the files are bit identical (and hence should sound the same) something else may be producing differences in sound. I'm betting that the differences could be caused by fluctuations in timing. The thread above that Neil linked to contains a short discussion about this.

Personally I want to get down to what could be causing such differences becaue I hear them myself. I intend to do a lot of ABX testing over the course of the next month. We'll see how I do.

 

[dknightd]

As you know, in theory a .wav should sound the same as an apple lossless. People above have suggeted good tests to help determin why you hear a difference. Another potential issue is jitter introduced by having to expand the alac file back to wav so it can be played.

 

[ADS]

Quote:

Originally Posted by dknightd As you know, in theory a .wav should sound the same as an apple lossless. People above have suggeted good tests to help determin why you hear a difference. Another potential issue is jitter introduced by having to expand the alac file back to wav so it can be played.

The sound data is stored in a buffer by the soundcard, so as long as his computer is keeping the soundcard fed, jitter should be identical. Seriously folks, short of an undiscovered bug in apple lossless or a configuration error (hardware or software), it's pretty much impossible for these two formats to sound different.

 

[Patrickhat2001]

Quote:

Originally Posted by ADS The sound data is stored in a buffer by the soundcard, so as long as his computer is keeping the soundcard fed, jitter should be identical. Seriously folks, short of an undiscovered bug in apple lossless or a configuration error (hardware or software), it's pretty much impossible for these two formats to sound different.

Can the use of this buffer be modified? The ASIO plugin (in Foobar, not iTunes) allows users to adjust the buffer size. Could, perhaps, iTunes eliminate or reduce the size of this buffer? Personally I have noticed a difference in sound between wav files and FLAC and APE files in Foobar while my ASIO buffer was set to "0" (ie turned off). Perhaps eliminating the buffer (if the ASIO buffer and the soundcard buffer are one in the same) is what caused these differences. I plan to do comparisions after the holidays to see if I still notice a difference between wav and lossless files and if changing the buffer size effects things.

 

[Icy006]

Ah, the wonders of digital audio. If I may, I'd like to throw out a few tid-bits for thought.

Music CDs contain audio encoded via 16-bit Pulse-Code Modulation, at 44.1KHz. So may the WAV files we rip from the CD. Some might think that it logically follows that there are identical 1's and 0's in that WAV file and its parent CD. The truth is a little more complicated..

CD technology is an amazing beast. For every one bit of data we're getting out of it, there are actually multiple bits corresponding to it on a burned or pressed CD. This is done in the name of error-correction, to make CDs robust - if there were no redundant bits, the slightest radial knick would render a lot of data useless. This is all at a low-level, and lots of intelligent circuits and software hide this ugliness from us mere mortals.

When the data is being ripped to a WAV file, the drive zooms along sucking up bits. If it misses some along the way, it is more than happy to back up and try again. The result is a nice linear PCM bitstream that represents the perfect digital audio sample the recording studio intended to provide us with.

So what I just said is irrelevant to this discussion. Except I'm trying to make an indirect point: a process is not always as simple as it may seem.

Say we take our favorite CD and encode it to .wav and Apple lossless files. Are these, and all lossless audio formats alike? Well, this may be drawing some assumptions:

1) All of the data is read off of CD, and if there are read errors, the drive keeps at it until it gets it right.
2) CD Audio decoding software receives a bitstream from the CD, and decodes all of it. This results in reproducible raw PCM data.
3) The encoding software for each codec takes the raw PCM data, and encodes it in a lossless manner.
4) When we play back the WAV or Apple lossless files, their respective decoders rip apart the bits and reconstruct that original raw PCM data we know and love.
5) The raw data is sent, unmolested, to the sound device (be it a PCI card in a computer, or as a little chip in a portable audio player).

Do we know for sure this is exactly what happens? Really, this is a brief summary of what actually happens with a focus on the expected end results. The audio decoding part, especially, might take some artistic license in how it deals with data starvation or corrupt blocks. Also, do we know that the audio player software is not somehow altering the decoded bitstream before it gets to the sound device? Is some obscure sound buffer configuration flag set for WAV decoding, but not for Apple lossless decoding? Are you assuming I actually know what I'm talking about?

Most processes can be safely simplified for understanding by a given layperson (though PR departments often take this to an extreme). The difference between the simplification and the whole story might be subtle, in both the means and the end result. Just like the difference between a $10,000 and $15,000 loudspeaker set. If some people are willing to drop serious cash toward subtle subjective sound nuances, then the same people damn well better trust their ears in all situations. Even if theory dictates they are nuts.

In practice, everyone else considers us nuts anyway

 

[lan]

Quote:

Originally Posted by Patrickhat2001 Can the use of this buffer be modified? The ASIO plugin (in Foobar, not iTunes) allows users to adjust the buffer size. Could, perhaps, iTunes eliminate or reduce the size of this buffer? Personally I have noticed a difference in sound between wav files and FLAC and APE files in Foobar while my ASIO buffer was set to "0" (ie turned off).

There are so many "buffers" inside the computer. Since iTunes doens't use ASIO, there is no ASIO buffer.

Low buffers for pro audio app use will yield a lower latency which is good for real time music production. But when there are interruptions in the system from other apps, it will hiccup. For music playback on a regular person's computer system, ASIO buffer set to zero doesn't make sense because they are doing other things on their computer which is dissruptive and certainly you don't need truly fast latency.

Quote:

Originally Posted by Icy006 Also, do we know that the audio player software is not somehow altering the decoded bitstream before it gets to the sound device?

For the sake of this discussion, we assume bit perfect operation. I asked before if Lynx is bit perfect in WaveOut/Directsound which will yield bit perfect in apps like iTunes. If you're using foobar2000, you can just use ASIO or kernel streaming. If your setup isn't bit perfect, you can start doing this comparision of lossless files yet.

 

[Mr.Radar]

Quote:

Originally Posted by Wodgy This is a very good observation. By default iTunes may enable its own form of ReplayGain (I can't remember what it's called, but it's not "ReplayGain"). There's information on the Apple Support website about configuring iTunes to be bit-perfect when used with the Airport Express. Those same settings should be used in iTunes even when you're playing back to a regular soundcard, otherwise what you're hearing isn't bit-perfect. Since .wav has no ReplayGain-style information by default, iTunes may be doing extra processing on ALAC and not WAV, and this could account for the difference. As a file format, ALAC is perfectly lossless. (This is also true for WMA lossless, despite what some people sometimes claim.) So there shouldn't be any difference if things are configured properly.

I would like to emphisize this. iTunes' "Sound Check" and "Sound Enhancer" features do degrade the SQ quite a bit on good systems.

 

[Woem]

Quote:

Originally Posted by Stephonovich Woem... I await the war

I'll believe the differences in output when I see some proof of it. There was paranoia that kmixer.dll was interfering with audio a while ago. Turns out that was disproven. I forget the link.

I acknowledge there may be a difference between the cd playback and the computer audio files. This is due to the difference in the DACs. However, if the CD is being digitally read and sent through the same output stream as the .WAVs and .ALACs, there should be no difference.

It's alleged "audiophiles" like the OP who manufacture artifacts (whether consciously or subconsciously) to demonstrate the superiority of their set-ups that annoys the hell out of me.

Placebo is an incredibly powerful beast. Disagree with that? Fine. Take some medical science at a post-secondary level. You'll see things differently very quickly.

 

[Wodgy]

Quote:

Originally Posted by Woem I'll believe the differences in output when I see some proof of it. There was paranoia that kmixer.dll was interfering with audio a while ago. Turns out that was disproven. I forget the link.

While I agree with your placebo comments, the kMixer interference is very real. You can verify it in several ways, either by loopback digital recording and comparing with the original signal, or attempting to play back a PCM encoded DTS/DD file to a home theatre receiver (if kMixer is munging the signal, you hear static, not music -- this test does not involve any qualitative assessment).

I have done both experiments on several configurations and can state that kMixer does indeed have an effect in the majority (but not all) driver/hardware combinations I tried. This is one reason why the Sonica + 1.2.05 drivers combination is so popular here on Head-Fi, as is the Transit + ASIO combination.

I expect my digital outputs to be bit-perfect -- equivalent to a good standalone CD player. kMixer munged audio with -92dB dynamic range instead of -96dB may or may not be audible, but it is not acceptable to me when there is no necessity for it to be this way.

 

[Patrickhat2001]

Quote:

Originally Posted by Woem It's alleged "audiophiles" like the OP who manufacture artifacts (whether consciously or subconsciously) to demonstrate the superiority of their set-ups that annoys the hell out of me.

Why would we make up this stuff just to be harassed from people like you. If anything stating that you hear such differences can only lower your creed around here. We're paranoid, not vain.

Quote:

Originally Posted by Woem Placebo is an incredibly powerful beast. Disagree with that? Fine. Take some medical science at a post-secondary level. You'll see things differently very quickly.

Actually I have a B.S. in Psychology (fitting, eh

) and am quite experienced with the effects of placebo. What you're talking about are medical studies where there is a control group who takes a placebo and a test group who takes the drug being tested. You're right in saying that many people in the control group will notice an improvement from nothing but a placebo but this is a different situation from what we're talking about. Because we're testing files on a computer we can A/B back and forth between files unlike groups of participants taking a drug and we also know which is the control file (the wav.) and what is the new file being tested (the lossless file). This is somewhat comparable to people in a drug study being able to switch back and forth on the fly between the effects produced by the real and fake placebo drug (actually such studies do exist--they're called A-B-A design studies but they occur over a longer period of time and the participants are usually not privy to what they are taking). If study participants could do such a thing I believe they'd notice very quickly which is the real and which is the fake drug (well most of the time, it depends on the drug). So what you're saying is because us audiophiles are privy to the differences between files we're imagining a difference between files that should sound identical. Is it possible that we're just imagaining everything? Yes, definately. I'm only asking that you entertain the notion that there might be something more going on. We do know that the files should sound identical and we can switch between files instantly to compare them yet we're hearing differences that shouldn't be there. I have faith that myself and others can notice real differences without imagining them it's only a matter of finding evidence to support it. Personally I intend to do ABX comparisions between wav and lossless files in the near future.

 

[philodox]

In my opinion the only limiting factor between CD and any sort of digital format is the computer itself. Obviously a CD can sound better than a WAV or ALAC if it is playing on a superior source.

Likewise the only limiting factor between WAV and lossless compression is the program used to do the ripping/encoding and the program that is used for playback.

So in a sense I agree with the rude newcomer... but I think that he needs to calm down a bit.

PS. My suggestion is listen to the music and upgrade or change your system if you find it is limiting your enjoyment. If something is only limiting your enjoyment when you are actively searching for a flaw, it is not your system that is the problem... its you.

 

[thomaspf]

To my knowledge the Lynx 2 is not bit perfect with either WAV or directsound APIs.

For the people who own an RME card, or an M-audio Sonica that bypass kmixer and play bit perfect. Can you verify that Itunes on a PC can actually be set-up to play bit perfect?

Cheers

Thomas

 

[Mr.Radar]

Quote:

Originally Posted by thomaspf To my knowledge the Lynx 2 is not bit perfect with either WAV or directsound APIs. For the people who own an RME card, or an M-audio Sonica that bypass kmixer and play bit perfect. Can you verify that Itunes on a PC can actually be set-up to play bit perfect? Cheers Thomas

It is bit perfect. Check out this page for configuring iTunes for bit-perfect playback.