Apple Lossless vs iTunes Plus

Let me ask you a question, since you're convinced that they can sound different.

Take a folder full of Word documents. ZIP it. Heck, re-ZIP it several times. Extract all the files. Do your Word documents look different? Did the formatting change? What about the letters? Did new words get inserted, or some others deleted?

No?

Lossless media compression works in the same way as lossless file compression does, excepting optimizations for seeking/scrubbing and streaming; things that aren't necessary with whole-file compression. At no point does a lossless compression algorithm discard data as irrelevant (that it can't reconstruct later during the decompression). This is unlike AAC or MP3 where temporal filters are applied and resolution discarded depending on variables like bitrate, profile, etc.

The same is true for PNG image files that is true of FLAC or ALAC audio files. No matter how times you compress or decompress a PNG file, all the original image data stays preserved accurately and faithfully. That's what makes it a lossless format.

By claiming that ALAC vs. FLAC vs. even AIFF or WAV can, or even will, sound different is to claim that the compression format isn't lossless. That is exactly the claim you're making, that ALAC and/or FLAC are losing data. This is very, very easy to test for.

- Take a raw uncompressed WAV or AIFF.
- Encode it as FLAC or ALAC (doesn't matter which, but you can try both).
- Decode it back to an uncompressed WAV or AIFF (which ever you started with).
- Perform a diff on the original compared to the compressed/decompressed file.
- You will note that there are no differences marked.

Why? Because at the binary format level of the file, everything was preserved in the process of compressing and decompressing. Which is precisely what makes it lossless.

Anyway, these tests for binary preservation are very easy to perform, and they don't require golden-ears, esoteric audio equipment, or belief in the wafting hands of some sound-spirit. A very basic computer, a WAV or AIFF file, and a program that can encode/decode FLAC and/or ALAC, and a simple binary diff program (lots of free and opensource ones) will do the trick. Any Mac or modern Linux machine will come with all the necessary tools, and any Windows PC can have the requisite software setup in a few minutes.

Cheers!

Let me ask you a question, since you're convinced that they can sound different.

Take a folder full of Word documents. ZIP it. Heck, re-ZIP it several times. Extract all the files. Do your Word documents look different? Did the formatting change? What about the letters? Did new words get inserted, or some others deleted?

We are not talking about a Word file. It does not have to be converted in the same way to convey timing information as well as content through various electronic devices to an electromechanical device. Regardless of your reasoning, my ears tell me different. I cannot explain it beyond that. I am not an expert in such matters by any stretch of the imagination. My friend did a rip via EAC to WAV and converted that rip to ALAC in iTunes. We compared that to a rip of the same tune directly via iTunes to ALAC. I can tell you with high certainty that on my system I could identify the files blindly 10/10 times. My friend felt the same way on his (very resolving) system. Yet in theory they should be bit-for-bit identical files. I can upload those two files should you care to compare them yourself and see if you agree, or you could try the same experiment if you have EAC and iTunes. On my office system, which is far less resolving, I could not tell the difference at all in the two files. There's plenty of discussions on similar topics on this and other sites. Choose whatever you'd like to believe, and use whatever works for you. I get your reasoning, and on face value it looks good on the page, but in real life, to my ears, it doesn't work that way. Enjoy the music!

Timing information? You pull down the audio from a CD into a WAV, the timing information is encoded as samples in the WAV file. The exact same samples are encoded into the FLAC or ALAC.

Timing information discrepancies are going to come from how your hardware deals with the PCM data, well after the software encode/decode stages. The PCM data itself will be binary-identical.

Can you explain to me how two binary-identical PCM data sets can sound different? Unless your hardware is locking on to one signal at 44.1 khz and the other at something not-44.1 khz (48, 96, etc.) they will sound the same.

You can believe that you can tell the difference 10/10 times. Fundamentals of mathematics and computer science prove unequivocally that you cannot in reality do so, regardless of what you believe. This is evident by lossless video, lossless image, and lossless _insert_filetype_ methods elsewhere that repeatedly and accurately reproduce the source from the compressed format. Audio is not special in this regard. Audiophilia is however.

FLAC and ALAC are either lossless, or they are not. If they are lossless, then they will produce identical results. If they are not, then they will have differences. So, simply test for wether or not they are lossless (as I outlined before). You can perform this test empirically and objectively, quite easily disproving subjective differences as bias.

Naschbac,

What you say makes a lot of sense. As long as the software is "bit transparent" then there should be no difference. There is some info on Headfi discussion forums about this issue under the Benchmark DAC1 threads.

You can believe that you can tell the difference 10/10 times. Fundamentals of mathematics and computer science prove unequivocally that you cannot in reality do so

I believe that in reality I can do so. Mathematics or computer science may have nothing to do why, and perhaps I don't understand it. Yet one file sounds consistently better to me. He actually supplied me with three files initially, and I also had a WAV file I had ripped myself of the same song. I picked out the single file from the four as sounding better, and I had no idea of what he was giving me or why. He just asked that I listen to the three and that I might be surprised by what I heard. The other three files all sounded the same, or too close to make a distinction, but one kept standing out with greater clarity, separation and slightly better bass.

I will give a blind test a try next time I have someone over who's willing to spend the time to conduct it. Heck, I'll give the test to someone who has no clue as to what they're listening to and see what happens. Hey, I could be wrong, but I'm pretty sure I'll be able to pick it out each time. I care as much about mathematics proving me wrong as I would about a tube amp not measuring up well by conventional standards. Ultimately you have to listen to it, and if you like what you are hearing better on one over another, what does it matter? FTR I am not re-ripping my entire library to conform to the convention my friend used. It's actually a royal PITA and I would not consider it. The difference is not that profound, even though I can clearly hear it.

Ping me off the thread if you are interested in the files in question. I can post them later on and give you a link to download, listen and analyze the two files yourself. As I said, my system is pretty resolving and I'm quite sure I cannot tell any difference at all on my office system. So it will not surprise me at all if you cannot hear a difference on a less resolving system. I'd be curious to hear what you come up with. Maybe I need to up my medication!