• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Apple Lossless vs. WAV and other Formats

shacky

shacky

Addicted Member
I was reading a stereo magazine a few months ago (Absolute Sound I think) and there was an article talking about digital formats.

From what I recall it basically said Apple Lossless was "inferior" to I think WAV. Also said FLAC was inferior. Talked about soundstage and such.

For life of me I can't find the issue.

Is there really that much difference in formats? Should I consider re-ripping my CD collection for posterity? :scratch2:
 
Register to hide this ad
FWIW, I've run tests.

It works this way...what you put in to flac, you get out of flac...sample-to-sample. WAV headers from decoders MAY vary...but after compensating for it...every bit of sample data comes out the same...both flac and alac. I don't know if that guy was aware that digitql audio is dependant on sample data...if its the same...how is it different?

I've heard people claim there's a difference...so I ask them if zipping a text file amounts to a different text file.

Part of it is I think people get this physchlogical idea that any compressed audio is bad...and they carry this idea over to flac. Its compressed...its different somehow.

There may be logical explanations. Perhaps he's using different players\decoders that are doing something. You'd think they wouldn't...but I've had people claim the same...and it boiled down to they were using player a for wav, player b for lossless...and player b had DSP effects going.

I've ripped audio cds with cue/wav methods...compressed the wav to level 8 flac...then decoded using standard command line tools...burned the cd...then reripped it. The WAV files were bit-for-bit. Same hash results...bit comparison tools found 0 differences.

Going from both audible expierience and forensic-like analyzing...they're the same. WAV, FLAC, ALAC, APE, TTA (TrueAudio), they all produce the same quality.

Anyone who says different probably thinks coloring the edge of your cds black...or using a different interconnect for digital, changes the audio. Look...we're all nuts, I'm a tube guy...but I stand by my findings. Lossless formats do nothing to the audio stream to affect its quality.
 
This kind of talk comes from people who have a hard time grasping the concept of digital, or who try to compare digital as analog. These are 2 completely different technologies, and if someone doesn't know or understand it they should not comment on it. FLAC and ALAC or lossless formats - meaning that not one bit of digital information is lost from input to output. The file is smaller than WAV because the 1's and 0's are arranged more orderly. It looks like this (very simplified, of course):
WAV file: 1110000011111110000001010000111100000010000000111100000
There are 21 1's and 33 0's.
That one note in WAV looks like a seiers of 54 bits.
FLAC, and ALAC looks at it like this:31,50,61,60,1,1,1,40,41,60,1,70,41,50
That one note in FLAC or ALAC now looks like a series of 24 bits - a lot less space, and much better packageing of that note compared to WAV, yet the information is exactly the same as the source! WAV, FLAC, and ALAC will all sound identical identical to the source - bit perfect. But there is a caveat: the encoders/decoders MUST be working properly and be of high quality or else, yea, it could sound like crap (as well as any CD WAV playback). iTunes is not a good player for audiophile playback because it has it's DSP functions ON all the time by default - even if you choose OFF in the preference they are still ON. Just like an EQ where you can push the NO EQ button but leave the unit ON and noise is heard - same with iTunes. J. River, Foobar, and a few others will completely bypass the DSP functions allowing for bit perfect playback.

Now on mp3 lossy encodings things go south much faster - literally. mp3 analyzes the music signal and using an algorithym it decides how much information it believes most people will not be able to hear - KEY PHRASE: "it decides how much information it believes most people will not be able to hear". It will then throw away that information from the song. How 'bout that! The different rates of mp3 encoding will determine how aggressive the amount of information thrown out will be, but on average a 320k mp3 will have thrown away 40% of the original music's information. For some people this is fine as they may not have that good of hearing, or they're equipment may not be that resolving (portable mp3 players, cheap headphones, noisy inviroments, etc). But there still are some people who have traained there ears to notice these difference's. People like musicians, conductors, and of course people like us who love music and know what real music sounds like. Personally, I don't want ANY part of my music thrown out, whether I can hear it or not, I want it all there.

One last thing to remember when talking about which is better, WAV or FLAC/ALAC, WAV cannot hold any metadata information, but FLAC/ALAC can. Metadata is the information like artist name, song name, album cover art, etc that we NEED when we play music on a PC. WAV files were never designed to handle these things, whereas FLAC, ALAC, and mp3 can. You would have to manually enter each song, artist info, cover art, etc for each WAV file you have, and the worst part is that the PC cannot save it to the WAV file directly so it must save it in another location, thus if you copy the songs to another drive, player, or whatever that information will not be there and you'll have to re-enter it all over again!

FLAC/ALAC is really the only way to go when using the PC as a music playback system.

Jimmy
 
Last edited:
I stated something about tags last week.

WAV has no offical support for tags. Some programs I've got will write tags in the header. Goldwave for example invokes a LIST property. I'm not sure where its stored in the file.

Not many programs will find this LIST property. Its non-standard. I've seen apps write id3v2 to the end of a wav.

The only advantage of wav is that, by now, its become standard and supported across a number platforms and usually contains PCM data.

I should also note that WAV is based on RIFF and is actually a container format. One popular thing from the mid 90s was .wav files that contained layer-3 (mp3) data.

MP3 also contains most of its data in frequency domain...rather than time domain..which is what we're used to. Getting time-domain in to frequency domain, and vise versa, isn't that easy. In fact, its been said one of the best ways to downsample from a higher bit-depth is to throw the audio in to frequency domain, then convert it back to time domain. Avoids messy quantization...but has other issues.

That's also a reason why, technically, mp3 isn't 16-bit. Since it relies hevily on frequency domain (through the iDCT method), it technically doesn't have a bit-depth. You can feed it 24-bit PCM and, some encoders, will use that audio in 24-bit..decoding it as 16 or 24-bit.

I'll also add the physcoacoustic removal of audio is, in fact, VERY similar to modern noise reduction methods, removing frequencies according to a model.
 
dewdude said:
I stated something about tags last week.

WAV has no offical support for tags. Some programs I've got will write tags in the header. Goldwave for example invokes a LIST property. I'm not sure where its stored in the file.

Not many programs will find this LIST property. Its non-standard. I've seen apps write id3v2 to the end of a wav.

The only advantage of wav is that, by now, its become standard and supported across a number platforms and usually contains PCM data.

I should also note that WAV is based on RIFF and is actually a container format. One popular thing from the mid 90s was .wav files that contained layer-3 (mp3) data.

MP3 also contains most of its data in frequency domain...rather than time domain..which is what we're used to. Getting time-domain in to frequency domain, and vise versa, isn't that easy. In fact, its been said one of the best ways to downsample from a higher bit-depth is to throw the audio in to frequency domain, then convert it back to time domain. Avoids messy quantization...but has other issues.

That's also a reason why, technically, mp3 isn't 16-bit. Since it relies hevily on frequency domain (through the iDCT method), it technically doesn't have a bit-depth. You can feed it 24-bit PCM and, some encoders, will use that audio in 24-bit..decoding it as 16 or 24-bit.

I'll also add the physcoacoustic removal of audio is, in fact, VERY similar to modern noise reduction methods, removing frequencies according to a model.
Click to expand...

I use a mac (itunes) or Ipods with Wadia transports >>tube dac>>tube buffer>> amps.

Had great results with 16/44-24/48 apple lossless on the ipods (dac set to 24/96) or full strength .aiff files on the macs (16/44 to 24/192) all ripped from FLAC files. I'm a fussy listener and have been very happy with the results.

MP3 files (320/lossy, etc) suck for critical listening but are fine for background music or using in the car.

Just my 2 cents.
 
You must log in or register to reply here.
Back
Top Bottom