Is Apple Lossless really lossless? Here are my observations.

First my setup: Apple dual-processor G5 running iTunes 6.0.4 under OS
10.3.9 connected via Toslink optical cable to a Rotel RSDX-02
integrated amp driving B&W 805 speakers in an 11X14 foot furnished
room.

Experiment one - I imported a test track (Copland's "Fanfare for the
Common Man" from
Telarc CD-80339 [DDD 1997]) in two ways - first using Apple Lossless,
then again as a straight .wav file. (Both with "error correction"
turned on.) This piece begins with tympany and symbols played forte,
followed by solo trumpet played mezzo forte then forte. The loud
dynamics, bright brass, and deep bass make it an excellent test CD, one
which has sent many pairs of speakers down in flames at my local audio
dealer.

I then burned two audio CDs from iTunes; the first from the AL version
of the file, the second from the .wav version. I then compared the
original CD and the two burned discs by playing them on the Rotel amp's
built-in CD player, listening carefully at loud and soft volumes. Here
are my observations:

The CD iTunes ripped from the .wav file sounded identical to the
original CD at all volumes.

The CD iTunes ripped from the AL file was clearly not as loud as the
original CD; when I turned up the volume on the Rotel to compensate,
there seemed to be a little loss of bass depth at the lower
frequencies, and the treble seemed to have less definition at the
brightest passages.


Experiment two - I played the two files from iTunes through my Mac's
optical digital output to the Rotel amp over the toslink cable. Again
these two files sounded slightly different from the original CD, but
exactly like eachother. Interestingly, both files played over the
optical cable from the computer to the amp were indistinguishable from
the CD burned from the apple lossless version of the file when played
on the Rotel amp's built-in CD player.

My Conclusions - It appears that iTunes performs some processing on
the files when they are converted to Apple Lossless, and this same
processing is applied when the files are played from iTunes regardless
of whether the source file is stored on the hard drive in AL or .wav
format.

CDs ripped from .wav files from iTunes sound identical to the
originals. CDs ripped from apple lossless files do not.

My Question - What kind of processing is iTunes (or more precisely
the AL codec) performing? My guess that it is at least some form of
dynamic compression, but I am not sure. What do you think?

(If you would like to confirm my observations, I would encourage you to
perform the experiments for yourself, using your favorite high-fidelity
setup and test CD.)

In article ,
"feedback" wrote:

Is Apple Lossless really lossless? Here are my observations.

First my setup: Apple dual-processor G5 running iTunes 6.0.4 under OS
10.3.9 connected via Toslink optical cable to a Rotel RSDX-02
integrated amp driving B&W 805 speakers in an 11X14 foot furnished
room.

Experiment one - I imported a test track (Copland's "Fanfare for the
Common Man" from
Telarc CD-80339 [DDD 1997]) in two ways - first using Apple Lossless,
then again as a straight .wav file. (Both with "error correction"
turned on.) This piece begins with tympany and symbols

Actually, there are no cymbals; what you are hearing is a tam-tam, just
FYI.

played forte,
followed by solo trumpet played mezzo forte then forte. The loud
dynamics, bright brass, and deep bass make it an excellent test CD, one
which has sent many pairs of speakers down in flames at my local audio
dealer.

I then burned two audio CDs from iTunes; the first from the AL version
of the file, the second from the .wav version. I then compared the
original CD and the two burned discs by playing them on the Rotel amp's
built-in CD player, listening carefully at loud and soft volumes. Here
are my observations:

The CD iTunes ripped from the .wav file sounded identical to the
original CD at all volumes.

The CD iTunes ripped from the AL file was clearly not as loud as the
original CD; when I turned up the volume on the Rotel to compensate,
there seemed to be a little loss of bass depth at the lower
frequencies, and the treble seemed to have less definition at the
brightest passages.

Ah! Sighted testing! Tsk! ;_)


Experiment two - I played the two files from iTunes through my Mac's
optical digital output to the Rotel amp over the toslink cable. Again
these two files sounded slightly different from the original CD, but
exactly like eachother. Interestingly, both files played over the
optical cable from the computer to the amp were indistinguishable from
the CD burned from the apple lossless version of the file when played
on the Rotel amp's built-in CD player.

My Conclusions - It appears that iTunes performs some processing on
the files when they are converted to Apple Lossless, and this same
processing is applied when the files are played from iTunes regardless
of whether the source file is stored on the hard drive in AL or .wav
format.

CDs ripped from .wav files from iTunes sound identical to the
originals. CDs ripped from apple lossless files do not.

My Question - What kind of processing is iTunes (or more precisely
the AL codec) performing? My guess that it is at least some form of
dynamic compression, but I am not sure. What do you think?

(If you would like to confirm my observations, I would encourage you to
perform the experiments for yourself, using your favorite high-fidelity
setup and test CD.)

Seriously, thanks for this, Feedback. Worth looking into for sure.

"feedback" wrote in message
...
Is Apple Lossless really lossless? Here are my observations.

First my setup: Apple dual-processor G5 running iTunes 6.0.4 under OS
10.3.9 connected via Toslink optical cable to a Rotel RSDX-02
integrated amp driving B&W 805 speakers in an 11X14 foot furnished
room.


There's actually a really very simple test you can run to verify the
"losslessness" of any lossless codec. First, take the source CD and rip a
wav file to the pc using EAC. This will result in a duplicate of the data
on the CD. Second, take your favorite lossless codec, (FLAC, Monkey's
Audio, Apple, Microsoft, whatever) and compress the file. Third, decompress
the file to a wav file using the codec's decompress. Lastly, run a file
compare on the before and after wav files. If they're identical, they're
identical.

Bruce

On 5 Mar 2006 18:04:10 GMT, "feedback" wrote:

Is Apple Lossless really lossless? Here are my observations.

First my setup: Apple dual-processor G5 running iTunes 6.0.4 under OS
10.3.9 connected via Toslink optical cable to a Rotel RSDX-02
integrated amp driving B&W 805 speakers in an 11X14 foot furnished
room.

Experiment one - I imported a test track (Copland's "Fanfare for the
Common Man" from
Telarc CD-80339 [DDD 1997]) in two ways - first using Apple Lossless,
then again as a straight .wav file. (Both with "error correction"
turned on.) This piece begins with tympany and symbols played forte,
followed by solo trumpet played mezzo forte then forte. The loud
dynamics, bright brass, and deep bass make it an excellent test CD, one
which has sent many pairs of speakers down in flames at my local audio
dealer.

I then burned two audio CDs from iTunes; the first from the AL version
of the file, the second from the .wav version. I then compared the
original CD and the two burned discs by playing them on the Rotel amp's
built-in CD player, listening carefully at loud and soft volumes. Here
are my observations:

The CD iTunes ripped from the .wav file sounded identical to the
original CD at all volumes.

The CD iTunes ripped from the AL file was clearly not as loud as the
original CD; when I turned up the volume on the Rotel to compensate,
there seemed to be a little loss of bass depth at the lower
frequencies, and the treble seemed to have less definition at the
brightest passages.


Experiment two - I played the two files from iTunes through my Mac's
optical digital output to the Rotel amp over the toslink cable. Again
these two files sounded slightly different from the original CD, but
exactly like eachother. Interestingly, both files played over the
optical cable from the computer to the amp were indistinguishable from
the CD burned from the apple lossless version of the file when played
on the Rotel amp's built-in CD player.

My Conclusions - It appears that iTunes performs some processing on
the files when they are converted to Apple Lossless, and this same
processing is applied when the files are played from iTunes regardless
of whether the source file is stored on the hard drive in AL or .wav
format.

CDs ripped from .wav files from iTunes sound identical to the
originals. CDs ripped from apple lossless files do not.

My Question - What kind of processing is iTunes (or more precisely
the AL codec) performing? My guess that it is at least some form of
dynamic compression, but I am not sure. What do you think?

(If you would like to confirm my observations, I would encourage you to
perform the experiments for yourself, using your favorite high-fidelity
setup and test CD.)

The above is utterly irrelevant to the thread title - even though it
is the originating post.

These are *digital* data files, so all you have to do is run a digital
comparison to discover if the system is lossless.

Bits is bits, dude....

--

Stewart Pinkerton | Music is Art - Audio is Engineering

Jenn wrote:
In article ,
"feedback" wrote:

Is Apple Lossless really lossless? Here are my observations.

First my setup: Apple dual-processor G5 running iTunes 6.0.4 under OS
10.3.9 connected via Toslink optical cable to a Rotel RSDX-02
integrated amp driving B&W 805 speakers in an 11X14 foot furnished
room.

Experiment one - I imported a test track (Copland's "Fanfare for the
Common Man" from
Telarc CD-80339 [DDD 1997]) in two ways - first using Apple Lossless,
then again as a straight .wav file. (Both with "error correction"
turned on.) This piece begins with tympany and symbols

Actually, there are no cymbals; what you are hearing is a tam-tam, just
FYI.

played forte,
followed by solo trumpet played mezzo forte then forte. The loud
dynamics, bright brass, and deep bass make it an excellent test CD, one
which has sent many pairs of speakers down in flames at my local audio
dealer.

I then burned two audio CDs from iTunes; the first from the AL version
of the file, the second from the .wav version. I then compared the
original CD and the two burned discs by playing them on the Rotel amp's
built-in CD player, listening carefully at loud and soft volumes. Here
are my observations:

The CD iTunes ripped from the .wav file sounded identical to the
original CD at all volumes.

The CD iTunes ripped from the AL file was clearly not as loud as the
original CD; when I turned up the volume on the Rotel to compensate,
there seemed to be a little loss of bass depth at the lower
frequencies, and the treble seemed to have less definition at the
brightest passages.

Ah! Sighted testing! Tsk! ;_)

I see that Jenn has learned something from the more technically inclined
from this newsgroup after all .

Bruce Adams has described a method to check the losslessness of the
codec in another post. Basically, you can do a file compare of the two
..wav files: the one extracted from the original CD, and the one
extracted from an iTunes burned CD that has gone though a
coding/decoding operation through the Apple lossless codec (original CD
to .wav to .m4a to burned CD to .wav). I recommend using a utility
called "Beyond Compare 2" for file comparisons; there is a free 30 day
trial. For ripping from CD to .wav, I use CDEx, but EAC works well, too.

The last time I did this, I found that on one CD track Apple lossless
compresses the .wav file to about 60% of its original size. Oh, by the
way, there is no difference in the two .wav files. Aren't sighted tests
wonderful?


Experiment two - I played the two files from iTunes through my Mac's
optical digital output to the Rotel amp over the toslink cable. Again
these two files sounded slightly different from the original CD, but
exactly like eachother. Interestingly, both files played over the
optical cable from the computer to the amp were indistinguishable from
the CD burned from the apple lossless version of the file when played
on the Rotel amp's built-in CD player.

My Conclusions - It appears that iTunes performs some processing on
the files when they are converted to Apple Lossless, and this same
processing is applied when the files are played from iTunes regardless
of whether the source file is stored on the hard drive in AL or .wav
format.

CDs ripped from .wav files from iTunes sound identical to the
originals. CDs ripped from apple lossless files do not.

My Question - What kind of processing is iTunes (or more precisely
the AL codec) performing? My guess that it is at least some form of
dynamic compression, but I am not sure. What do you think?

(If you would like to confirm my observations, I would encourage you to
perform the experiments for yourself, using your favorite high-fidelity
setup and test CD.)

Seriously, thanks for this, Feedback. Worth looking into for sure.

Now here's the interesting question. If you know that the files are
identical, can you still detect audible differences?

Bruce Abrams wrote:
"feedback" wrote in message
...
Is Apple Lossless really lossless? Here are my observations.

There's actually a really very simple test you can run to verify the
"losslessness" of any lossless codec...file compare.

I've done this test withy Apple Lossless. "Before" and "after"
AIF files are bit-for-bit identical, showing that ALC doesn't
eliminate data.

John Atkinson
Editor, Stereophile

In article , Chung
wrote:

Jenn wrote:
In article ,
"feedback" wrote:

Is Apple Lossless really lossless? Here are my observations.

First my setup: Apple dual-processor G5 running iTunes 6.0.4 under OS
10.3.9 connected via Toslink optical cable to a Rotel RSDX-02
integrated amp driving B&W 805 speakers in an 11X14 foot furnished
room.

Experiment one - I imported a test track (Copland's "Fanfare for the
Common Man" from
Telarc CD-80339 [DDD 1997]) in two ways - first using Apple Lossless,
then again as a straight .wav file. (Both with "error correction"
turned on.) This piece begins with tympany and symbols

Actually, there are no cymbals; what you are hearing is a tam-tam, just
FYI.

played forte,
followed by solo trumpet played mezzo forte then forte. The loud
dynamics, bright brass, and deep bass make it an excellent test CD, one
which has sent many pairs of speakers down in flames at my local audio
dealer.

I then burned two audio CDs from iTunes; the first from the AL version
of the file, the second from the .wav version. I then compared the
original CD and the two burned discs by playing them on the Rotel amp's
built-in CD player, listening carefully at loud and soft volumes. Here
are my observations:

The CD iTunes ripped from the .wav file sounded identical to the
original CD at all volumes.

The CD iTunes ripped from the AL file was clearly not as loud as the
original CD; when I turned up the volume on the Rotel to compensate,
there seemed to be a little loss of bass depth at the lower
frequencies, and the treble seemed to have less definition at the
brightest passages.

Ah! Sighted testing! Tsk! ;_)

I see that Jenn has learned something from the more technically inclined
from this newsgroup after all .

LOL Actually, I've ALWAYS seen the advantage of blind testing. What
I've questioned is the practical application of it in the shopping
process.