Re: [ARSCLIST] Lossy compression losing quality (was Re: [ARSCLIST] Pristine Audio and the Milllennials . . .)

[Mike Richter <mrichter@xxxxxxx>]

Howard Friedman wrote:

> The loss is essentially in the high frequencies, but not uniformly so. 
> Note that compression may be selected in a range of about 3-30 with the 
> usual choice 11:1 (128 Kbps MP3 from redbook WAV). The more compression, 
> the greater the loss.
>
> Note, too, that LAME offers a wider range in its not-MP3 MP3 than do 
> Fraunhofer compressors.
>
> Mike


-- Windows Media Player provides 6 possible formats for ripping audio CD 
tracks. The Windows Media Audio (WMA) is stated to be lossless, yet 
these files are about one-third as large as the corresponding WAV files. 
Are they equal in quality to WAV? In other words, for the BEST 
conversion, which format should one use to RIP audio tracks from a CD to 
a hard drive? Also, there is a slight, usually 1 second, difference in 
timing of these various formats per track. Howard

************
Sorry about the indenting problem; it occurs because you did not edit 
out my signature.

I do not use WMP for conversion, but suspect that much the same 
compression would be achieved if you used Monkeys Audio or FLAC on the 
same file. The factor of two applies with a typical source with 
representative tape or LP noise. A monaural source captured in two 
channels, a low-noise source, or one with very little high-frequency 
content for the sample rate will compress more than that, as will spoken 
material which has lots of silence.

There is one meaningful test for a lossless compressor. Compress the 
source file, decompress that, and compare the results. If the 
decompressed file matches the original, the compressor is lossless. 
Different compressors - indeed, different settings on a single 
compressor - achieve different amounts of compression and take different 
times for processing. In addition, the standard value is usually the 
most consistent from version to version and the best behaved. (Less than 
standard compression has little payoff; more compression tends to save 
little more than standard.)

As usual, there is no definition of "best" except your own.

Any form of compression can lead to differences in the estimated and 
actual playing time of the decompressed file. There have been some 
players that were thrown by that and would cut off playback at the 
estimated duration (from the header) rather than when the file ran out 
of information. That problem seems to have been eliminated from WinAmp 
and other players that have lasted more than a year or two. However, the 
displayed time may differ from the actual time with any compressed format.

Mike
--
mrichter@xxxxxxx
http://www.mrichter.com/