A question about FLAC

[Johnny Vinyl]

Whenever I rip a CD to my hard drive I save it as a FLAC file to save space. I noticed that FLAC has various levels of compression and I've used a few different settings over time. Not once did I hear a difference in playback, so why are these multiple compression levels there?

[hometheatergeek]

Originally Posted by Johnny Vinyl

Whenever I rip a CD to my hard drive I save it as a FLAC file to save space. I noticed that FLAC has various levels of compression and I've used a few different settings over time. Not once did I hear a difference in playback, so why are these multiple compression levels there?

From Wikipedia

Compression levels[edit]

libFLAC uses a compression level parameter that varies from 0 (fastest) to 8 (smallest). (The compressed files are always perfect "lossless" representations of the original data.) Although the compression process involves a tradeoff between speed and size, the decoding process is always quite fast, and not very dependent on the level of compression

IOW, there is no difference between the levels. If you are in a hurry you would use 0 and if you want to use as little space as possible you use the 8 setting. When we create images for computers we use the highest setting to save as much space since PC loads take up a lot of room.

[Johnny Vinyl]

Originally Posted by hometheatergeek

From Wikipedia



IOW, there is no difference between the levels. If you are in a hurry you would use 0 and if you want to use as little space as possible you use the 8 setting. When we create images for computers we use the highest setting to save as much space since PC loads take up a lot of room.

I figured that the various levels had to do with filesize only, but wasn't completely sure...hence the question.

[bhampton]

Awesome info.

I never knew there was levels as I use Apple Lossless only and always convert if I encounter some FLAC.

[pentatonic]

Yep, the way the algorithm works, if you want the smallest file size, it will only preserve the music. If speed is needed, and should not be if you do keep all your music, then it also encodes the "noise", the blank parts. Some do seem to believe that not having to recreate white noise, empty tracks etc by having it also encoded "could" be better to the limits of the actual player used as if much information is actually missing, some do believe different players may "possibly" be better at recreating those.

As for me, I've actually tried to tell through different players if I think there is any validity to such claims, and either I'm limited by my gear and ears (which I doubt) but I see no point.

[pjb3]

Originally Posted by pentatonic

Yep, the way the algorithm works, if you want the smallest file size, it will only preserve the music. If speed is needed, and should not be if you do keep all your music, then it also encodes the "noise", the blank parts. Some do seem to believe that not having to recreate white noise, empty tracks etc by having it also encoded "could" be better to the limits of the actual player used as if much information is actually missing, some do believe different players may "possibly" be better at recreating those.

As for me, I've actually tried to tell through different players if I think there is any validity to such claims, and either I'm limited by my gear and ears (which I doubt) but I see no point.

What you describe would be true for lossey codecs but not for FLAC. A decoded FLAC file is bit for bit the same as the original, nothing is omitted no matter the compression level.

[pentatonic]

Originally Posted by pjb3

What you describe would be true for lossey codecs but not for FLAC. A decoded FLAC file is bit for bit the same as the original, nothing is omitted no matter the compression level.

Yes, trust me I know that, but the idea of compression only works if you find methods of actually removing info that is not pertinent to the end result as well as "zipping" that file. The decoder will generate a file that is 100% accurate by using an algorithm to rebuild it, but trust me there is a reason to why your end result being encoded can vary so much in output file size, it's not just "zipping it" better with no difference, just that decoders that are to spec will very rapidly rebuild the audio.

[pjb3]

Originally Posted by pentatonic

Yes, trust me I know that, but the idea of compression only works if you find methods of actually removing info that is not pertinent to the end result as well as "zipping" that file. The decoder will generate a file that is 100% accurate by using an algorithm to rebuild it, but trust me there is a reason to why your end result being encoded can vary so much in output file size, it's not just "zipping it" better with no difference, just that decoders that are to spec will very rapidly rebuild the audio.

There is no such thing as "removing info that is not pertinent" in lossless compression, all data is considered relevant. Some data can be compressed at a greater rate than other data and some compression algorithms are better than others but at no time is data ever removed in lossless compression.

Johnny,
If you are interested in an experiment, rip the same track to two different FLAC files with different compression. You will have files with two different sizes. If you look in the metadata there should be a value for MD5 which is the checksum of the audio stream of the file. Both of the MD5 values should be the same indicating that both files will decode exactly the same. Another test, if you use Foobar2000, would be to add the Binary Comparator component and then select the two files and tell it to "Bit-Compare Tracks". This will disregard any non-audio data and compare only the audio within the files and no difference should be found.