Quote:
Originally Posted by tintin47 /img/forum/go_quote.gif
Then how do you explain that different songs ripped lossless have different bitrates? Also, how would you explain that lossless files take up only 50-60% of the space that the information would require were it directly recorded onto a harddrive?
|
As for your second question, they take up a fraction of the original space for the same reason as a zip or rar file takes up a fraction of the space occupied by the original files in it. I'm not going to get into the nitty gritty details, but at least in FLAC for sure, no information at all is discarded. If you want to know more, google it. Lossless data compression is nothing new, the lossless audio codecs are just compression schemes that are tailored for audio streams.
If you can't believe it's true, that's your problem. Zip up a spreadsheet and then unzip it. Your data is still there. The math is different but the theory is the same.
As for differing bitrates, for any given lossless compression system - whether it's compressing spreadsheets or music - the compression ratio necessarily dependent on how easy it is to find redundant bits in it and then describe them, and most compression systems can be configured variably for how hard to look.
zip for example has 9 levels of 'compression' that are really just 9 different settings for how wide a net it should cast. On higher levels, it spends more cycles on the search, and has a harder time describing it (with the result that zip -9 sometimes creates a slightly larger file than a lower setting might have). I don't have intimate familiarity with all the lossless formats, but the existence of this sort of functionality would come as no surprise.
If someone were to apply a bandpass filter to the audio before compressing it, as you suggest, this would reduce the size of the resulting file somewhat - but then the file could not truly be called lossless, and whoever created it would rightly be made the recipient of much scorn.
Some lossless codecs, like OptimFROG, offer a hybrid mode where some data is stored in a lossy format. FLAC and ALAC do not.
What i propose to you is to rip a short track from a CD as WAV, encode that WAV as FLAC, and then decode the FLAC back to another WAV file. This should be easy to do with mplayer/mencoder or other free tools. Then you can use any of innumerable free binary file comparison tools to determine what, if any differences exist between them. You will find that file header information differs between the wav files, but as long as both wavs are in the original S16LE format as used on the CD, the audio stream will be identical.