I thought bitrate did not matter with FLAC but Im comparing a cd I encoded to FLAC to one a buddy downloaded and there is a huge difference so Im wondering why. For example a song I encoded using React2 ended up being 34.6MB and 964kbps bitrate. the same song that was downloaded is 10MB and 298kbps. The whole album is like that. Actually, Ive seen other albums like that as well. So what gives here?

The download may have been lossyflac

Damn, I can see this coming: some people will upload lossyFLAC rips to BitTorrent sites without mentionning their lossy nature, they'll get caught ("zomg transcode!!1!"), all lossyWAV variations will be banned, and the codec will get a bad reputation for no good reason.

Or downloading FLAC files from BitTorrent sites will get a bad reputation

Couldn't it be due to volume differences?

Older release with more quite songs - smaller files
Newer release with louder songs - larger files

I'm not sure if this can cause such huge differences in bitrate, but if the difference is small, you might want to run ReplayGain on them and check their volumes.

It is clear the intention, if this was lossywav*, was to get the smallest file possible. This bitrate probably was obtained with -q 0 or -q 1 at most (default is still -q 5, being around 400~500kbps).

With this, i mean that the correct usage of lossywav should not get bad reputation in any sense. If someone wants to get the smallest size and not tell anyone it has used lossywav, then the file sizes would tell so anyway.


It will end (if continued) being like it was with mp3 back then. Everyone trying to find the highest bitrates (Which then put VBR in a second place while it shouldn't). Nothing to really worry about.



*which, btw, could not be by encoding a lossy mp3 or the likes.

Wishful thinking. I'd wager piracy has significant impact on the way software without corporate backing, like FLAC, is perceived. The iPod and the iPhone overwhelmingly dominate the market of portable media players, yet FLAC, which is not supported on those devices, is the de-facto codec for lossless music on P2P. Why? I have no idea, but it's a fact. And when someone mentions other lossless codecs such as WavPack, the usual answer is "wtf is that crap, get off my lawn".

Matroska, also without huge corporate backing, was headed to oblivion, when pirates started ripping HDTV broadcasts and chose Matroska for distribution, rejecting corporate-backed standards. Look how popular it is now! Hardware appliances have even started to support it, like DivX started getting support, for no other reason than its popularity due to pirated content.

I bet we'll start seeing posts where the OP dismisses suggestions to use lossyWAV on the grounds that they want "quality, not some half-assed transcodes".

That's assuming the P2P world acts rationally (which it doesn't by any stretch).


Sure, then again we still get threads asking "whether this file I downloaded is a transcode or not".


I hope you're right. LossyWAV is a fine piece of software that serves its purpose well, and deserves to be known.

I hadnt heard of lossyflac before but that does make sense after reading up on it. Listening and comparing I can tell no difference between the files (at least through my office stereo). One is just smaller than the other.

So is lossy flac out there to replace vbr mp3? I wouldnt think any lossy format would be an archive solution.

I was personally against the idea of Lossy FLAC even being allowed to exist (I don't recall if I ever posted on this). Reason being is that, how do you know if your FLAC files are actually lossless, unless you made them yourself, and/or have the original files to compare.

If there was some kind of signature or tag set in the files to distinguish them, that would be helpful. There may be, and I may have just missed reading about it. I need to search the older threads on the topic before posting further

There isn't. And in the end tags could be left out or faked. At first the file is a .wav, after that it becomes a lossy.wav; with perhaps a minimal quality and/or bitrate. From that point on it's a source just like any other wav and one can convert it to flac, wv, tak or whatever.
And I even have 'flac' files (self)made from mp3's. Flac is losslessly converting the source, so if the source sounds good enough, the flac would'n sound worse ?
Yeah, cheaters can be a problem, always.

The wave file that lossywav generates contains a tag with information about the version and commandline used to generate it. The "problem" is that FLAC doesn't add it by default (i.e. without the use of the --keep-foreign-metadata)

Also, it is relatively easy to detect if data has been processed (it is like the file had variable bitwidth)

yes, I don't know if such a detector exists, but all it has to do is scan the 'wasted bits' number in each frame. unprocessed flacs will show little to no variance in the sequence; lossyWAV-processed ones will vary.