Please choose your favorite lossless audio format.

-Ed

There's no subjectivity in lossless, which makes this comparison kind of moot IMO. Whichever provides the smallest filesize is best.

The way I encode my MP3s-- VBR is the key-- I doubt you could hear the difference in a double-blind headphone test.

Wrong.

There's also the following to consider:
1) Ease of encoding
2) Ease of tagging
3) Playback support (which softwares can play it back)
4) Speed of encoding

So it's not just about size.

I don't mean to sound harsh; I just felt that this was sort of obvious. I'm not stupid enough to compare two things when there's nothing to compare. There is a reason for comparison here.

-Ed

EDIT: PS: BTW (again, no offense intended), in order to get a noticeable difference between 100% VBR MP3 and lossless formats, you need, both, good, "ears," (as audiophiles call it), as well as sufficiently good equipment. Again no offense, as I am not saying your hearing is no good or your equipment is no good (I've no idea how good/bad your hearing or equipment is, and thus am in no position to judge), but back when I was using a set of Creative Labs FPS2000's hooked up to My SB Live! via Digital-DIN, there wasn't any tangible difference between WAVe audio and a 100%VBR MP3. With my current setup (Echo MiaMIDI, S/PDIF to Onkyo TX-SR501, driving Paradigm Reference Studio/20s via silver cabling), however, the difference is night and day. If you're anywhere near NY, I invite you to come over and hear for yourself.

"Favorite Lossless Audio Format"?!?

That's edging a bit into the esoteric, isn't it?

Keep this up and I'll figure out a really obscure poll to see if I can get no votes at all

Maybe something like "Favorite 2D Image Convolution":
Sobel -vs- Difference Of Gaussians -vs- FFT -vs- Walsh-Hadamard Transform -vs- Binary Moments

Or perhaps "Favorite Capitalized Consonant Letter With Curved Lines":
B, C, D, G, J, P, Q, R, or S?

Of course I could be completely off base and you could get dozens of votes. I'm wrong about things all the time.


Scott

Well, I own $500 worth of Sennheiser headphones (two sets), so let the equipment pissing match begin!

Lossless is kind of a waste, unless you MUST have archival copies (in which case I, shock and horror, buy the original CD). LAME VBR at a high enough bitrate is really indistinguishable from the original-- and I mean "double-blind headphone audiophile listener" indistiguishable. In other words, you can randomly pick the raw WAV and the high quality MP3, and the listener cannot tell which is which.

Like so:

It happens to be archival I seek.

It also happens that I wish to avoid flip-flopping discs every time I go to listen.

It might just be the fact that I was encoding with MusicMatch instead of LAME; either way, I'll stick to what's, for a fact, precise reproduction. I've the space to spare.

-Ed

Oh man, you really REALLY need to be using LAME. It is hands down the best MP3 encoding engine out there. Can't say enough good things about LAME.

For me, having the original CD (or archival copy I guess) gives me the luxury of using smaller encoding formats so I can fit more music on my portable devices.

The best "bang for the byte" for me personally is VBR MP3s on a moderately aggressive setting; I usually end up with files around ~160kbps average but it depends heavily on the content. VBRs take a lot longer to encode (like 4-5x longer) but they do a dramatically better job at capturing the dynamic range of the music-- eg, they use more bits to capture the hard stuff, while wasting less on the silence/easy stuff. Sounds much more "alive".

And don't even bother with VBR unless you're using LAME!

Actually, audiophile-ears aren't particularly helpful in distinguishing lossy audio encodes from the original. And neither is equipment, for a large percentage of artifacts you'll find in lossy encodes.

Most importantly is training your ears to hear the artifacts particular to lossy encoding. I can't distinguish high-end audio equipment for beans, but I'm good at picking out the various artifacts in lossy audio because I've taken many listening tests. Also, as long as the listening equipment isn't absolute junk, you can hear just about every artifact imagineable, except perhaps for high-frequency errors, if your equipment can't reproduce it.

Dibrom, the guy who tuned the Lame --alt-presets, did most of his work with some cheap headphones and the speaker out of his laptop! I've done listening tests with crappy headphones plugged into the crappy headphone-out on my friend's crappy speakers (so much hiss!), and I was just as capable of distinguishing lossy from lossless.


This is all very true. It's nitpicking to all but the most hardcore music collectors, though. I guess I belong in that group, too.

I picked FLAC, due to the mysterious fifth "thing to consider": speed of decoding. My other consideration was cross-platform support, and FLAC delivers in that respect, too.

Hey, sbabb?

I like:

S

I agree, there are many reasons one could have for choosing a "favorite" format.

Personally, I nowadays just use EAC to pull best-quality disk images directly to .wav/.cue. I prefer the simplicity/compatibility, have no problem with space, and never want to worry about it any further than that.

i also use exact audio copy (eac) and encode the files to .mp3 using the lame encoder... i have no complains but i dont really notice the difference between good and real good when talking about audio quality

I wish I had your drive space. You DO realize that lossless means that you can get the exact wave format out as you did going in. I mean, use either flac or mac, and it'll save you some space, probably around 50%

As for me, I voted for Monkey's Audio, but truthfully, I think I should investigate OptimFrog and LosslessAudio, which have the best compression ratios, but the worst encoding/decoding speed.

Anyone got benchmark results, or filesize results comparing them all?

Not mine, but I trust the results.
http://web.inter.nl.net/users/hvdh/lossless/lossless.htm

http://home.wanadoo.nl/~w.speek/comparison.htm

Looks like a relatively tight cluster around 55-60 percent compression for all of the ~12 lossless contenders. Considering a raw WAV file for a single song would be around (off the top of my head) 40 megabytes per 4 minute song, that's ~23mb per song. A high quality VBR version of the same file (~160kbps avg) would be 4.5mb.

Worst of all, the sound-specific compression algorithms exibit the same asymptotic response of general purpose compression algorithms (wizip, rar, etc)-- beyond a certain "sweet spot", you can expend hours of compression time for almost no improvement in size whatsoever. No magic bullet here..

Too rich for my blood! I'd rather keep the lossless versions offline on optical media, and the high quality compressed versions handy for use on a variety of devices.

If you feel daring and/or want to go completely nuts, you could try WavPack Hybrid or OptimFrog DualStream. What they can do is produce a lossy file and then produce a 2nd file of corrections. The lossless stream can be constructed from these two files. The idea behind this is so one could keep the lossy file on their computer and back up the correction file to CD.

The only disadvantage I know of is that it's not psycho-acoustically tuned (I think that's the right phrase). Meaning that it does not necessarily throw away the parts that you don't hear.

That's a nice comparison! Thanks for sharing the link.

From the graphs, you can see why FLAC is the most popular lossless format for portable devices--It decodes very quickly.

The most recent discussion of these "lossy-lossless" encoders at Hydrogenaudio is this thread, where several listeners compare Wavpack's new "hybrid" encoding with OptimFrog's and give feedback to David Bryant, the Wavpack developer. To summarize, bitrates are about twice as high as a good lossy encoder, but less than half of a completely-lossless encoder. Because these hybrid encoders don't have tuned psychoacoustic models (yet...), and because they encode music in a different manner from most lossy codecs, they have different distortions and artifacts. The most common problem with hybrid encodes is the addition of faint background noise, which can be heard by sensitive listeners. Occasionally, weird warbles and other problems show up. Extensive testing would likely reveal other problems as well. Hybrid encoders are still in heavy development, and plenty of work to be done before these hybrid encoders can compete with existing encoders for everyday encoding.

There are a couple of practical applications for hybrid encoders, though. One, of course, is the smaller "difference" files, which when combined with the hybrid file, recreate the original lossless file. Some people like the idea of archiving difference files to CD, rather than a whole lossless audio file. Another advantage for hybrid files is their transcoding friendliness: these hybrid encodes are reportedly good sources for transcoding to MP3, AAC, Vorbis, etc., much moreso than an already lossily-encoded file, and almost as good as a lossless file that's 2-3x larger. Basically, transcoding from one lossy file to another generates many more artifacts than simply encoding from the original lossless file. So, if you have a sizeable hard drive, but not one that's big enough for lossless archival of your entire collection, and you do a lot of portable-player listening, then hybrid encoders could be of some interest to you.

i hear that aac is supposed to be the next big thing in lossy audio, from what ive read its the next generation MP4 to replace MP3's. Reading audiophile website aac ripped at 225kbps is their standard and compaires to 360kbps mp3. For excellent quality vs size.

I want to get into high quality audio but am struggling trying to find good quality aac audio.

Seal, you're in luck! The latest round of 128kbps AAC audio tests has just completed at Hydrogenaudio. Check the news headlines on the front page for more information.

You know, I am the guy that attempted the transcode test several months ago. Although I haven't received enough data to put out results, I can say that this small advantage appears to be BS.

Oh, didn't see your transcoding test (I was on HA hiatus at the time). I must have been thinking about Den's initial findings last April. I'll go read your thread now, to see what you had to say. Is this thread the one I should read?

APE for lossless encoding of my CDs. I also keep everything in wmv format for use on my portable player. Eventually I'll have all my games and DVDs on harddrive.

I'm lazy and I don't like changing CDs or Movies. Its worth it, to me, to get additional hds to store everything.

That's the discussion thread. I opened a listening test thread after that here.