Does this thing play WMA-Lossless *as* lossless?

April 26, 2013 4:28 PM

Hi, I have a new Cowon iAUDIO 9 portable. The lossless files I rip from Windows Media Player end up the same size on the device after synced. However, the device doesn't have a way to check the bitrate of any song. But it's acknowledged (not yet directly to me by the slow-to-reply makers in South Korea) that the device does NOT support WMA-Lossless, only everybody's precious FLAC. It will PLAY WMA, but usually on players like this the bitrate is 320.

Well, I'm starting to convert to FLAC with Foobar after a few days of waiting, including tagging w/ Mp3tag. A long process for my 3,000 ripped CDs...

So, am I wasting my time? In other words, if the size is the same, can the bitrate nevertheless sync as reduced? If not reduced, then maybe someone needs to tell the buyers of this discontinued item that it's actually playing losslessly without needing to convert to FLAC. Thanks.

posted by noelpratt2nd to Technology (17 answers total)

If the file is the same size, the audio data is probably just being copied to your player byte-for-byte. Lossy is smaller than lossless, after all.

But it looks like WMA lossless hasn't been publicly documented, so there seems to be no way to know if, say for example, WMA lossless consists of a lossy-encoded piece and a losslessly-compressed version of the difference, which is then recombined to create the original waveform (at least one lossless format does this, I don't remember which). If that were the case, your player might just play the lossy part. This is all just speculation, though.

Saying that it doesn't support WMA lossless doesn't necessarily mean that it can't do it, just that they don't want to have to field questions about it. But sometimes no support actually means that the device won't do it.

If it's any consolation, you really can't tell the difference between properly-encoded lossy and lossless anyway (that's the point), not even with the world's best mp3 player and world's best headphones. This is a fact.
posted by grahamsletter at 4:48 PM on April 26, 2013

I'm not sure what would mean for a player to play a losslessly-encoded file at a "bitrate" of 320 kbps (unless it were a "hybrid" format like grahamsletter suggested, but this page says it is not).

Are people suggesting that it decodes the WMA lossless file, and instead of just playing the decoded audio it then does a bunch of useless extra work to re-encode it at 320 kbps, then decode it again, and then play it? That seems very unlikely to me, since it would be pointless and a huge drain on the battery. (My own experience is this area is as a developer on software like Audacity that handles decoding, processing, playing, and encoding audio in many formats.)

If you put a lossless file on the player and it plays, then that's almost certainly the best quality you are going to get. I think it's highly unlikely there would be any difference in output after converting from one supported lossless format to a different supported lossless format. Anyone claiming they noticed such a difference would need some actual evidence to convince me, like a dump of the software from the device, or signal analysis of the audio output.
posted by mbrubeck at 4:56 PM on April 26, 2013 [2 favorites]

Response by poster: You guys are great -- trippin me out. I'll await more input.

Muchly,
N.
posted by noelpratt2nd at 5:05 PM on April 26, 2013

I'm going to agree with mbrubeck here. If these WMA Lossless files are transferring to your player in a "speedy" fashion, then they have to be playing as WMA Lossless files. The player itself is not going to have a decoder built in, so the only way I could see these being decoded to 320kbps WMA files is if it takes longer to "transfer" these files than you would expect (i.e., Cowon has software that decodes then transfers).
posted by kuanes at 6:13 PM on April 26, 2013 [1 favorite]

I agree, it seems like a lot of work to purposely not play a lossless file as such. But the possibility exists that the file format is interleaved, like JPGs that download with progressively better quality, and that the player can only see as "deep" into the file as the 320kbps "layer".

Another possibility is that 320kbps is just the highest number it knows how to display. When I play my precious FLAC files, Winamp shows a weird bitrate that doesn't seem right as well.

I wouldn't worry about it. Playback isn't nearly as hard as encoding, and the less compressed a file is, the easier it is to play back. I'm sure it is doing the best it can. And if it can't, there really isn't anything you can do about it.
posted by gjc at 6:30 PM on April 26, 2013

Response by poster: Except to go for placebo of continuing to "be sure" by converting all to FLAC, which it says it supports in no uncertain terms.

I mean, SanDisk (Sansa) converts lossless files before syncing them to 320, so they spend the time/energy. But you're right, this Cowon is speedy in that it just goes right to "syncing." I also am considering now that perhaps "they all say" yes-yes, we support WMA, but the manual and some reviews say it supports WMA "~320." Meaning, I suppose, that anything up to 320. If that at least was a VBR in Sansa, I might have not spent the bucks for the Cowon. But hey, it's also gapless and sounds very good.
posted by noelpratt2nd at 6:51 PM on April 26, 2013

Seems to me that setting up a randomized blind listening test to compare a few of the tracks in question with their FLAC-converted counterparts would involve far less work than converting your entire collection and might well remove your need to "be sure".
posted by flabdablet at 8:04 PM on April 26, 2013

Response by poster: Oh yeah, but the ear is almost always very tricky (and the brain). This is, I know, a matter of splitting hairs. But for some people and for some times, it takes us over a bit. I have surely done blind tests in by which a 128-bitrate was deemed to be the lossless music over an actual WAV file. Busted! Everyone goes through that, but then also overall I've come to hear what was missing before a greater percentage of the time listening to higher-quality files.
posted by noelpratt2nd at 8:19 PM on April 26, 2013

Best answer: This advice, by the way, is coming from a point of view that says that point of lossless compression is to prevent stacked compression artifacts from creeping in after multiple successive conversions.

As long as your main collection is losslessly encoded and properly backed up so that you're never going to find yourself needing to recover pieces of it from your player devices, there is really no point at all in worrying about whether what's on those devices is also losslessly encoded.

If the only trips your audio ever takes are one-way from a losslessly encoded archive out to a player device, then your workflow generates no opportunities for compression artifact stacking; you'll either be hearing genuinely losslessly encoded audio or at worst audio that's been lossily encoded once.

I don't believe there is any audible difference between 320kb/s lossily compressed audio and its lossless source. You should run enough blind A/B tests to inform your own opinion. Make sure they really are blind, though - the history of high quality audio is littered with expensive placebos.
posted by flabdablet at 8:23 PM on April 26, 2013 [1 favorite]

128kb/s is of course a different kettle of fish. It's usually not very hard to pick the difference between a 128kb/s MP3 and its losslessly encoded source. My personal opinion is that 128kb/s encoding should only ever be used for commercial pap whose lack of musical quality completely masks any lack of audio quality.
posted by flabdablet at 8:27 PM on April 26, 2013

I've come to hear what was missing before a greater percentage of the time listening to higher-quality files.

Again, my personal experience is that repeated listening to just about any recording will eventually lead to my hearing things in it that I've never noticed before. This is why, when you're setting up a blind A/B comparison, you want to do it over multiple sessions with the tracks shuffled a different way each time, record your guesses as you go, and check them only at the end of the trial.
posted by flabdablet at 8:30 PM on April 26, 2013

I have surely done blind tests in by which a 128-bitrate was deemed to be the lossless music over an actual WAV file. Busted!

Busted? Maybe. It might just be that for that particular track you actually prefer the sound of the 128kb/s MP3; compression artifacts are not necessarily unpleasant. See also: valve sound.
posted by flabdablet at 8:33 PM on April 26, 2013

Response by poster: I'm pretty sure I've done that a few times at least. Results have varied, but if, as you say, 320 and Lossless are indistinguishable, that could be expected. Well, the point of this is only about what most readily converts TO lossless on a portable, since I do most of my listening on commutes and love my Brainwavz earbuds. And since this thread I've noticed any worry factor or visions of labor recede.
posted by noelpratt2nd at 8:35 PM on April 26, 2013

Response by poster: Yes to the post on 128s sometimes sounding better for the moment...
posted by noelpratt2nd at 8:36 PM on April 26, 2013

I do most of my listening on commutes and love my Brainwavz earbuds

I would be astonished to find the background noise level on a commute insufficient to mask compression artifacts for any encoding at least as good as 192kb/s MP3, even using high quality noise blocking earbuds.
posted by flabdablet at 8:44 PM on April 26, 2013 [1 favorite]

Response by poster: Well, it's in some of those those cases that, depending on my ears/brain at the time, I've heard simply better-sounding output. And when you get somewhere quiet, you don't want 192.
posted by noelpratt2nd at 6:38 AM on April 27, 2013

Only reason I mentioned 192kb/s MP3 is that it's almost always adequate for a portable device and earbuds, especially with background noise; in my experience it's far better sounding than 128kb/s MP3.

320kb/s MP3 or WMA is another huge leap better. Even if your portable device really is playing at "only" 320kb/s, I can't see any way compression could be the cause of any quality difference you're hearing between that device and something higher-spec being fed directly from a lossless source; at 320kb/s, compression artefacts will be totally swamped by small deficiencies in the player's audio amplifier and the earbuds themselves.

If you do take the time to recode terabytes of WMA-lossless audio as FLAC, you might well experience better sound on your commute - but that improvement will absolutely be placebo-effect-driven; we're firmly into wooden volume knob and oxygen-free speaker wire and green marker territory here. Only you can decide whether there''s an equally effective placebo involving less effort.
posted by flabdablet at 10:58 PM on April 27, 2013