Spectrograms before and after conversion to flac

[hello_hello]

PS An irrational debate has broken out in the "CD to MP3" thread now where claims are being made that a certain well known lossless format isn't lossless, and that it's also impossible to make perceptually transparent MP3 encodings. I'm done there, but you might want to take a look.

Yeah I tried to resist, but couldn't help myself. It makes the wave vs MP3 debate almost appear rational.
Not that I'm even remotely close to being an expert on audio encoding, compression, or decoding etc, but if playing a FLAC file directly somehow changes the sound of the audio compared to uncompressing it first and then playing it, I'd be keen to know to know how that works.

[Slipster]

Yeah I tried to resist, but couldn't help myself. It makes the wave vs MP3 debate almost appear rational.

I know. Fun here, innit?!

...if playing a FLAC file directly somehow changes the sound of the audio compared to uncompressing it first and then playing it, I'd be keen to know to know how that works.

As even a CD-A or WAV file still needs decoding/converting to a bitstream to feed into a buffer of some kind on-the-fly during playback, there is no difference, but it will still be interesting to read the explanations of why this nonexistent difference allegedly exists. I wouldn't be surprised if the word 'jitter' comes into the conversation somewhere.

[Hoser Rob]

... AAC, AC-3 and MP3 are all crap if used incorrectly, ie, forced into a non-default encoding mode or starved of bitrate. They're also all capable of perceptual transparency (ie, sounding the same as the lossless source) under most circumstances if used correctly by a competent individual at the encoding end...

I don't really agree that aac or mp3 can be indistinguishable from lossless if the source recording is decent. With recordings made in the last 10 or so years, which are obscenely compressed, I agree you can't tell the difference really. With the crap so compressed there's only about 1dB of range there's no level information to lose anyway. But with a reasonably good recording there's no way in hell lossy audio codecs sound as good as lossless.

I definitely agree that you need a good encoder and you need to know how to use it. Not unlike video encoding but the complexity level's not quite so hideous.

You also need to configure your audio player and/or OS to get decent audio. With windows you need to use asio or wasapi drivers or the system will constantly try to resample the audio. Even lossless will sound bad when that happens. Linux is a bit easier. You just have to configure pulse/alsa to not muck around with it and use a decent player.

And, speaking as someone who still has their turntable and is knowledgeable about phono reproduction, I am absolutely sick of vinyl snobs, who are generally pretty clueless.

I know perfectly well vinyl can sound as good or better than cd but you'd have to be a moron in this day and age to spend the kilobucks a good TT costs and then try to find decent vinyl. The used selection has been pitiful for 20 years and almost all new LP releases are just made from the CD anyway.

All that 'warmth' vinyl people with their half assed 'tables talk about is noise modulation. Plain and simple.

[jman98]

Hoser Rob - The vast majority of human beings cannot tell a difference between the original source and MP3 files with high bit rates. Now if you want to argue that YOU can, that is one thing, but honestly most people cannot and really do not care at all about being able to do this. This really cannot be some kind of shock to you unless you have had your head in the sand. Have you never noticed that most North Americans just watch EVERYTHING in 16:9 on their HDTVs? Yet we are supposed to believe that because you can (or at least claim you can) tell a difference between high bit rate MP3 and the original source that everybody else can too? Sorry man, but you're on your own with that one.

I strongly disagree that vinyl can EVER sound "as good or better than CD". Certainly this is untrue against a properly mastered CD. This is just what the old school vinylphiles think, but I would just simply argue that given vinyl's inherent lower dynamic range compared to CD that this cannot be correct. A more accurate statement, which all vinylphiles refuse to accept, is that some percentage of the population has become used to how analog formats like LPs sound and they PERCEIVE it to be "better". All this stuff is mostly subjective anyway. Vinylphiles typically have more money than sense and they make a LOT of noise on various forums and spend a lot of time reinforcing each other's beliefs, so this has led to them drawing lines in the sand that vinyl is always better than CD and other such nonsense. I find your final comments a bit contradictory in that I give you full credit for honestly admitting that "vinyl snobs" are clueless, yet you feed the standard lie that "vinyl can sound as good or better than CD".

[hello_hello]

I don't really agree that aac or mp3 can be indistinguishable from lossless if the source recording is decent. With recordings made in the last 10 or so years, which are obscenely compressed, I agree you can't tell the difference really. With the crap so compressed there's only about 1dB of range there's no level information to lose anyway. But with a reasonably good recording there's no way in hell lossy audio codecs sound as good as lossless.

Well Slipster provided a sample of audio where he could tell if he was listening to the encoded version. So apparently could a few others here. If memory serves me correctly he said using a higher bitrate for that particular audio sample then makes it transparent for him, but anyway.....

Rather than offer generalizations, why not offer us a sample of this "decent source" to which you refer? Something we can all download, convert to a lossy format, then hear for ourselves how there's no way in hell a lossy audio codec could sound transparent. We could all try different encoders and different bitrates etc to see if we can get your "decent source" to sound transparent to us. Just one or two samples will do. I assume you have at least a couple of samples you could offer... unless of course you're just offering unsubstantiated theory as fact.

[Slipster]

Well Slipster provided a sample of audio where he could tell if he was listening to the encoded version. So apparently could a few others here. If memory serves me correctly he said using a higher bitrate for that particular audio sample then makes it transparent for him, but anyway.....

Thanks. It's worth bearing in mind that there will always be a few extreme cases that can catch any lossy encoder out, and this happened to be one of them. Note that the CD this was taken from is the only one from a collection of 200+ where '-V2' didn't more than suffice for me. YMMV.

Rather than offer generalizations, why not offer us a sample of this "decent source" to which you refer? Something we can all download, convert to a lossy format, then hear for ourselves how there's no way in hell a lossy audio codec could sound transparent.

Great idea! I'm up for it. Sounds like fun.

[pandy]

This thread is a disaster.

Why? Good thread, issue explained, workaround working - i would say 100% success, additionally good discussion.