Hi guys, as MP3 and Opus have very similar objectives, I think the original poster's question was a valid one: Why does Opus have more artefacts in the lower frequency ranges than MP3 has? The spontaneous suspect that lower frequency artefacts may be more noticeably than higher frequency artefacts seems plausible, also. Is it a matter of energy (which is higher for higher frequencies)? When your own ears are no longer in their best possible condition, you may try a spectrogram, just to make sure you don't miss anything. Regards, Ulrich>>> Jean-Marc Valin <jmvalin at jmvalin.ca> 31.10.17 3.09 Uhr >>>Just to be clear, my goal here wasn't to make fun of anyone, but to drive the point that spectrograms should *never* be used to demonstrate quality. The only case where they can sometimes be useful is for diagnostic purposes. If you hear something and you're not sure what you're hearing exactly, then sometimes a spectrogram can help you figure out what it is. That's pretty much it. If you can't hear any artefact, who cares what the spectrogram looks like? Also, looking at the difference signal (either as a spectrogram or as actual audio) is particularly dangerous. There are many things you can do to an audio signal that are completely inaudible and yet will cause quite large differences (e.g. flip the sign or delay by X samples, but there's many more). Cheers, Jean-Marc On 10/30/2017 08:16 PM, Orestes Zoupanos wrote:> Jean-Mark sarkasm. > > Jean-Markasm. > > (Bonus points for providing an actual noisy WAV! ^_^) > > On 30/10/2017 20:28, Jean-Marc Valin wrote: >> Hi, >> >> Before I comment on the graphics you posted to visualize the difference >> between two audio signals, I'd like to ask for your help in evaluating >> my JPEG encoder. I've encoded an image with JPEG and then computed the >> difference with the original. I then converted the difference to sound. >> You can listen to the image difference on this clip: >> https://jmvalin.ca/misc_stuff/diff.wav >> >> Can you hear how good the visual quality is? Do you think it could be >> improved to make JPEG sound better? Personally, I think JPEG could do >> better on my subwoofer. >> >> Cheers, >> >> Jean-Marc >> >> On 10/18/2017 07:08 PM, encrupted anonymous wrote: >>> Good morning. >>> >>> I've ran a test against MP3 format. >>> >>> Code: (first convert tested audio file to 16 bit 48khz with sox.exe if >>> needed) >>> lame.exe -b 320 48khzfilein.wav -o fileout.mp3 >>> lame --decode fileout.mp3 -o fileout.mp3.wav >>> opusenc.exe --bitrate 320 48khzfilein.wav fileout.opus >>> opusdec.exe fileout.opus fileout.opus.wav >>> wavdiff.exe 48khzfilein.wav fileout.mp3.wav -diff fileout.mp3.delta.wav >>> wavdiff.exe 48khzfilein.wav fileout.opus.wav -diff fileout.opus.delta.wav >>> >>> Results: (compare two deltas with spek.exe - i've attached graphic file >>> from my test) >>> MP3 much better at 0-4 kHz, Opus little better at 12-20 kHz. >>> Plus I think 0-4 kHz is more important than 12-20. >>> >>> Current Opus 1.2.1 is the best at 32 kbit/s for music. >>> But if you input 44100 Hz audio and give 96-512 kbit/s, Opus pretty badly >>> spends that much bitrate because of frame rate conversion. >>> >>> That's all I wanted to say for now, good evening. >>> >>><https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>>>> Без вирусов. www.avast.ru >>><https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>>>> >>> >>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> >>> >>> >>> _______________________________________________ >>> opus mailing list >>> opus at xiph.org >>> http://lists.xiph.org/mailman/listinfo/opus >>> >> _______________________________________________ >> opus mailing list >> opus at xiph.org >> http://lists.xiph.org/mailman/listinfo/opus > > > _______________________________________________ > opus mailing list > opus at xiph.org > http://lists.xiph.org/mailma_______________________________________________opus mailing list opus at xiph.org http://lists.xiph.org/mailman/listinfo/opus
On 10/31/2017 07:08 PM, Ulrich Windl wrote:> as MP3 and Opus have very similar objectives, I think the original poster's > question was a valid one: Why does Opus have more artefacts in the lower > frequency ranges than MP3 has?I'm not sure, but my best guess would be "because MP3's window is very leaky and MP3 has to waste a lot of bits in the LF because of that". It could also be just the MP3 encoder being silly, or other things. Most logical explanations would be related to MP3 being bad than anything else.> The spontaneous suspect that lower frequency > artefacts may be more noticeably than higher frequency artefacts seems > plausible, also. Is it a matter of energy (which is higher for higher > frequencies)?Most signals have more LF energy than HF, so it's normal for the noise to look like that as well. If the noise is flat, then you have too much HF noise and you're wasting bits in the LF. In fact, that's exactly what I'm noticing in the spectrograms that are posted.> When your own ears are no longer in their best possible condition, you may try > a spectrogram, just to make sure you don't miss anything.Actually, that's the wrong way. Especially when the spectrogram is computed on a signal difference. For example, some codecs can alter the phase (or add a small delay) in a way that's imperceptible, and yet causes a large difference signal. At the same time, artefacts such as pre-echo will not be noticeable on the spectrogram of the difference signal -- even when it's audible and annoying. The least bad way of estimating how good a codec is at very high bitrate is to just measure the point where you can't ABX and assume that all codecs improve by about as much per kb/s once that point is reached. And that's mostly true. Cheers, Jean-Marc
>>> Jean-Marc Valin <jmvalin at jmvalin.ca> schrieb am 01.11.2017 um 06:08 inNachricht <b73e61f9-e689-72bc-ddc0-b74031ce8183 at jmvalin.ca>:> On 10/31/2017 07:08 PM, Ulrich Windl wrote: >> as MP3 and Opus have very similar objectives, I think the original poster's >> question was a valid one: Why does Opus have more artefacts in the lower >> frequency ranges than MP3 has? > > I'm not sure, but my best guess would be "because MP3's window is very > leaky and MP3 has to waste a lot of bits in the LF because of that". It > could also be just the MP3 encoder being silly, or other things. Most > logical explanations would be related to MP3 being bad than anything else. > >> The spontaneous suspect that lower frequency >> artefacts may be more noticeably than higher frequency artefacts seems >> plausible, also. Is it a matter of energy (which is higher for higher >> frequencies)? > > Most signals have more LF energy than HF, so it's normal for the noise > to look like that as well. If the noise is flat, then you have too much > HF noise and you're wasting bits in the LF. In fact, that's exactly what > I'm noticing in the spectrograms that are posted. > >> When your own ears are no longer in their best possible condition, you may > try >> a spectrogram, just to make sure you don't miss anything. > > Actually, that's the wrong way. Especially when the spectrogram is > computed on a signal difference. For example, some codecs can alter the > phase (or add a small delay) in a way that's imperceptible, and yet > causes a large difference signal. At the same time, artefacts such as > pre-echo will not be noticeable on the spectrogram of the difference > signal -- even when it's audible and annoying. > > The least bad way of estimating how good a codec is at very high bitrate > is to just measure the point where you can't ABX and assume that all > codecs improve by about as much per kb/s once that point is reached. And > that's mostly true.Hi! Unfortunately (as I pointed out) young children (those with the good ears) are rare among the developers (and rare as a resource for developers, I'm afraid). Is there any analytic tool that can substitute good ears? Or don't we know how audible perception works, still? ;-) Regards, Ulrich
On 2017-11-01, Jean-Marc Valin wrote:> I'm not sure, but my best guess would be "because MP3's window is very > leaky and MP3 has to waste a lot of bits in the LF because of that". > It could also be just the MP3 encoder being silly, or other things.Was the original poster speaking about the SILK or the CELT derived mode? Because at least wrt SILK (and the rest of the LPC derived codecs) there is an additional explanation: perceptually matched synthesis. If you look at things like the GSM/UMTS series codecs, they can seem *really* bad in a naive noise floor comparison, while sounding better at low bitrates and when applied to speech-like signals. That's pretty much the reason LPC derivatives are used in the first place, and in OPUS as well: the algorithm itself somewhat models human speech production, so that it encodes salient information not easily measured by simple-minded, even if highly developed and complex, spectral and/or statistical coding methodologies.> Most logical explanations would be related to MP3 being bad than > anything else.One obvious problem at the lowest end is its filter bank. It simply hasn't the resolution to model what's happening at the lowest of the low end, perceptually speaking. After all at LF we get the heavy attenuation of the cochlea, combined with *extremely* compressed pitch sensitivity. No sane MDCT-based coder allocates many bits there, and LPC-based ones can often do even better because of the pitch sensitivity side of things.> Most signals have more LF energy than HF, so it's normal for the noise > to look like that as well. If the noise is flat, then you have too > much HF noise and you're wasting bits in the LF. In fact, that's > exactly what I'm noticing in the spectrograms that are posted.Yes. And of course many of the fundamentals live in the LF range. It'd be more useful to post spectrograms which normalized any residual noise by the utility signal reconstructed by the codec; i.e. time-varying S/N ratios per band. Of course that's still rather naïve as well. But it'd be a start at least.>> When your own ears are no longer in their best possible condition, >> you may try a spectrogram, just to make sure you don't miss anything. > > Actually, that's the wrong way. Especially when the spectrogram is > computed on a signal difference. For example, some codecs can alter the > phase (or add a small delay) in a way that's imperceptible, and yet > causes a large difference signal.My favourite is what happens with ambisonics, and especially NFC-HOA. You're simply not *allowed* to do time-coherent detection with that stuff.> The least bad way of estimating how good a codec is at very high > bitrate is to just measure the point where you can't ABX and assume > that all codecs improve by about as much per kb/s once that point is > reached. And that's mostly true.That's also the reason why you can't -- as of now at least -- make do with just one single coding concept at all bitrates and with all utility signals: doing it the empirical way leads to things like OPUS which just happen to work much better in practice. -- Sampo Syreeni, aka decoy - decoy at iki.fi, http://decoy.iki.fi/front +358-40-3255353, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2