Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: 'listening Fatigue' (Read 31890 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

'listening Fatigue'

I just read a thread on ogg vorbis on the server of the german c't magazine.
The post talks about a phenomenon he expereienced with ogg vorbis at high bitrates he calls 'listening fatigue'.

Quote-translated:
Quote
...above 196kbs the situation gets comlicated. You can listen to MP3s endoded at that high a bitrate easily for several hours without suffering from listening fatigue. Comparable Ogg Vorbis files (size, bitrate) make me nervous. Somehow the dynamics are missing and the high frequencies are being kind of 'cut off' and if you continue to raise the bitrate everything is getting worse.


Anyone with a comment? Garf?

P.S: Here is his LAME command line: -h -k -m s -v -V [2|0] (0 for CDs, 2 for radio recordings)

sic transit gloria mundi...

'listening Fatigue'

Reply #1 –
Quote
Anyone with a comment? Garf?

I think he's just nuts, but then again, I listen for hours on an end to high-bitrate oggs, so I probably suffered braindamage already.

Quote
Somehow the dynamics are missing


That is provably wrong...sigh

'listening Fatigue'

Reply #2 –
I was very tired after my listening tests. But it was with few thousands sessions with castanets.wav 

But, I noticed something like a poorer dynamic with ogg vorbis, but at low bitrate only. More noise, different noise, and somethink like a reduced dynamic (I tested classical music, with a very large dynamic range). Obvious at -q3  and below. I noticed it with abc/hr, when I compared ogg vorbis with LAME and mp3proVBR. And I prefer this as many distorsions of mp3 codec family at these bitrates.

After that, I tried to abx this at -q6 : it was for me impossible.
Wavpack Hybrid: one encoder for all scenarios
WavPack -c4.5hx6 (44100Hz & 48000Hz) ≈ 390 kbps + correction file
WavPack -c4hx6 (96000Hz) ≈ 768 kbps + correction file
WavPack -h (SACD & DSD) ≈ 2400 kbps at 2.8224 MHz

'listening Fatigue'

Reply #3 –
My nose starts to bleed after several hours of high bitrate OggVorbis...

You will find such screwballs all the time. The thing is if you ignore it, it'll create more screwballs due ppl taking it seriously what they say, if you're jumping in and start to straighten it out, you'll usually get tagged as idiot or smartass most of the time. Weird world it is in internet forums.

With the louder HFs in Ogg files. Frankly, if you don't know the original or aren't ABXing it, you won't really notice it IMO. At least I don't find it annoying.

'listening Fatigue'

Reply #4 –
Quote
With the louder HFs in Ogg files. Frankly, if you don't know the original or aren't ABXing it, you won't really notice it IMO. At least I don't find it annoying.

Well, to me what you say is similar. Ogg doesn't give files louder HF's. There exist a few clips where the stereo modes do badly at low quality levels (mostly < -q6, though I did nail down at least one problem case at higher bitrates) and cause more HF noise. I've also seen a few clips where there was preecho on cymbals or hihats and people described it as 'louder HF's'.

Ogg doesn't boost HF's, really

There's a few bad cases where it boils down to sounding something like that, but if you say it like that it sounds like a general problem, which isn't true, IMHO.

It's like saying MP3 preechos.

'listening Fatigue'

Reply #5 –
Quote
But, I noticed something like a poorer dynamic with ogg vorbis, but at low bitrate only. More noise, different noise, and somethink like a reduced dynamic (I tested classical music, with a very large dynamic range). Obvious at -q3  and below.

Hmm, as I said above, I know it can't actually be dynamics, because Vorbis doesn't touch them, so it'd be interesting if you could describe more closely what the problem is that you hear.

I'm guessing you're noticing the lowpass at the very low bitrate levels.

'listening Fatigue'

Reply #6 –
I won't pretend expertise on this particular subject, but for text fonts it is known that redundancy on the fonts (i.e. adding form to a letter that is already perfectly readable) contributes to reducing 'reading fatigue'.

At a symposium on audio coding I went to some time ago, it was mentioned that this effect could take place with audio coding as well, i.e. by reducing the redundancy of the audio, it would be more tiresome to listen to, even though the quality would be the same.

Of course this generated the well-deserved discussion, but the fact remains that there never went much effort in studying this subject...

'listening Fatigue'

Reply #7 –
Quote
At a symposium on audio coding I went to some time ago, it was mentioned that this effect could take place with audio coding as well, i.e. by reducing the redundancy of the audio, it would be more tiresome to listen to, even though the quality would be the same.

Hmm. As far as I know, a great deal of the removed information is information that cannot physically be perceived due to the contruction of the hearing(lowpass/highpass/temporal effects).

Comparing this to reading the shape of a letter is very hazy, at best.

And of course, it's no more than a speculation anyway that the effect exists at all.

If the choice is between assuming a complicated, unproven and unlikely effect exist versus a very simple explanation, (the person that said was just talking crap) which we've repeatedly observed to be true in practise, then I know what I'm going to assume.

Remember that science is coming up with simple explanations that can be observed and proven right, not complicated ones that exist merely because we think they may exist (aka magic).

'listening Fatigue'

Reply #8 –
Quote
Hmm, as I said above, I know it can't actually be dynamics, because Vorbis doesn't touch them, so it'd be interesting if you could describe more closely what the problem is that you hear.

I'm guessing you're noticing the lowpass at the very low bitrate levels.

Yes, of course, the lowpass was obvious : music color was poorer. But it perform very well at -q4

Sorry, but I can not be precise on « dynamic ». It was very hard for me to analyse the problem. Sound is different, but what exactly ? I can not really categorise the problem I hear. At moments (piano music if I remember correctly), I noticed that the music was I bit louder : obvious (maybe a decoding problem) - and brighter (noise tonality I think). Maybe that this loudness made me think taht dynamic was differnent, I'm not sure. As I said, it is more difficult to analyse correctly a problem than hearing it. And with my poor engliche, things are impossible to explain:P
But the piano music was not only louder I think ; difference between low and high volume seems to be less precise, less rich. I noticed something like this many time at -q1, -q2 and -q3... I should test it again, more precisely.
But as I said it before : it is more pleasant than many distrosions you can haer at the sae bitrate with mp3pro, wma, LAME, mpc or aac, more « analogic » sound.
Wavpack Hybrid: one encoder for all scenarios
WavPack -c4.5hx6 (44100Hz & 48000Hz) ≈ 390 kbps + correction file
WavPack -c4hx6 (96000Hz) ≈ 768 kbps + correction file
WavPack -h (SACD & DSD) ≈ 2400 kbps at 2.8224 MHz

'listening Fatigue'

Reply #9 –
Quote
Yes, of course, the lowpass was obvious : music color was poorer. But it perform very well at -q4


Hmm, -q3 lowpasses at 18kHz. That should not be 'obvious'...

Quote
At moments (piano music if I remember correctly), I noticed that the music was I bit louder : obvious (maybe a decoding problem) - and brighter (noise tonality I think). Maybe that this loudness made me think taht dynamic was differnent, I'm not sure


At low bitrates, it's very possible that Vorbis makes things a little brighter. But changing loudness (and hence dynamics), I don't think so.

I'd be interesting if you have a clip where you think this effect is most obvious.

'listening Fatigue'

Reply #10 –
Quote
Remember that science is coming up with simple explanations that can be observed and proven right, not complicated ones that exist merely because we think they may exist (aka magic).


The whole point is that science never came up with any answers on that subject.
And science is coming up with highly theoretical and unprovable explanations as well, and although subject to heavy debate, they contribute to our knowledge of the world as well as the observable and provable part.

Merely because we can't prove things exist, we can't prove that they don't...

'listening Fatigue'

Reply #11 –
Quote
The whole point is that science never came up with any answers on that subject.


Of course not, it wasn't proven the effect _exists_. It's a pure hypothesis that very maybe it does. Until we know it does, why try to come up with explanations? That makes no sense.

Quote
And science is coming up with highly theoretical and unprovable explanations as well, and although subject to heavy debate,


Exactly, and as long as they're unproved, they remain unsubstantiated hypothesi which aren't generally accepted.

Quote
Merely because we can't prove things exist, we can't prove that they don't...


If there's no proof, the scientific way is to choose the simplest explanation that has most evidence going for it.

You can't prove magic exists, and you can't prove that it doesn't, either. As far as science is concerned, it doesn't.

'listening Fatigue'

Reply #12 –
Quote
Hmm, -q3 lowpasses at 18kHz. That should not be 'obvious'...

Ooopss... mistake.  I said that -q4 performs well, and not that -q3 was bad. I'm not very sensitive to HF lowpass. I focused my earing on artifacts. I prefer poorer sound that an accidented one 


Quote
At low bitrates, it's very possible that Vorbis makes things a little brighter. But changing loudness (and hence dynamics), I don't think so.


Brighter : this is not only possible, it is really true. I can not only ABX vorbis with that particularity, but identify the codec. Note that at high bitrate, on castanets.wav (the one with the guitar), the problem is still present, with a lot of pre-echo too.

Quote
I'd be interesting if you have a clip where you think this effect is most obvious.


OK, i must did test again and found the samples (and time enough : my holidays are finished  )
Wavpack Hybrid: one encoder for all scenarios
WavPack -c4.5hx6 (44100Hz & 48000Hz) ≈ 390 kbps + correction file
WavPack -c4hx6 (96000Hz) ≈ 768 kbps + correction file
WavPack -h (SACD & DSD) ≈ 2400 kbps at 2.8224 MHz

'listening Fatigue'

Reply #13 –
As far as everbody except Galileo was concerned, the earth was flat.
The simplest explanation was that a ship would just fall of the edge at some point, since some strange pulling force holding ships onto the surface could of course never exist. With so many ships disappearing, evidence was abundance.

BTW, should we move this thing OT? Or create a special dir for 'Audio Rumors'?

'listening Fatigue'

Reply #14 –
Quote
With so many ships disappearing, evidence was abundance.


That wasn't evidence.

Your 'simplest explanation' argument doesn't hold either.

You don't have to move this to OT, I'll stop discussing here anyway. If you want to continue arguing we move our scientific scepticism back to the dark ages, it will be with someone besides me.

Oh, and if you look for an audio rumors site, there's plenty of 'audiophile' forums around.

'listening Fatigue'

Reply #15 –
Quote
Hmm, -q3 lowpasses at 18kHz. That should not be 'obvious'...

A question about that. This lowpass seems to be very high for a so low bitrate. If I remember correctly, RC3 cut the HF at 18500 even at -q10. I found that behaviour strange for this high bitrate (mpc and aac have complete range, and LAME --insane is higher too).

But the new Vorbis I made me sceptic too. Is that new behavior really adequate to that low bitrate ? Is vorbis I like mp3pro, with a SBR like system encoding, which make a so high lowpass possible ? In other words, could Vorbis did a better job (less artifacts) with a less optimist lowpass (16000 for -q3 ou -q4 are more common value) ?


Don't be angry, but in my opinion, this high lowpass was deceided in order to seduce a lot of people. For those people who never did listening test (and would not do), but a graphical comparison. Vorbis is so wonderful compared to LAME on CoolEdit - and most people choose their codec on that fallacious criterion. I read on french forum that vorbis is better than other codecs for that reason. I have read that people prefer vorbis -q5 at mpc --standard for that reason too.


Of course, it is just an opinion - mine - and without any proof (I don't know how changing lowpass frequency with Voribs). Can you enlight me on this particularity of Vorbis ? Do you think that a lower lowpass could affect artifacts reduction with the vorbis codec ?
Thanks...
Wavpack Hybrid: one encoder for all scenarios
WavPack -c4.5hx6 (44100Hz & 48000Hz) ≈ 390 kbps + correction file
WavPack -c4hx6 (96000Hz) ≈ 768 kbps + correction file
WavPack -h (SACD & DSD) ≈ 2400 kbps at 2.8224 MHz

'listening Fatigue'

Reply #16 –
The hf problem is quite apparent to me. I listen to mostly trance and it can be heard quite easily on those samples. Maybe because they use a lot of filtered noise instead of real world cymbals. I don't know if I would say the highs are louder, they are just more present and sound different. Hard to put in words. Sounding different is probably the best explanation I can give you. It seems like the hf noise of cymbals has a different tonality. Or like there is more noise in the hf region alltogether than in the original. To me the difference is apparent up until q6. Didn't try anything higher (but I suspect I will fail to abx at >= q7). The sample where I think it can be heard best is: "Inqdine - Death End Street" from the "Eye Trance 04" sampler. Contact me if you want to host the sample. The FLAC is about 2,2mb.
I tried ogg q4 on a lot of samples and on almost all, the first thing I heard wrong with the oggs was the changed high frequencies. The artifact per se is not annoying, but I personally seek transparency and not just unannoying music compression.

'listening Fatigue'

Reply #17 –
I made a post a while ago where I mentioned that I encoded some whitenoise for fun (due boredom) and noticed that the Ogg output is brighter and had a diff 'tonality'/pitch. I guess that's the same what's happening to the HFs on your cymbals.

Well, as said before, if you have never heard/dont remember the original song/file, or you aren't ABXing it, you will not notice the different HFs.

'listening Fatigue'

Reply #18 –
Quote
My nose starts to bleed after several hours of high bitrate OggVorbis...

You will find such screwballs all the time. The thing is if you ignore it, it'll create more screwballs due ppl taking it seriously what they say, if you're jumping in and start to straighten it out, you'll usually get tagged as idiot or smartass most of the time. Weird world it is in internet forums.

With the louder HFs in Ogg files. Frankly, if you don't know the original or aren't ABXing it, you won't really notice it IMO. At least I don't find it annoying.

If he had stopped at stating fatigue (and not mentioned high frequencies, etc) it would be a fair assessment... perceptual audio codecs are just that (perceptual, subjective) and if this guy felt fatigued after listening then maybe something doesn't agree with him.  It could even be he's just more used to the MP3 sound.

I'll say that listening to vinyl records is less 'fatiguing' to me than listening to CD's, and I hope nobody thinks I'm a screwball... it's just my perception (shared by many).  I make no claims or guesses beyond this whatsoever as to why.

'listening Fatigue'

Reply #19 –
 
i think that using some vitamines can enhance your listening abilities
espacially vitamine B1 ans B6 (only one cachet) per a day (of 250mg)
dont use more

'listening Fatigue'

Reply #20 –
Quote

i think that using some vitamines can enhance your listening abilities
espacially vitamine B1 ans B6 (only one cachet) per a day (of 250mg)
dont use more

Actually, lots of unusual things can affect listening abilities and/or enjoyment.  Very few people realize that ear wax buildup can have a major effect!  So can listening soon after either complete silence or a noisy background (residual effects), and the effects of various background noises (computer fans, anyone?). 

As for enjoyment, one's general mood can have a powerful effect.  The guy who claimed fatigue after listening to .OGG, it could have been as simple as not getting enough sleep that night... who knows?

'listening Fatigue'

Reply #21 –
Quote
It could even be he's just more used to the MP3 sound.


We call that the Placebo effect. 

Quote
I'll say that listening to vinyl records is less 'fatiguing' to me than listening to CD's, and I hope nobody thinks I'm a screwball... it's just my perception (shared by many). I make no claims or guesses beyond this whatsoever as to why.


In my opinion it's all just a matter of objective fidelity no matter which way you split it.  B)
budding I.T professional

'listening Fatigue'

Reply #22 –
Quote
Quote
It could even be he's just more used to the MP3 sound.


We call that the Placebo effect.

That's not really the same thing.  The placebo effect is where you perceive an effect where there is none due to your expectation that there would be one (for example thinking you feel better after taking a pill or thinking audio sounds worse after it's been compressed).  It's at least conceivably possible that someone could be used to the particular artifacting common in MP3, so listening to music that doesn't have such artifacting (or has different artifacting) seems strange and possibly fatiguing due to the unfamiliarity.  Of course there might be no such phenomenon and it might just be a placebo effect, but if it is in fact caused by being used to the MP3 sound, that's a different phenomenon from the placebo effect entirely.

'listening Fatigue'

Reply #23 –
Quote
Of course there might be no such phenomenon and it might just be a placebo effect, but if it is in fact caused by being used to the MP3 sound, that's a different phenomenon from the placebo effect entirely.


I see where you are getting at. It just seem's a little to physcological for me.
budding I.T professional

'listening Fatigue'

Reply #24 –
Quote
I see where you are getting at. It just seem's a little to physcological for me.

Hehehe a little too psychological - placebo ... hehehe get it?? 
You can't kill the Elephant.