[Jazic]

I'm not trolling or anything. My intent is just to share an interesting video I found on reddit.com/r/audiophile
 
https://www.youtube.com/watch?v=-Bd99cADk70

 

[Strangelove424]

Maybe they consider this trolling on other sub-forums, but most everyone in these neck of the woods would agree with that sentiment. I've seen other videos from that YouTube channel, that guy is pretty knowledgeable and honest with his statements. He's also quite balanced, I've heard him admit that high-resolution can be good for editing/mixing but has zero benefit for distribution. So he's not coming from an extremist view-point, just an extremely practical one... and you could say as much (hopefully) about those of us who hang out in this forum. 

 

[Jazic]

Agreed. For mixing and editing having the absolute highest quality possible is important. I need to do some digging and see if he shows the differences between those 16bit wav files and high quality mp3's or the like. Just to show the difference using the phasing. 

 

[bigshot]

The reason you need 24 bit for recording is because you are adjusting volume levels radically at times. For instance, almost all vocals are heavily compressed to make the words not get hard to hear under the music. If you apply compression to a 16 bit recording, the compressor is pulling up the noise floor as it boosts the consonants. It sounds like a sputtering steam pipe. But with 24 bit, the noise floor is so low, the compression is totally inaudible and the vocals are natural sounding and clear.

 

[RRod]

So now "stochastic resonance" is why people can hear the difference between 24 and 16 bit, even though when you ABX test them they can't?

 

[esldude]

  The reason you need 24 bit for recording is because you are adjusting volume levels radically at times. For instance, almost all vocals are heavily compressed to make the words not get hard to hear under the music. If you apply compression to a 16 bit recording, the compressor is pulling up the noise floor as it boosts the consonants. It sounds like a sputtering steam pipe. But with 24 bit, the noise floor is so low, the compression is totally inaudible and the vocals are natural sounding and clear.

That also gets to why I would prefer 24 bit for distribution at least for some music.  I use digital room correction and a similar thing can happen (though it has surprised me how rarely) that correction will pull up something a bit audible that wasn't before.  Though the correction itself may just be unmasking things.  And to reiterate for some music.  Very little commercial music is unmolested enough for this to matter.

 

[Strangelove424]

  So now "stochastic resonance" is why people can hear the difference between 24 and 16 bit, even though when you ABX test them they can't?

 I'm not sure I follow, what does "stochastic resonance" have to do with mastering or manipulating data in higher bit depths? One involves the boosting of the signal frequencies by resonating them with noise, while the other aims to boost the signal in as clean a way as possible to reduce the presence of noise. The two concepts are at odds with each other. 
 
It's the audible headroom for noise and distortion in hi resolution data that makes it beneficial for data manipulation (mixing/mastering) and simultaneously non-beneficial for final delivery. Nobody is saying you can hear 16bit vs 24bit differences on the original recorded file, but drag those files through a large signal boost or EQ shift and the 16bit version will pick up more noise and distortion. The final file delivery is a different thing. Nobody is changing the data during playback, so the headroom you get with high resolution has zero benefit at that point.

 

[bigshot]

 
That also gets to why I would prefer 24 bit for distribution at least for some music.  I use digital room correction and a similar thing can happen (though it has surprised me how rarely) that correction will pull up something a bit audible that wasn't before.  Though the correction itself may just be unmasking things.  And to reiterate for some music.  Very little commercial music is unmolested enough for this to matter.

You're probably right. Also, if your room correction is going so far that it's bumping into the noise floor of redbook, odds are it is being messed up in a bunch of ways. In my theater, I tried to adjust the room as close as I could get it to sounding good before I started making EQ corrections. It doesn't pay to do massive corrections to fix grossly off room acoustics. You just fix one thing and it causes another.

 

[Strangelove424]

 
That also gets to why I would prefer 24 bit for distribution at least for some music.  I use digital room correction and a similar thing can happen (though it has surprised me how rarely) that correction will pull up something a bit audible that wasn't before.  Though the correction itself may just be unmasking things.  And to reiterate for some music.  Very little commercial music is unmolested enough for this to matter.

 

 
You're probably right. Also, if your room correction is going so far that it's bumping into the noise floor of redbook, odds are it is being messed up in a bunch of ways. In my theater, I tried to adjust the room as close as I could get it to sounding good before I started making EQ corrections. It doesn't pay to do massive corrections to fix grossly off room acoustics. You just fix one thing and it causes another.

 
You guys bring up good points about room correction. I have a room correction function on my receiver, so I guess data manipulation is actually quite normal on a lot of systems even for playback. But if the boost is big enough that it brings up the noise floor of the material over the receiver's own noise floor, then I think bigshot is probably right, you're better off addressing the acoustic problem directly. My receiver always tries to boost the bass in my subwoofer a ton, so I set up the receiver with the subwoofer on a higher setting than normal, then turn it down once calibration is finished. I probably have some phase issues with bass in my room, but compensating by boosting the levels too much causes more harm than good.  

 

[bigshot]

The thing to remember though is that there are crossovers all over the place. Each multi driver speaker has a crossover, and there is a crossover to the LFE channel. Even if the sub was boosted a ton, the only noise would be below 80Hz. I don't think it's very likely that room correction would bring the redbook noise floor up. More likely the noise floor from a crummy amp, or distortion from overdriving would be the culprits of noise.

 

[RRod]

   I'm not sure I follow, what does "stochastic resonance" have to do with mastering or manipulating data in higher bit depths? One involves the boosting of the signal frequencies by resonating them with noise, while the other aims to boost the signal in as clean a way as possible to reduce the presence of noise. The two concepts are at odds with each other. 
 
It's the audible headroom for noise and distortion in hi resolution data that makes it beneficial for data manipulation (mixing/mastering) and simultaneously non-beneficial for final delivery. Nobody is saying you can hear 16bit vs 24bit differences on the original recorded file, but drag those files through a large signal boost or EQ shift and the 16bit version will pick up more noise and distortion. The final file delivery is a different thing. Nobody is changing the data during playback, so the headroom you get with high resolution has zero benefit at that point.

 
Oh, I was referring to the reddit thread; probably should have posted the link ^_^ :
http://www.reddit.com/r/audiophile/comments/2lfnzk/the_difference_between_24bit_16bit_audio_is/

 

[Strangelove424]

   
Oh, I was referring to the reddit thread; probably should have posted the link ^_^ :
http://www.reddit.com/r/audiophile/comments/2lfnzk/the_difference_between_24bit_16bit_audio_is/

Oops, sorry. I see the context now. I probably should have clicked the reddit link a while ago. After I've seeing a few 24vs16 threads turn into civil wars that rage on for years, I kind of avoid those scenes nowadays.  Except for the discussions on Sound Science, which tend to be more sensical and down-to-earth, which I am always happy to engage in.
 
    

  The thing to remember though is that there are crossovers all over the place. Each multi driver speaker has a crossover, and there is a crossover to the LFE channel. Even if the sub was boosted a ton, the only noise would be below 80Hz. I don't think it's very likely that room correction would bring the redbook noise floor up. More likely the noise floor from a crummy amp, or distortion from overdriving would be the culprits of noise.

 
I agree that it's far more likely it's the amp or speakers hitting the noise or distortion limit before the CD. With top of the line receivers having an SNR of 95-100db, that's right in line with CD's SNR.
 
In my case, I reduce the bass not because of noise but boominess. The bass is not evenly distributed across my room, and the calibration can;t account for the fact I walk around alot when listening to music. The mic also can't pick up the low frequencies going through the love seat next to the subwoofer. Audyssey seems to exaggerate the LFE level to compensate for the null spot which is unfortunately just in front of my listening position (and the location I am forced to position the tripod-mounted microphone). I need well-placed bass traps like mad, but it can be expensive. 

 

[bigshot]

I don't think most automatic room correction systems work worth a damn.

 

[Strangelove424]

The regular distance and loudness measurement function that balances levels between fronts and surrounds is very accurate in my experience. But when you start getting into EQ, that's far more complex and I don't think an automated system can measure all the acoustic variables in a room, or completely compensate for acoustic problems.

 

[bigshot]

The problem is that the EQ and the relative volumes are interrelated, so if you adjust the EQ, it affects the volume and you have to readjust that. When I was calibrating all that went in circles as I fine tuned more and more accurately over a period of over a month.