[bigshot]

Test sent.

 

[FFBookman]

  If you want to do the real test of 24bit masters verse 16bit downsamples I can offer up some materials and connections. I have about 10 albums in open FLAC @ 24bit, can do the low-end Apple conversion and the mid-range focusrite conversion to 16bit, and know someone with the converters to do the high-end versions.  I don't know that he will work for free though, I'd have to ask if he's interested in helping in the name of science.
 
I'd love to throw in DSD but it introduces different mastering sessions as a variable. I'd rather stick with a single source file at 24bit and test multiple down samples of the same source.

My 24bit albums I'd offer up for test material (we should pick something we all know and love)
 
Rolling Stones - Sticky Fingers
The Who - Who's Next
The Cars - The Cars
Sam Cooke - Portrait of a Legend
Led Zeppelin - III
Led Zeppelin - Physical Graffiti
Slave - Slave
Red Hot Chili Peppers - Blood Sugar Sex Magik
Bootsy's Rubber Band - Ahh The Name Is

 

[RRod]

  If you want to do the real test of 24bit masters verse 16bit downsamples I can offer up some materials and connections. I have about 10 albums in open FLAC @ 24bit, can do the low-end Apple conversion and the mid-range focusrite conversion to 16bit, and know someone with the converters to do the high-end versions.  I don't know that he will work for free though, I'd have to ask if he's interested in helping in the name of science.
 
I'd love to throw in DSD but it introduces different mastering sessions as a variable. I'd rather stick with a single source file at 24bit and test multiple down samples of the same source.

 
I offered what I offered. If you have a friend who can do conversions then he can help you out with your own tracks. I'd bet that most tracks wouldn't even need dither, so no fancy algorithms necessary; just chop off the bits.

 

[StanD]

   
I offered what I offered. If you have a friend who can do conversions then he can help you out with your own tracks. I'd bet that most tracks wouldn't even need dither, so no fancy algorithms necessary; just chop off the bits.

What no rounding? 

 You're gonna hear about that.

 

[RRod]

  What no rounding? 

 You're gonna hear about that.

 
I know, I've moved on from anti-scientific to sacrilegious!

 

[bigshot]

 
My 24bit albums I'd offer up for test material (we should pick something we all know and love)
 
Rolling Stones - Sticky Fingers
The Who - Who's Next
The Cars - The Cars
Sam Cooke - Portrait of a Legend
Led Zeppelin - III
Led Zeppelin - Physical Graffiti
Slave - Slave
Red Hot Chili Peppers - Blood Sugar Sex Magik
Bootsy's Rubber Band - Ahh The Name Is

Those are all analogue recordings, aren't they? And some of those aren't particularly well recorded. Shouldn't it be a native high bitrate recording?

 

[StanD]

  What no rounding? 

 You're gonna hear about that.

 
 

   
I know, I've moved on from anti-scientific to sacrilegious!

Perhaps the extra bits can be dealt with by an exorcism.

 

[FFBookman]

 
Those are all analogue recordings, aren't they? And some of those aren't particularly well recorded. Shouldn't it be a native high bitrate recording?

Yep. I think that's important. 
 
See, this "sound science" goes both ways.  Producers, mixing engineers, and mastering engineers adjust for the marketplace. The marketplace is the dominant format and the mainstream playback devices of the era.
 
When CD took over in the 80's it allowed for far more overall volume and subwoofer kick without skipping the needle. Producers took advantage of this. Music production changed. If you believe redbook has any weaknesses, producers also adjusted for those.
 
I find it's soundstage suspect, it's center undefined, and it's mid-high's completely unnatural and crispy. Its sub-lows are OK but it's mid-lows (like bass guitar or standup) are weak.  So producers worked ways to compensate for this.  Also, are you really going to hear a 24bit advantage if all your plug-ins operate at 16bit, if all the samples are 16/44 and if most of your digital instruments all create 16/44?  No.
 
When MP3 took over 15 years go it happened again, and modern muscs seems more alive, louder, and more present, but it's all due to creative ways to cope with bandwidth restriction. There's a lot less data there, trying to sound like more.
 
Going back to analog material allows us to hear music without any digital trickery, without any digital compensation for other degradation. It allows to hear real instruments, real analog reverbs, real room delays, real instrument timbre, analog EQ, and of course fine, professional microphones and recording techniques.
 
My entire belief is that 24bit PCM sounds closer to the analog tape masters than does 16bit PCM. So when I do listening tests I try to keep my material pure and obvious, and avoid modern material with it's multiple layers of deception.

 

[castleofargh]

   
Science must adapt and grow to learn how to perform proper music sound quality tests. 
 

like your uncontrolled ones that are the only way to get a result that agrees with you? you're sure that's how science should work. we should forget blind testing and rent a studio to make sighted listening? wow that's got to be the only real way, the method is so clean and objective.
please don't talk about what is science itself, a non stop questioning and evolving way to understand stuff, while anything you write screams the denial of method and the enforcement of your own bias.
 
you look for anything that is stereo to prove your idea is right, found the idea that dither might not be perfectly stereo and hang onto that idea without any concern for the fact that we're talking sound at -96db. plz tell me what sound system you use that has such a perfect left/right balance that you notice the -96db difference while everything else is perfectly aligned? we all want those devices and speakers/headphones(no really we do!). you're fighting the ant while carefully avoiding the elephant in the room.
but 16bit has to be the culprit because you decided it.
blind test doesn't say the same? well blind tests are wrong.
measurements say you're wrong? there is more than measuring a soundwave in audio!
you always have an excuse not to look at the truth.
the album was made with a few microphones and the guys most likely recorded one after the other in the same studio room on the same chair at the same place(at least for most modern stuff), the mix generated the synthetic stereo placement and probalby 2 microphones with 2different signals and different phases from where they are, where mixed to get one instrument to sound "better". but when you don't get all the great soundstage, it has to be 16bit. and in fact you can't know because you reject blind testing. so you're just stating your faith in 24bit while giving us advices about science. amazing!
 
 
now efeuvete talks about the recording of 2 different sounds as if it would make any difference to the movement of the membrane in a microphone. as if anything recorded was more than 1pressure at one moment at one point. as if the air could contract and expand instantly without limits and let you hear stuff like a pure square wave.
of course anything in the air is at least a portion of a wave. it has to else how would it travel? so yes we could record sounds that are not waves(generated on a computer for example). and then yes maybe we would need another format and 16/44 or even PCM would fail. but no speaker or headphone could make those sounds without massive distortions, because the air would desagree a little^_^. the result is that we hear waves(or portions of them) because of how the sound travels.  and that's why digital audio works.
but again it's as if recording in 16bit was the reason why sound is messed up. not microphones, mix, sound system, our ears. no all this is 100% perfection with all the care of phase alignment from A to Z and perfect retrieval of uncorrupted data thanks to magic unicorns.
who cares about all the stuff that are worst than 16bit? who cares that people fail to actually hear it? efeuvete has his idea and build up a theory that matches it, based on his very own understanding of audio. I see no science there.
 
analogsurviver also is by association anti 16/44, as he is pro anything way up. but this time it's for the ultrasounds and getting all the music by getting all the frequencies. rejecting the fact that most humans old enough to not believe in santa will most likely not hear them. so his foundation theory comes from something false. otherwise he does have a very rational reasoning. by that I mean that if ultrasounds were beneficial to music, then most of his arguments would stand. that's really not the case for efeuvete and you IMO.
 
so here are 3 people saying that 16bit is audibly inferior but not even for the same reasons ... and of course without an scienctific method to show it.
while the "we don't listen to music we just read graphs" secret society, made of most of the others in this topic(not arguing that numbers gives me truth, that's BS. just saying we agree for the same reasons at least), are saying 16bit vs 24bit isn't, or at least shouldn't be audible in normal listening. and we all explain it with:
1/ the difference between the 2 resolutions is at about -96db and below, everything above is the same. when it isn't, it has to do with the sound system not with 16bit.
2/ people repeatedly fail to identify differences this low in blind tests.
therefore people repeatedly fail to identify 16bit => we say the errors from dithered 16bit aren't audible to people on usual albums at usual listening levels. and we're still waiting for people to show us otherwise in a convincing way (because we do believe in the scientific method and would accept a repeatable method showing we were wrong).
 
we could give you the subconscious part where we still get it but we don't know we do. but if it is so, how did you know 24bit sounded better? see the slight flaw with that idea? if it's not conscious, well it isn't. and if the effect could manifest on a conscious level, then people wouldn't fail blind tests so consistently. it's one or the other but we can't bend reality just to fit in that 16bit is audibly bad.
 
 

Standard fail #1 -- music and sound quality is about way more than frequency.  Think outside of the waveform. Frequency is but one parameter of sound, why can't you people understand that?

ok.
a waveform includes amplitude and timing, there is zero need to think outside of a waveform. frequency obviously is just one parameter.  on a record we get rid of the timing and only keep the amplitude because it will be up to our replay device to put the samples at regular intervals.
the result is again a waveform with time and amplitude(else it wouldn't be a waveform). if there is more to sound than this, then it is on no record on earth because a microphone doesn't know better than making a waveform in volts.
and stereo is but 2 waveforms at the same time. if you have timing errors and you do at least in a measurable way, 16bit isn't your problem in that regard.
just try to get rid of your tunnel vision and look at the entire audio process and the individual precision of all parts. pretending that -96db noise is the reason for bad stereo rendering is simply ludicrous. just by looking at right/left balance of a headphone you should understand how insignificant dither is.

 

[StanD]

:
now efeuvete talks about the recording of 2 different sounds as if it would make any difference to the movement of the membrane in a microphone. as if anything recorded was more than 1pressure at one moment at one point. as if the air could contract and expand instantly without limits and let you hear stuff like a pure square wave.
:

C'mon haven't you heard of the Quantum Microphone. That's one of Prof. Emmet Brown's finest works.

 

[manbear]

Manbear, from my first post up to this one I had no intention of disturbing anybody but I'm afraid this idea is not correct:

"Nyquist applies to B in the exact same way it applies to A. There is no difference. Both are sounds with frequencies in the finite range of human hearing"

Sorry buddy. It's exactly correct. There is nothing magical or different about the instrument noise you're talking about. Sound is sound and it's as simple as that. All of it can be sampled the same way (by your ears too, not just by ADCs).

If you don't believe me, explain how your ears can physically hear these sounds that are impossible to sample.

 

[manbear]

Standard fail #1 -- music and sound quality is about way more than frequency.  Think outside of the waveform. Frequency is but one parameter of sound, why can't you people understand that?

My grandma is 99. I honestly have never tested mp3's on her.  I was exaggerating, you caught me. She likes polka cassettes.

But my 65 year old father in law and the rest of his friends he's played it for all hear 24bit improvement on the pono player. One guy heard it in 5 seconds, said "it's like the whole thing". His other friend, also in his 60's, said "it's like being in a recording studio". 

I am aware they are reacting partly to the signal chain ,that's why i'm interested in a 24bit vs 16bit downsample shootout. 

All that your ears physically pick up are frequency and amplitude. There is nothing more to sound. Stereo image, timing cues, etc. are simply the result of having two ears at once.

What is so special about Pono? You talk about it as if its the only way to experience 24 bit music

Have you considered that your 24 bit recordings are using different masters? Have you tried compressing one to 16 bits?

 

[lamode]

  Recorded music is all about the stereo presentation, timing cues, and the timbre of the tones. Your waveforms tell you nothing important as far as sound quality.

 
If the difference between the original waveform and the reconstructed waveform is under -100dB then the difference is inaudible in the real world. In other words, the recorded sound is inaudibly different from the original. I would say that's an absolute indicator of SQ.

 

[lamode]

 
I'm not trying to take you on a ride. I'm answering the accusation that 24bit PCM audio is a "myth" or "snake oil".  It's neither. It's been used in production

 
This has been raised and answered at least 20 times in this thread already. 24 bit is used in production for other reasons.

 

[lamode]

  Can anyone here PROVE to me that dither is perfectly stereo matched?

 
Inaudible in left channel.
Inaudible in right channel.
 
Yep, seems perfectly matched to me.