Some thoughts on digital recording

[miroslav]

But how music affects us emotionally is outside the realm of science. Or at least the type of science I address. That's more psychology, and possibly psychoacoustics. All I ever address is how fidelity is defined, and what properties of sound are and are not audible. If we can't hear stuff higher than 20 KHz, then ultra-high samples rates and more expensive equipment to support them are not needed.

Right....your interest and argument stops at the purely physical and currently measurable., and that's your choice.
However, if the point of audio/music is what overall effect it has on us, and that when listening to music, we don't just *measure* incoming audio for the sake of the science, rather we receive it for that effect ....
....then how can we be so sure that after some point (aka 20 kHz), nothing else is important for said effect?
I think that's where your disconnect is.

You've made up your mind that stuff beyond that is not important, but again, I say there is no absolute proof of that yet, and there has been very little study done with that in mind. Most of science has just focused on the physical, which is a good foundation, but with music there's more than that....IMO....and I do hope there is.

It's funny how physics and science can cling to a given "proof" for a very, very long time as an absolute...and then one day someone invents a new way to measure things or makes yet one more new discovery, and then everyone runs around rewriting their theories....

 

[grimtraveller]

Miroslav, while I partially agree on where you're coming from, I think that yours and Ethan's core points are coming from two very different places and actually address two very different things. There are obviously overlaps because in the end, we're talking about sound. But one is talking about measurable things in order to demonstrate that there are certain myths or questionable information paraded as fact in terms of recording and recorded sound while the other is talking about things that can't really be objectively measured but which many, many humans, with all our flaws and fluctuations feel and perceive and which we'll talk about forever and a day because the individual input doesn't require science, just passion and that's what humans are more the stuff of. Both sides are valid, both sides are {at least, to me} interesting but I don't see any conflict between the two because they are paradoxically, dealing with different channels.

 

[Ethan Winer]

how can we be so sure that after some point (aka 20 kHz), nothing else is important for said effect? I think that's where your disconnect is. You've made up your mind that stuff beyond that is not important, but again, I say there is no absolute proof of that yet, and there has been very little study done with that in mind.

There has been a ton of study on the perception of ultrasonics! And every test has concluded that we do not hear it, nor do we "perceive" it through non-auditory means. The only test I'm aware of that concluded we can hear ultrasonics (Oohashi et al) was later proven to be flawed. When the test was repeated without the flaw, nobody could hear it. So there is indeed absolute proof. Again, the disconnect is yours for not knowing about the existing research. Dude, if you had my Audio Expert book you'd know all of this.

then one day someone invents a new way to measure things or makes yet one more new discovery, and then everyone runs around rewriting their theories....

Repeat after me: Null test

If you don't understand why a null tests reveals all differences, including those you might not even know about, let me know and I'll explain it again.

--Ethan

 

[Ethan Winer]

I think that yours and Ethan's core points are coming from two very different places and actually address two very different things ... I don't see any conflict between the two because they are paradoxically, dealing with different channels.

Exactly. Thank you.

--Ethan

 

[miroslav]

I am aware of the Oohashi study from awhile back....but that there have been "tons" of other modern studies on psychoacoustics ...

....which ones?

I know that Oohashi was/is the only one ever mentioned by name, over and over in audio forum discussions.

 

[miroslav]

There are obviously overlaps because in the end, we're talking about sound.

My point exactly.
I know Ethan is just talking about audio fidelity, but I'm also talking about how it applies to the appreciation of music for each person.

Seems like there's a hard line being drawn as to what is important and what is not. All this "if I can't hear it/measure it in the 20-20kHz range it has zero significance"... is a bit myopic, IMO.
I don't argue against the obvious measured stuff....I just say there has to be more when it comes to music than just that.
Perceptions, emotions, feelings....hard to measure, but part of every piece of music, and delivered via sound.
What's really important....?

 

[Beck]

Nobody is trying to tell anyone else what they can or cannot hear. All that us "nerds" are trying to do is get people to understand that what they think they hear may not be real. Now, it might be real! But until you test yourself properly you can't know for sure. Speaking of which, I'm still waiting for your replies to this post:

https://homerecording.com/bbs/gener...iques/24bit-vs-16bit-hz-355305/5/#post4036793

I'm especially waiting for an email from you showing that you can hear what you claim regarding the test files linked here:

https://homerecording.com/bbs/gener...iques/24bit-vs-16bit-hz-355305/5/#post4037164

It's one thing to find you can't defend your position and leave a thread while questions are outstanding. But it's quite another to realize you can't defend your position, then continue to spout the same misinformation a week later anyway.



Hey, that's been my point all along! Test gear does indeed trump "the experience of professionals" when it comes to establishing what is real and what is imagined.

--Ethan

No conspiracy, I'm simply in the process of moving and the hard drive with the outlook pst file that contains the email about the HHB CDR-850 is in a box. Remember, from my perspective Meyer-Moran is old news (2007) and long discredited by myself and others. In computer terms that was several hard drives ago.

The email was from Chris Heap of HHB UK. The 16-bit loop was not 16-bit, so there's nothing about that experiment that can be redeemed. The converters are equivalent to 20-bit and the monitor signal is not converted to Red Book 16/44.1. But you shouldn't have to ask anyone now that you have the service manual. I say equivalent because that's exactly how Chris Heap worded it... There is no direct comparison between standard PCM and 1-bit schemes like Pioneer's Legato Link, which the HHB CDR's use.

Pop quiz: Which component/components on the circuit board of the HHB CDR-850 or the 24-bit CDR-830 readies the signal to write to CD-R at Red Book standards? (Obviuosly not the input ADC)

You can't answer that, and neither can Moran. This is the difference between speaking in broad theoretical terms, as you are, and speaking about the design and function of specific models, as I am. That's why my first instinct as an investigator was to examine the so-called "16/44.1 bottleneck." My sort of curiosity and skepticism solves "Perfect" crimes... and that ain't gonna change. I only get better at it as I get older.

When a researcher chooses to use a CD recorder as a converter he better know how it works. If he doesn't have a copy of the design patent perhaps he doesn't know enough. And why did he use a CD recorder anyway and not just dedicated converters? Red Book is a CD standard. Nothing is Red Book until something is actually written to CD, which in the Meyer-Moran debacle nothing was.

What's the rush anyway? You've been mistaken about Meyer-Moran since it came to your attention. What difference will a couple more weeks make? I don't live on these forums. If I drop out for a while its because I'm busy... or at times a thread becomes pointless when the participants are obviously and irrationally unmovable. You've already revealed what you will do if/when you can see the fatal errors in Meyer-Moran. You'll simply ignore it and reference other experiments... though you once refereed to Meyer-Moran as, "Groundbreaking!"

Its more important for me that people see the comedy in the Meyer-Moran style approach rather that I must discover and expose the errors in an endless list of these misadventures into audio. We have a fundamental disagreement on the value of short-term controlled listening tests compared to real life over time. I hope that point is not missed. Consequently I'm not interested in your audio samples, but thank you anyway.

Be patient! I could kill you quickly (figuratively speaking in debate terms), but what's the sport in that?

On another note... I've spent my recording career becoming very good at choosing and using comparatively modest equipment to achieve fully professional results. If I had a dog in this fight it would be hoping that Meyer-Moran were correct. And as an analog advocate I would like nothing better than to show digital has not progressed at all, but it has.

 

[Mo Facta]

Ethan, what about the report in the 1993 Guinness Book of Records that subjects of a study in Russia were able to perceive oscillations up to 200 kHz when the oscillator was pushed up against their head? That's at least some indication that while we may not hear infrasonic sound it may in fact be subliminally perceived.

Babies with pneumonia have been known to hear well above 20 kHz.

Cheers

 

[Ethan Winer]

I know that Oohashi was/is the only one ever mentioned by name, over and over in audio forum discussions.

Yes, audiophiles love that study because it confirms their beliefs. But the tests were flawed. This is from my Audio Expert book:

There was also a study by Tsutomu Oohashi that's often cited by audiophiles as proof that we can hear or otherwise perceive ultrasonic content. The problem with this study is they used one loudspeaker to play many high-frequency components at once, so IM distortion in the tweeters created difference frequencies within the audible range. When the Oohashi experiment was repeated by Shogo Kiryu and Kaoru Ashihara using six separate speakers, none of the test subjects were able to distinguish the ultrasonic content. This is from their summary:

When the stimulus was divided into six bands of frequencies and presented through six loudspeakers in order to reduce intermodulation distortions, no subject could detect any ultrasounds. It was concluded that addition of ultrasounds might affect sound impression by means of some nonlinear interaction that might occur in the loudspeakers.

I know there have been many studies reported in the Journal of the AES and various acoustician publications, but I can't put my finger on them at the moment. Surely this test proves the point:

Audibility of a CD-Standard Loop Inserted into High-Resolution Audio Playback

They tested 60 people over a period of one year in 554 separate trials, and nobody was able to tell with statistical significance when a 20 KHz "bottleneck" was added to the signal path. So with Kiryu and Ashihara that's two, and I know there are many more. I'll see if I can pin some more down and provide links.

--Ethan

 

[Ethan Winer]

Seems like there's a hard line being drawn as to what is important and what is not. All this "if I can't hear it/measure it in the 20-20kHz range it has zero significance"... is a bit myopic, IMO.

Where is your evidence that things that can't be measured affect the fidelity of audio devices? And if you have no evidence, why do you choose to believe such a theory? How is your "theory" different from typical magical thinking?

--Ethan

 

[Ethan Winer]

Ethan, what about the report in the 1993 Guinness Book of Records that subjects of a study in Russia were able to perceive oscillations up to 200 kHz when the oscillator was pushed up against their head?

Google was no help. Got a link? I'll guess what was heard was mechanical vibration of some sort. Maybe the transducer created in-band IM products in the same way as the flawed Oohashi tests. Again, Meyer and Moran proved people can't hear ultrasonics, and so did Kiryu and Ashihara mentioned above. Have you ever tested this for yourself? It's not difficult with audio editor software running at 96 KHz and a pair of headphones that can go past 20 KHz which some do.

Babies with pneumonia have been known to hear well above 20 kHz.

I doubt pneumonia is even needed with such young (and small) ears. But how much is "well above?"

In the grand scheme of things, even if it were discovered that it's possible for some people to just barely tell when ultrasonic content is present or not, who cares? How does this affect the enjoyment of music, or the perception of sound being high quality? Have you ever heard a CD that sounds so amazing you can't imagine better sound? If so, that proves that content above 20 KHz is not needed.

To my way of thinking, high fidelity is a flat response with minimal resonance, low distortion, and a low noise floor. Those are the impediments to good sound, not a lack of ultrasonics. That people obsess over ultrasonics as proof of [who knows what] seems silly to me when so many more important things are well known and fully understood.

--Ethan

 

[Bobbsy]

I tried to stay out of this thread...I really did!

However, I have to say that even Ethan would concede that there may be things happening in audio beyond what we've learned to measure...yet. What's important, though, is to make sure that the things you're sure you're hearing really are there and not just a case of talking yourself into hearing something.

I've never been sure why phrases like "double blind test" seem to have become ridiculed around this forum. Even the most discerning listener, be they from the musician side of the forum or the technical side, can easily fall victim to hearing what they believe they should rather than what's really happening. I've had this proved to me time and time again through the use of double blind tests. I'll be positive I can hear one thing being better than another (generally led by price or a good name on the box) only to discover that, when I don't know which is which, I can't hear any difference at all.

None of this takes away at all from the magic and emotion music can generate. But, at the same time, don't assume the whole world will think something is better due to things where you've willingly allowed your perception to be deceived.

 

[Ethan Winer]

The 16-bit loop was not 16-bit

Your problem is that you write too much and read too little. What part of "Brad connected the recorder with the A/D/A conversion in the loop, and he measured the degradation to confirm that it was indeed a 44/16 'bottleneck'" from this post do you not understand?

This is a direct quote from Brad Meyer's email to me:

(1) However the signals at the analog inputs are first converted to digital, the recorder has to generate a 16-bit, 44.1 kHz digital stream to send to the disc. Conversely, that signal coming off the disc has to be converted back to analog with whatever limitations that 16-bit "bottleneck" imposes. (2) Assume for the moment that the initial conversion is to 20-bit words, which are then reduced by some carefully thought-out scheme to 16 before recording. To use those 20-bit words instead of 16, and to reconvert the (presumably dithered) bit-stream to analog for those outputs, would (besides constituting a very strange and difficult-to-implement design decision) give you a 0dBFS-to-A-weighted-noise-floor ratio of about 116 dB. That's not what appears at the outputs. What does appear is absolutely typical CD-quality audio, with a slight droop in the frequency response before 20 kHz from the playback filter and an A-weighted noise floor a full 24 dB higher, at -92 dBA /re/ full scale, right where you would expect it.

Beck, you have failed to show any evidence to support your claim that Meyer & Moran's bottleneck was 20 bits and not 16 bits, so I won't waste any more time with you. Of course you're welcome to continue talking to yourself.

--Ethan

 

[Ethan Winer]

I tried to stay out of this thread...I really did!

LOL, it's addicting isn't it?

I agree with all you said except this:

I have to say that even Ethan would concede that there may be things happening in audio beyond what we've learned to measure...yet.

Here it is again: Null test

See my Post #56 in this thread for a more complete explanation of why a null test proves there's nothing more to fidelity than what's already known.

--Ethan

 

[miroslav]

I know there have been many studies reported in the Journal of the AES and various acoustician publications, but I can't put my finger on them at the moment. Surely this test proves the point:

Audibility of a CD-Standard Loop Inserted into High-Resolution Audio Playback

They tested 60 people over a period of one year in 554 separate trials, and nobody was able to tell with statistical significance when a 20 KHz "bottleneck" was added to the signal path. So with Kiryu and Ashihara that's two, and I know there are many more. I'll see if I can pin some more down and provide links.

Nope....that only proves that "normal-to-loud noise" wasn't being *audibly heard* above 44.1 kHz ( at louder levels the noise cut-off WAS noticeable).
That's not a test that addresses the psychoacoustic effects of sound.

Like I said, the Oohashi was/is the only one that ever gets mentioned...and I don't think there have been any other modern tests done on that subject.
Every test (including all your double-blind, null tests) is only focused on the ability of subject to *audibly hear* things. The lack of any tests/studies on psychoacoustic effects doesn't mean that consideration of them is "magical thinking".

There are many things in the universe that are yet not provable and haven't been seen, heard or touched...and yet scientists are convinced of their existence.
You being so convinced that there is nothing more is equally "magical thinking"....or maybe it's even a bit of man's occasional arrogance that he's 100% sure he has the whole universe already figured out.

<EDIT>
I'm referring to psychoacoustic studies with infra/ultrasound. There haven't been many serious studies in those frequency ranges and their effects on humans, only psychoacoustic studies in the 20-20kHz range.

I even did some Google searches and could not find a "ton of studies"...or really anything serious/scientific.
I did find these links which are not true studies, but certainly make you stop and think before saying anything outside of 20-20kHz is insignificant and can be discarded.

http://icliverpool.icnetwork.co.uk/...47768&method=full&siteid=50061#story_continue

http://ppjg.me/2009/01/31/silent-so...e-all-digital-tv-broadcast-signal-connection/

 

[Beck]

Your problem is that you write too much and read too little. What part of "Brad connected the recorder with the A/D/A conversion in the loop, and he measured the degradation to confirm that it was indeed a 44/16 'bottleneck'" from this post do you not understand?

This is a direct quote from Brad Meyer's email to me:

"Blah blah blah"

Beck, you have failed to show any evidence to support your claim that Meyer & Moran's bottleneck was 20 bits and not 16 bits, so I won't waste any more time with you. Of course you're welcome to continue talking to yourself.

--Ethan

LOL!!! The evidence is in the design of the HHB CDR-850! Yeah ok... I agree you shouldn't be wasting any more time with anyone. Your time would be better spent learning to read a schematic! You're too all about Brad Meyer and his misconceptions. You have no objectivity whatsoever when you quote the musings of the person you should be scrutinizing.

Meyer's explanation shows a total failure to understand how the device works beyond the theoretical just as I was saying. He's speculating about the function of a device where no speculation is necessary because the design parameters are well documented.

I tried to explain the impact of the analog circuitry and the design decisions, like where to set the low pass filter cutoff, or as Meyer calls the output filter. Those things limit the theoretical potential of a given Bit depth and thus you can't determine the things Meyer claimed to determine by measuring dynamic range or any other analog parameter. LOL

It should be obvious at this point to everyone I'm talking over your head. To be fair most people are at the wiki level in understanding anything more than general theory, so you're by no means alone.

We all end up talking to ourselves when you cut people off who are shaking your paradigm. That's a brilliant! (Then you're always right... in your own mind that is)

By the way, I saw that quote from Meyer that he allegedly sent to your email on a forum in 2008 where he was debating with people about his methodology. It made me laugh out loud then too.

 

[Ethan Winer]

Ethan is clearly the more knowledgeable in this area so I relent and admit correction now. I don't agree with everything he says because I have my own convictions based in my own experience but that will never come to light because I can't prove it scientifically.

In a lot of ways, I'm on your side and in other ways, I find your conclusions difficult to believe which may be based in bias and/or misinformation ... I will obviously now venture into my own experiments and give your conclusions more thought.

And yet, today I saw a post by you in another forum that undid all we accomplished here.

My posts are censored in that other forum because they can't defend their position with science or logic either, so I'll reply here:

I unwittingly got into an argument with Ethan Winer on another forum and he just refuses to admit that poor analog components distort badly when they are pushed to the limits. My assertion was that running signals too hot produces distortion that, after stacking multiple tracks, accumulates.

I handed to you on a silver platter several short articles with test files you can download that prove what I said. You asked folks in that other forum what they thought, but why don't you just test this for yourself? You obviously have a recording studio and the ability to record sine waves at various levels right up to 0 dBFS. You even said above that you'd "venture into my own experiments." If only you would spend the 20 minutes to do that, you'd know for sure!

Same for stacking. Record different frequency tones on different tracks, mix them together, then use an FFT display to see if the mixed result contains significant levels of added THD or IMD frequencies. Or just watch my AES Audio Myths video which I'm certain I've linked for you many times. The section that shows clearly why stacking is a myth starts at 28:28 into the video. Have you ever watched this video? If not, why not? And did you even read my Perception article I also linked several times that addresses stacking and summing? You will learn a lot, I promise!

If you willfully refuse to even try to understand, then you are operating from a belief system. If that's the case for you, please just admit it and I will gladly give up. We all know it's impossible to change someone's religious beliefs with logic and evidence.

In my experience, depending on the source material, some circuits produce distortion WAY below clip point.

Yes, some circuits do that. Mostly gear that's nonlinear to add analog "color" intentionally. Or tube gear, or "old radios" you mentioned. But in the context of this discussion in this thread, we're talking about sound cards and outboard converters that almost always aim for a clean sound. Again, in less than half an hour you could measure this for yourself and then you'd know for sure.

Ethan cited his null tests and scientific experiments as proof that I was wrong.

They do prove you are wrong! If you don't understand that, then say so and I'll try to explain it again more clearly.

--Ethan

 

[miroslav]

 

[Ethan Winer]

^^^ LOL, exactly.

 

[Beck]

By the way... topics like this are the only thing keeping these interest-specific forums alive and preventing the total domination of social networking by facebook! On firearms forums its 9mm vs. 40mm or 45 cal. On motorcycle forums its Honda vs. Harley. Some forums have literally gone under because overzealous moderators tried to put these fires out. Everyone complains about this topic, but look at the hits compared to other topics! Trust me! Let it burn!