Master Thread: Are measurements Everything or Nothing?

[Deleted member 21726]

The ONLY people this product lets down is...audiophiles. To the other 95%, its perfect.

Nope.
I’m very, very satisfied with RME. I’m subjectivist.
What is behind your comment? I don’ quite understand.

 

[Thorsten Loesch]

Hi,

> I agre with the entire post.
> But… there is some “boner de jur” with people lusting after the SINAD and SNR.

Sure. Some people buy car's by the numbers, others by looks, others by how they drive. I drive a Chopper bike in Captain America paintjob. Corners like a pig, but the looks I get from the chicks are worth it!

Some people buy gun's by numbers, others try, shoot and buy what get's them to hit the target reliably, quickly and without effort. A beat up Siam Mauser with 'scope and .357 special snub nose for self defence, heavily tactical air pistol for varmints here.

Some people buy impressive digital SLR's. I use an original 1930's pre-war Leitz Lense on an old Sony alpha 5000 body,

So some people pick Hifi for numbers, others for looks others for sound quality.

What is important is that as a consumer we know what we want/need and then put our money where the least money gets us the best result for our needs. The least money may still be a lot. But it will be value for money.

> At some point the distortion profile becomes important, which vanishes as the SINAD heads towards infinity.

There is more to that than just profile. Distortion needs to be actually pretty gross to be perceived, by modern (say 1970's HiFI) standards.

We can see that from the simple fact that few speakers or headphones are judged as being subjectively heavily distorted by listeners, despite the fact that it will be hard to find any speaker or headphone that exceeds 60dB SINAD at 105dB SPL.

Another thing, what is that SINAD fad? It is just THD & N. Can we just stick to established conventions? 60dB SINAD = 0.1% THD&N.

Yes, if we have all distortion below audibility, we will not have audible distortion. But is harmonic/intermodulation distortion, noise and frequency response all that may audibly alter our listening experience?

But let's take an example. The OPA2604 Op-Amp (external Class A biasing mandatory due to internal design) and the NE5532 (external Class A biasing optional but beneficial due to internal design) Op-Amp both have distortion and noise that under most conditions is reliably is inaudible. NE5532 actually has lower distortion and noise objectively speaking. Yet to me there are perceivable differences in sound quality, especially hard to quantify factors like emotional engagement. And wherever there is a difference there is a preference, in my case OPA2604.

Of course, we can take the NE5534 (single of 5532) with some interesting internal connections available and replace the input stage with high transconductance dual J-Fet's, increase the second stage current significantly and take the output from compensation pin with an external class A buffer with around 10mA Iq, now the "Franken 5534" will have more of the qualities that make the OPA2604 in my books preferable over NE5532, with lower noise and lower distortion than both originals.

It is tempting and east to jump to the conclusion that the lower distortion is the cause, but given that in all cases the harmonic distortion and noise is well below any audibility limit, it would seem that other, more subtle factors are at work here.

> It was seeing no chips on the board with 32 (or more) spider legs on it that had me wondering.
> And I did not recall seeing 32 bit DACS in the past.
> Routinely the sales are stored as 32bit float or integer, and most computers do not have native 24 bit registers, so 32 or 64 is a sensible container to hold 24 bit within.
>
> That chip is advertised as 24 bit, which is certainly enough.
> But a calling a 24bit DAC a 32bit DAC, is a bit of a marketing overbyte.

Maybe. But a DAC is not just a chip.

If a DAC (Device) accepts direct input via USB of a 32 Bit audio sample without requiring the source device to convert the 32 Bit Sample to 24 Bit, is that a 32 Bit DAC or not? If not, then what is? I can only accept the "194dB+ Dynamic Range" as alternative definition.

To me it is just one more number fetish. How many bit's are actually in a given music recording.

Say we use two Microphones with 20dB(A) noise floor to record a symphonic orchestra with 110dB peaks and 85dB average SPL for louder (tutti) passages and lowest actual musical passages at around 30dB (piano, solo).

In my experience this is about the greatest real world dynamics we get. Commonly it is necessary to "gain-ride" such a recording to make it listenable in domestic conditions, either live (like when we recorded to magnetic tape with 84dB dynamic range (-70dB noise floor and +14dB head room) in the 1980's. Nowadays you would use post production and high resolution (18 Bit or more) recordings to capture the original dynamics and do the "gain-riding compression" in post.

How many Bit's are required to represent this classical recording, uncompressed?

110dB - 20dB = 90dB = 15 Bit

So recording with a system that offers 18-19 Bit of real world resolution will make sure we miss nothing and capture everything. THen we can work on making it more suitable to home consumption in post.

Thor

 

[Reynaldo]

@Thorsten Loesch

It is true that a DAC is not just the chip but the whole set that was designed.
But I was able to try several DACs and almost all of them that have the Burr-Brown chip have a different sound. Is it a placebo effect or is there something that really changes?

The impression I have (I could be wrong) that somehow devices like iFi and Musical Fidelity, the engineers and designers were very concerned with the sound that was closer to that of an analog device.
As I wrote earlier, I already had Gustard's DAC and SMSL are excellent, but for my taste it seems that something is missing that gives life to the music.

This is a more technical forum where people look at test numbers, this is important but it seems to me that from a certain level it will not change anything in the sound quality.

Once again, congratulations on your posts.

 

[Killingbeans]

It is tempting and east to jump to the conclusion that the lower distortion is the cause, but given that in all cases the harmonic distortion and noise is well below any audibility limit, it would seem that other, more subtle factors are at work here.

But I was able to try several DACs and almost all of them that have the Burr-Brown chip have a different sound. Is it a placebo effect or is there something that really changes?

If I were to make a bet, I'd put my money on placebo.

 

[Reynaldo]

If I were to make a bet, I'd put my money on placebo.

It might be.

 

[McFly]

Its satire/in jest. Because they don't market it with all the audiophile buzzwords and bullshit, therefore there's no magic for audiophiles to feel all warm and fuzzy.

But in reality, its perfect for everyone.

 

[RonSanderson]

I apologize for not reading the previous 4,630 posts, but…

IMHO, if we are missing any measurements they are on the human side, not so likely in the electronics.

Mathematical treatments of waveforms have leaned heavily on Fourier transforms, for example. The math can break down a complex waveform into a base frequency plus a stack of harmonics. Some of these exceed the upper threshold of hearing - based on listening to a straight isolated tone at that harmonic frequency. But, does the ability to hear an isolated tone, as the math might imply, really correlate to the way we hear? Or can we respond to the sharpness of the wavefront instead, and still be affected by the overall shape of the pressure wave? Putting it another way, does scraping off frequencies we can’t hear rob the music of its punch? (Not a fan of Fourier transforms!)

Another example is in the low frequencies. I have Fried Studio V speakers. Bud Fried (using the research of others) relied on psychoacoustic research which reported that directional perception of low frequencies is governed not by relative loudness from each speaker, but by a travelling wave that moved from one speaker to the other. (I have a link to the patents somewhere...) The design uses a specially wound voice coil that feeds an L-R or R-L signal to each channel to create this traveling wave. (See the link for tech details.)

Are we measuring the wrong thing in the lab, if the way we hear is not the way we measure?

• Should we be measuring rise time instead of frequency response?
• Should we be measuring the propagation of bass wavefronts instead assuming that bass is non-directional?

I still have my doubts about both ends of the frequency spectrum. I have not seen any analysis that makes me think I am wrong - and I have been watching since 1968!

Regards,
Ron

 

[Killingbeans]

Or can we respond to the sharpness of the wavefront instead, and still be affected by the overall shape of the pressure wave? Putting it another way, does scraping off frequencies we can’t hear rob the music of its punch? (Not a fan of Fourier transforms!)

Nope. Your hearing is also bandwidth limited. You couldn't hear a perfect square wave if such a thing existed.

 

[Blumlein 88]

I apologize for not reading the previous 4,630 posts, but…

IMHO, if we are missing any measurements they are on the human side, not so likely in the electronics.

Mathematical treatments of waveforms have leaned heavily on Fourier transforms, for example. The math can break down a complex waveform into a base frequency plus a stack of harmonics. Some of these exceed the upper threshold of hearing - based on listening to a straight isolated tone at that harmonic frequency. But, does the ability to hear an isolated tone, as the math might imply, really correlate to the way we hear? Or can we respond to the sharpness of the wavefront instead, and still be affected by the overall shape of the pressure wave? Putting it another way, does scraping off frequencies we can’t hear rob the music of its punch? (Not a fan of Fourier transforms!)

Another example is in the low frequencies. I have Fried Studio V speakers. Bud Fried (using the research of others) relied on psychoacoustic research which reported that directional perception of low frequencies is governed not by relative loudness from each speaker, but by a travelling wave that moved from one speaker to the other. (I have a link to the patents somewhere...) The design uses a specially wound voice coil that feeds an L-R or R-L signal to each channel to create this traveling wave. (See the link for tech details.)

Are we measuring the wrong thing in the lab, if the way we hear is not the way we measure?

• Should we be measuring rise time instead of frequency response?
• Should we be measuring the propagation of bass wavefronts instead assuming that bass is non-directional?

I still have my doubts about both ends of the frequency spectrum. I have not seen any analysis that makes me think I am wrong - and I have been watching since 1968!

Regards,
Ron

I don't know about traveling waves. I know Fried liked transmission lines. Some info in the link is incorrect. We hear direction for frequencies below 800 hz by timing differences. Above 1500 hz by intensity and a mix between those two. Around and below 80 hz we seem to lose the ability to hear direction via timing. I've heard demonstrations where a very low frequency had phase varied between the two channels and you hear it move. I don't know if you are hearing the position of overtones or not however. And what is happening is comb filtering between the two sources at a significant distance from you. Of course back in earlier days with LP those frequencies are all mono anyway.

 

[RonSanderson]

Nope. Your hearing is also bandwidth limited. You couldn't hear a perfect square wave if such a thing existed.

No, not perfect. But, how close?

The question is, is the upper limit of hearing a steady tone exactly the same as the limit of detecting a rising wavefront. I recently saw a reference to the ear detecting short impulses that are steeper and briefer than a steady-tone analysis would predict. (Sorry, lost that link.)

A Fourier analysis says, breaking apart an sharp impulse, that no leading wavefront steeper than a 20 kHz wave should be detectable. Have we really proven this? Math may say one thing, but the ear is a complex system and it’s a lot more complicated than a speaker. There may be multiple complementary mechanisms at play. Just possibly the math is not relevant to modeling the full function of an ear.

 

[RonSanderson]

… Of course back in earlier days with LP those frequencies are all mono anyway.

Regrettable about the bass, since I have just dusted off my record collection. Where I have a digital copy I’ll be using that so I don’t have compromised bass.

Still digesting your other thoughtful comments.

 

[Killingbeans]

The question is, is the upper limit of hearing a steady tone exactly the same as the limit of detecting a rising wavefront. I recently saw a reference to the ear detecting short impulses that are steeper and briefer than a steady-tone analysis would predict. (Sorry, lost that link.)

That's interesting.

How loud were the impulses? It opens another question about whether or not comparable impulses can be found in music, and whether or not the brain recognizes them as critical information.

 

[antcollinet]

Nope. Your hearing is also bandwidth limited. You couldn't hear a perfect square wave if such a thing existed.

Or to put it another way - your hearing "scrapes off the frequencies you can't hear" in the same way an electronic filter does.

 

[Holmz]

Whole post was good…

But the key point:

If a DAC (Device) accepts direct input via USB of a 32 Bit audio sample without requiring the source device to convert the 32 Bit Sample to 24 Bit, is that a 32 Bit DAC or not? If not, then what is? I can only accept the "194dB+ Dynamic Range" as alternative definition.

To me it is just one more number fetish. How many bit's are actually in a given music recording.
…

In my mind… a DAC using a 24bit DAC chip, is not a 32 bit DAC.

I can do a bunch of 64 or 128 bit math on the computer and down sample it to 16 bits.
But if I call that a 128 bit it seems like it is dishonest and I risk Santa not coming down the chimney this year.

In this case, the “Bonor de Jur”, includes 8 imaginary bits that go between the true 24 bits and the imaginary 32 bits…

And it potentially gets even worse, as if the 24 bits are assumed to be in the upper end of the 32 bit register, and the sending unit had them at the bottom, then we may end up with 16 bits.

 

[RonSanderson]

I don't know about traveling waves. I know Fried liked transmission lines. Some info in the link is incorrect. We hear direction for frequencies below 800 hz by timing differences. Above 1500 hz by intensity and a mix between those two. Around and below 80 hz we seem to lose the ability to hear direction via timing. I've heard demonstrations where a very low frequency had phase varied between the two channels and you hear it move. I don't know if you are hearing the position of overtones or not however. And what is happening is comb filtering between the two sources at a significant distance from you. Of course back in earlier days with LP those frequencies are all mono anyway.

The system implemented in the Fried Studio V was known as m.a.r.s., for McShane Ambient Recovery System. The patent is listed at https://ppubs.uspto.gov/pubwebapp/static/pages/ppubsbasic.html. The PDF is complete with circuit diagrams for those who want to try in the lab.

US-4847904-A

Preview PDF

Ambient imaging loudspeaker system

McShane; Charles L.

1989-07-11

9

The design is meant to address the very relationships between loudness and phase that you cite above.

In keeping with the purpose of this thread, if a sound source does not stimulate the same mechanisms that interrelate loudness, phase, and frequency, could it really be said to be an accurate transducer? Perhaps technically accurate, but is it accurate at recreating the sound field that the listener needs?

This seems like something we can, indeed measure. Can we say a system has accurate bass if it does not embody these effects? Or can a digital room correction be correct if these wavefronts are not recreated so they match the original psychoacoustic stimulus?

 

[RayDunzl]

I recently saw a reference to the ear detecting short impulses that are steeper and briefer than a steady-tone analysis would predict.

You can easily hear the result of your system being fed a single non-silent sample.

"Click".

Looks like this here at my place when picked up by a microphone and recorded:

Turn that data into a wave - up/down/up/down repeat - and I sure wouldn't hear a tone. Too high.

 

[RonSanderson]

Or to put it another way - your hearing "scrapes of the frequencies you can't hear" in the same way an electronic filter does.

I understand what you are saying, but let’s pose a scenario.

The eardrum is a physical system, perhaps like a drumhead drawn tight. Assume it is hit by a 22kHz standing pressure variation. It may be too stiff, or have too much back pressure, to transmit much (or any) of that wave.

Does that mean it is impossible for a steep wavefront to provide a transient acceleration in that range?

I don’t think we can simply perform a Fourier transform and state that components above 18kHz, or 20kHz, are simply discarded. Not until we have independently mapped both transient response and steady-state response of the ear, can we decide whether this math describes the physical system.

And if it doesn’t, the steady state numbers we measure are the wrong numbers. And assumptions we make about the Nyquist cutoff point are also wrong, and the design of a lot of digital reproduction is based on a false model of hearing, and are wrong as well.

I kind of think the design of the system starts with understanding the ear and nervous system, not with the electronics.
Then, we can measure meaningfully.

 

[tomelex]

We record sound with transducers and electronics and we measure electronics with electronics. Basically, no matter how you look at it, we know completely the electronics and transducer side of things, nothing here is magical. We are slaves to our electronics and transducers no matter what else is "possibly" going on.

 

[RonSanderson]

You can easily hear the result of your system being fed a single non-silent sample.

"Click".

Looks like this here at my place when picked up by a microphone and recorded:

Turn that data into a wave - up/down/up/down repeat - and I sure wouldn't hear a tone. Too high.

So cool! Here we have something we can hear, but which we do not measure. But we should, and should understand why.

What would you estimate the sharp rise would correspond to in kHz? (I would do the math but I don’t know what the axes are!)

 

[RayDunzl]

So cool! Here we have something we can hear, but which we do not measure. But we should, and should understand why.

We do measure it.

It is Impulse Response.

Oddly enough, with a log swept sine tone - REW's "measurement" tone is an example. The tone sweeps from low to high frequency with a precise mathematical construction.

The recording above is of an actual "impulse".

Immediately afterwards, I performed the measurement test with REW's tone sweep and got this for it's "impulse response" measurement.

The calculation and the recorded waveforms are all but identical - close enough that I could put the error to different ambient noise during the two sessions.


See

Impulse Response

Presented for your amusement: Single Full Scale Bit sent through speakers, playback recorded in Audacity: Impulse response calculated from a ten second 10-24kHz sweep tone (three weeks ago) in REW: Well, I am amused.

www.audiosciencereview.com

and for extra credit

Impulse Response

Presented for your amusement: Single Full Scale Bit sent through speakers, playback recorded in Audacity: Impulse response calculated from a ten second 10-24kHz sweep tone (three weeks ago) in REW: Well, I am amused.

www.audiosciencereview.com

Somehow I didn't believe the mathematicians could pull "edges" - like impulse and step response - out of a swept sine wave - which has no edges in it.

But they can.

It's above my pay scale, of course.