DAC Measurements

[GMStudio]

Sorry if this technical question has already been answered elsewhere but maybe someone can help my understanding?

Let’s say we have a DAC with a dynamic range of say 124dB, which means that -125dB and lower all we’ve got is noise (probably thermal noise). How then can we measure for example jitter noise/artefacts at say -140dB (or lower in some cases) at the DAC’s output? Wouldn’t it be buried 15dB below the DAC’s noise floor?

TIA.

 

[HarmonicTHD]

Sorry if this technical question has already been answered elsewhere but maybe someone can help my understanding?

Let’s say we have a DAC with a dynamic range of say 124dB, which means that -125dB and lower all we’ve got is noise (probably thermal noise). How then can we measure for example jitter noise/artefacts at say -140dB (or lower in some cases) at the DAC’s output? Wouldn’t it be buried 15dB below the DAC’s noise floor?

TIA.

I would say yes. And at these levels everything is inaudible anyway so it wouldn’t even matter.

Audibility thresholds of amp and DAC measurements

Introduction A recurring theme on ASR is whether or not the various measured qualities of the devices are audible. In this post, I'll present some clear and visual thresholds for when those imperfections can be considered a potentially audible concern. I will not explain the basics of amp/DAC...

www.audiosciencereview.com

 

[GMStudio]

I would say yes. And at these levels everything is inaudible anyway so it wouldn’t even matter.

Thanks for your response but it doesn’t really help. I realise it’s way below audibility but I’m not considering audibility. The highest dynamic range DACs have about 126dB or so but we fairly often see jitter noise/artefact measurements substantially below that. I’m trying to understand how it’s possible to measure say -140dB artefacts at the DAC’s output. Wouldn’t we just see the DAC’s noise floor extending below -126dB (or higher), rather than the spectrum of jitter noise/artefacts many dB lower?

G

 

[NTK]

When reporting the signal to nose ratio, the power of the noise is the total power of the noise at all frequency bins (of the reporting measurement bandwidth). At each specific frequency bin, the noise power is a lot lower. Using Amir's usual measurement settings, the "noise floor" (i.e. the "grass") is usually down ~30-35 dB from the reported signal to noise ratio (this number is called the FFT gain). Therefore, you can often detect/measure spurs that are quite a bit below the reported S/N ratio. Actually, one way to increase sensitivity is to use longer FFT length, which will spread out the noise into more frequency bins and depress the heights of the grass.

 

[staticV3]

Here's one example:
The DAC I'm measuring (9038D) has a DR/SNR of around 122dB(unweighted), but using the Fourier transform I can examine distortion down to -160dB relative to the fundamental.

Basically, by dividing the signal into more and more frequency bins, the tones inside those bins keep their amplitude, but the noise decreases.


You can learn more about the FFT in these videos:

Spoiler: Videos

 

[GMStudio]

When reporting the signal to nose ratio, the power of the noise is the total power of the noise at all frequency bins (of the reporting measurement bandwidth). ….

Thanks that does indeed help!

A different but related question: Can these often very low measurements affect sound quality? Using the example in the OP, of say jitter noise/artefacts/spurs peaking at -140dB, can this have any effect on the output of the transducers (speakers or headphones)? Or will it be completely un-reproducible, lost/buried in the noise floor of the amp+transducers, even if we consider the best amp/transducers?

Again, I’m not talking about audibility, I realise we’re well below that, just about what would/could be in the actual sound waves produced.

TIA

 

[Sgt. Ear Ache]

Things that are inaudible can't affect "sound quality." Inaudible is inaudible. You ain't hearing it. It's not changing sound quality. If it's changing sound quality, then by definition it's audible.

 

[HarmonicTHD]

Thanks that does indeed help!

A different but related question: Can these often very low measurements affect sound quality? Using the example in the OP, of say jitter noise/artefacts/spurs peaking at -140dB, can this have any effect on the output of the transducers (speakers or headphones)? Or will it be completely un-reproducible, lost/buried in the noise floor of the amp+transducers, even if we consider the best amp/transducers?

Again, I’m not talking about audibility, I realise we’re well below that, just about what would/could be in the actual sound waves produced.

TIA

How do you define (perfect) sound quality?

I was always under the impression that if some distortion/noise is not audible, than sound quality is perfect?

Or do you mean that once you send the eg -140dB noise through, let’s say 20 to 30dB amplification (gain), it might become audible ? (I would say no, see research on audibility).

Maybe I misunderstood your question (as before ).

 

[staticV3]

There's a difference between perceivable sound quality and measurable sound quality. Both are valid definitions of sound quality imo.
Declaring that something is perfect just because our sensory organs are not able to perceive any faults is nearsighted. Imagine how far JWST would've gotten if our ears, eyes, and hands were the only measure of quality.

Going back to OP's question:
Audio analyzers can only measure what is actually part of the signal. Even if jitter or distortion is below the noise floor of the DUT, that doesn't mean that they don't exist.
Obviously that doesn't mean that they are audible or even worth paying attention to.

 

[GMStudio]

Things that are inaudible can't affect "sound quality."

Sure they can. For example we can have a component in a sound wave that’s say at 22kHz and 40dBSPL, this would be inaudible but it would still be part of the sound output, assuming the speakers/HPs can reproduce those freqs of course. There are numerous other examples/conditions of actual sounds or components within a sound wave which exist but are inaudible.

Maybe I misunderstood your question (as before ).

Yes, although it’s entirely possible you misunderstand because I didn’t explain it well

Maybe it’s easier to understand if I drop the word “quality” from the question. IE. Is it possible that the sound can be affected? Can artefacts/distortion (at say -140dB) at the output of a DAC, actually be transduced into sound by even the best amp+speakers/HPs? For example, instead of the DACs output from a J test being fed into an AP55, it was fed into top of the line amp+speakers/HPs, would the DACs noise/distortions (at -140dB) have any effect whatsoever on the sound wave produced or would it simply not exist because it is so far below the noise floor of the amp+speakers/HPs?

 

[Sgt. Ear Ache]

Sure they can. For example we can have a component in a sound wave that’s say at 22kHz and 40dBSPL, this would be inaudible but it would still be part of the sound output, assuming the speakers/HPs can reproduce those freqs of course. There are numerous other examples/conditions of actual sounds or components within a sound wave which exist but are inaudible.

Yes, although it’s entirely possible you misunderstand because I didn’t explain it well

Maybe it’s easier to understand if I drop the word “quality” from the question. IE. Is it possible that the sound can be affected? Can artefacts/distortion (at say -140dB) at the output of a DAC, actually be transduced into sound by even the best amp+speakers/HPs? For example, instead of the DACs output from a J test being fed into an AP55, it was fed into top of the line amp+speakers/HPs, would the DACs noise/distortions (at -140dB) have any effect whatsoever on the sound wave produced or would it simply not exist because it is so far below the noise floor of the amp+speakers/HPs?

It's part of the measurable sound "output" but it's not part of the sound "quality." Assuming by sound quality you mean something that we can listen to.

 

[staticV3]

would the DACs noise/distortions (at -140dB) have any effect whatsoever on the sound wave produced

No. Speakers are not transparent enough to resolve such small details.

 

[BDWoody]

would the DACs noise/distortions (at -140dB) have any effect whatsoever on the sound wave produced or would it simply not exist because it is so far below the noise floor of the amp+speakers/HPs?

It would exist in the electrical signal sent to the amp as 1:10,000,000 of the overall signal. Not sure what amp or transducer wouldn't bury that completely.

 

[Sgt. Ear Ache]

There's a difference between perceivable sound quality and measurable sound quality. Both are valid definitions of sound quality imo.
Declaring that something is perfect just because our sensory organs are not able to perceive any faults is nearsighted. Imagine how far JWST would've gotten if our ears, eyes, and hands were the only measure of quality.

Going back to OP's question:
Audio analyzers can only measure what is actually part of the signal. Even if jitter or distortion is below the noise floor of the DUT, that doesn't mean that they don't exist.
Obviously that doesn't mean that they are audible or even worth paying attention to.

The definition of perfection is relative to what we are trying to accomplish. Something that can't be heard by our ears and doesn't impact audible sound quality is for all intents and purposes irrelevant as far as listening to music is concerned.

 

[staticV3]

Assuming by sound quality you mean something that we can listen to.

No part of "quality" implies "perceivability". All it means is a characteristic or feature.

 

[Sgt. Ear Ache]

No part of "quality" implies "perceivability". All it means is a characteristic or feature.

I disagree. I think when we're talking about hifi and musical sound reproduction, the term "sound quality" pretty much always refers to perceivable characteristics - or what people think might be perceivable at least.

 

[staticV3]

Something that can't be heard by our ears and doesn't impact audible sound quality is for all intents and purposes irrelevant as far as listening to music is concerned.

From the standpoint of a consumer I 100% agree.
From the standpoint of a designer or evaluator however, doing things by ear is not the way to go.

 

[Sgt. Ear Ache]

From the standpoint of a consumer I 100% agree.
From the standpoint of a designer or evaluator however, doing things by ear is not the way to go.

Of course not. But when someone asks the question "can these very low measurements affect the sound quality?" what is the meaning? Can they affect the "measurable SQ?" Well yes, obviously...since we're talking about things that have already been measured. The only other way to interpret that question is "can these things that measure below audible levels impact SQ in some other audible way?" I mean if the question is "can these low measurements affect SQ in some other inaudible way?" then the answer really is...who cares?

 

[staticV3]

Going back to OP's question:

Can these often very low measurements affect sound quality? Using the example in the OP, of say jitter noise/artefacts/spurs peaking at -140dB, can this have any effect on the output of the transducers (speakers or headphones)? Or will it be completely un-reproducible, lost/buried in the noise floor of the amp+transducers, even if we consider the best amp/transducers?

I understood it as being of technical nature (can these artifacts be reproduced by a transducer, or will they be filtered/masked out?), not of practical nature (should I base my purchasing decisions on these artifacts?).

Imo, interpreting and assuming meaning behind words, beyond their textbook definition, breeds misunderstandings and unnecessary confrontation.

 

[Sgt. Ear Ache]

Going back to OP's question:

I understood it as being of technical nature (can these artifacts be reproduced by a transducer, or will they be filtered/masked out?), not of practical nature (should I base my purchasing decisions on these artifacts?).

Imo, interpreting and assuming meaning behind words, beyond their textbook definition, breeds misunderstandings and unnecessary confrontation.

I'm sorry, but I'm still not sure what the OP was asking. What does "artifacts being reproduced by a transducer" mean if it doesn't mean something audible? Or at least potentially audible...

It sort of seems to me that this is an attempt to find a way to account for "perceived" differences between dacs that seem to measure the same in any measurably-audible sense. "Well even though the jitter measures well below audibility, maybe it affects something that affects something else that then is audible?"