• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

What are all the variables that impact DAC quality?

Joined
Oct 31, 2020
Messages
27
Likes
15
I was having a conversation in a YouTube comment section and someone said that while a DAC measures excellently in the areas of noise and distortion, that doesn't tell the whole story because these measurements don't take ''transient reconstruction'' into account, and under another video a guy said that Chord dacs with their short digital filters reproduce the timing of music better than other dacs.

Is there some truth to this or is it all just snake oil?
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,204
I was having a conversation in a YouTube comment section and someone said that while a DAC measures excellently in the areas of noise and distortion, that doesn't tell the whole story because these measurements don't take ''transient reconstruction'' into account, and under another video a guy said that Chord dacs with their short digital filters reproduce the timing of music better than other dacs.

Is there some truth to this or is it all just snake oil?

If phase and amplitude are flat* in the audio band (i.e. <20kHz), nothing that happens beyond the audio band (i.e. >20kHz) will have the slightest audible effect on the "transient" performance of the DAC.

*Flat enough that any deviations are beyond human audibility, which is the case for any decent DAC.
 

Sgt. Ear Ache

Major Contributor
Joined
Jun 18, 2019
Messages
1,445
Likes
3,075
Location
Winnipeg Canada
I was having a conversation in a YouTube comment section and someone said that while a DAC measures excellently in the areas of noise and distortion, that doesn't tell the whole story because these measurements don't take ''transient reconstruction'' into account, and under another video a guy said that Chord dacs with their short digital filters reproduce the timing of music better than other dacs.

Is there some truth to this or is it all just snake oil?

They failed to note the critical element that truly distinguishes one dac from another - unobtainium! It's the musical magic pixie dust that never shows up in the measurements! :D
 

Jimbob54

Master Contributor
Forum Donor
Joined
Oct 25, 2019
Messages
8,332
Likes
10,121
With regard to the headline question- I think you need to define what you mean by "quality"- because I know where this thread is going if you mean "sound quality".
 

Wes

Major Contributor
Forum Donor
Joined
Dec 5, 2019
Messages
3,851
Likes
3,669
ask the guy what transient reconstruction is

But it may very well be that distortion plus noise does not fully capture all factors that can affect SQ. One issue is that different types of distortion may have different psychoacoustic weightings (in your mind). AFAIK, this has not been well explored, nor would it be trivial to do so.

Listening tests are the gold std. (and by tests, I do not mean comparisons alone)
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
14,316
Likes
23,066
They failed to note the critical element that truly distinguishes one dac from another - unobtainium! It's the musical magic pixie dust that never shows up in the measurements! :D
Is that how they implement the unmeasurablium into the sound output?
 
OP
J
Joined
Oct 31, 2020
Messages
27
Likes
15
I've been thinking about this some more and came to the conclusion that perhaps the testing methods Amir uses really are insufficient. It appears plausible to me that DAC speed could be a very real thing, i.e., that smearing could occur during fast transitions in the input signal. Same applies to amps.

https://www.picotech.com/library/oscilloscopes/rise-time

This could potentially go unnoticed when measuring the DAC/amp the way Amir does, because, please correct me if I'm wrong, the test tones he uses for the measurements do not contain any fast transitions. For instance, he doesn't test the gear with things like cymbal hits, etc.

Couldn't a DAC like the Chord Dave be faster than the best DACs Amir measured, in the exact same way an electrostatic headphone is faster than say a dynamic one?
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,204
I've been thinking about this some more and came to the conclusion that perhaps the testing methods Amir uses really are insufficient. It appears plausible to me that DAC speed could be a very real thing, i.e., that smearing could occur during fast transitions in the input signal. Same applies to amps.

https://www.picotech.com/library/oscilloscopes/rise-time

This could potentially go unnoticed when measuring the DAC/amp the way Amir does, because, please correct me if I'm wrong, the test tones he uses for the measurements do not contain any fast transitions. For instance, he doesn't test the gear with things like cymbal hits, etc.

Couldn't a DAC like the Chord Dave be faster than the best DACs Amir measured, in the exact same way an electrostatic headphone is faster than say a dynamic one?

Time and frequency are the same thing. If a DAC can reproduce 20kHz at full output, then it is fast enough to reproduce all audible transients. (20kHz, after all, is 20,000 cycles per second.)

If a cymbal hit has a risetime faster than 20,000 cycles per second (which is absolutely plausible, as you say), all that means is that it produces some frequencies that humans can't hear. If we can't hear these frequencies, we don't lose anything when they are absent. The cymbal sounds the same whether or not those frequencies are captured/reproduced (and btw, they are rarely even captured in the first place).
 

Wes

Major Contributor
Forum Donor
Joined
Dec 5, 2019
Messages
3,851
Likes
3,669
The problem with the above stmt. is the psychoacoustic assumption involved.

Human hearing has been traditionally tested with sine waves. The leading edge of a transient - in association with other music - might be more (or less detectable) than a single sine wave at 20 kHz.

The proof of the pudding would be to test two different DACs side by side - in an ultra-quality system, using trained teenage musical prodigies - violinists maybe. If they cannot hear the difference, then highly likely you cannot.

That said... a speaker or room tmts. upgrade is highly likely to be more effective at improving SQ.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,204
The problem with the above stmt. is the psychoacoustic assumption involved.

Human hearing has been traditionally tested with sine waves. The leading edge of a transient - in association with other music - might be more (or less detectable) than a single sine wave at 20 kHz.

The proof of the pudding would be to test two different DACs side by side - in an ultra-quality system, using trained teenage musical prodigies - violinists maybe. If they cannot hear the difference, then highly likely you cannot.

That said... a speaker or room tmts. upgrade is highly likely to be more effective at improving SQ.

You could take a recording that includes transients and that contains meaningful musical content above 20kHz and then high-pass filter it so that only content above 20kHz remains. If, when you play back that high-pass filtered content, nothing is audible, then I think this would put an end to that line of argument.
 

Wes

Major Contributor
Forum Donor
Joined
Dec 5, 2019
Messages
3,851
Likes
3,669
not to the issue of whether effects obtain when in combination

not that it wouldn't be a worthwhile grad. student thesis

I think a direct DAC comparison would be useful - a prospective buyer only care about whether they can hear a difference and input switching is easy. I dunno if a Chord has a return privilege or can be borrowed.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,846
Likes
13,197
Location
Riverview FL
I've been thinking about this some more and came to the conclusion that perhaps the testing methods Amir uses really are insufficient. It appears plausible to me that DAC speed could be a very real thing, i.e., that smearing could occur during fast transitions in the input signal. Same applies to amps.


Ok:



Analyze my little experiment as you like.



1. Audio Track. One full-scale bit. It doesn't get any "faster" - it is an "illegal" signal, though...
2. DAC output (RCA) - uses Focusrite ADC for capture
3. Amplifier output (differential) - sure does look like the DAC waveform - uses Focusrite ADC for capture
4. Martin Logan electrostat - using the amplifier - using UMIK-1 ADC - measured at listening position
5. JBL LSR 308 - using its own ADC and amplifier - then using UMIK-1 ADC - measured at listening position

Recorded traces amplified in REW as required to match amplitude.

Dots are individual sample values.

I notice the "ringing" at the DAC/Amp is not reflected in the speaker output.

Trace for JBL is inverted at REW for comparison.


1605476098528.png
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,204
not to the issue of whether effects obtain when in combination

not that it wouldn't be a worthwhile grad. student thesis

I think a direct DAC comparison would be useful - a prospective buyer only care about whether they can hear a difference and input switching is easy. I dunno if a Chord has a return privilege or can be borrowed.

If that's what you're testing for, then IMO testing two DACs side by side would introduce too many uncontrolled variables. Why not just test two signals played back at high sampling rates on the same DAC, one more band-limited than the other?
 
OP
J
Joined
Oct 31, 2020
Messages
27
Likes
15
Time and frequency are the same thing. If a DAC can reproduce 20kHz at full output, then it is fast enough to reproduce all audible transients. (20kHz, after all, is 20,000 cycles per second.)

If a cymbal hit has a risetime faster than 20,000 cycles per second (which is absolutely plausible, as you say), all that means is that it produces some frequencies that humans can't hear. If we can't hear these frequencies, we don't lose anything when they are absent. The cymbal sounds the same whether or not those frequencies are captured/reproduced (and btw, they are rarely even captured in the first place).

You appear to be correct. The ''fast transition'' hypothesis states that when you go from say 20 Hz o 20 kHz, some DACs might struggle with the transition, which doesn't make much sense if you can reproduce 20 kHz continuously. If you can reproduce 20 kHz, it shouldn't matter ''how'' that 20kHz is arrived at.

It's not like a car where accelerating from 20 to 200 mph is obviously different than accelerating from 190 mph to 200 mph. The first clearly takes a lot more time than the second.

With frequency, it's different. If you want to reproduce 20 000 Hz, you don't get any ''help'' from starting at say 19 000 Hz. It's the same thing as starting at 20 Hz, or dead silence for that matter.

You can easily visualize this.

Imagine a graph of a continuous 20 000 Hz sine wave. Pick any point on this graph where the wave intersects the X axis and mark it.

Next, imagine a second graph, of a wave which changes frequency from say 5 Hz to 20 000 Hz at some point which lies on the X axis. Mark this point also.

Now compare the portions of the two graphs beginning at the points you marked. Both are identical - constant 20 000 kHz, starting at zero. What came before doesn't matter in the least.

In other words.

Case 1. Take a car and accelerate it to 200 mph and brake to a standstill a bunch of times, and then, from standstill, accelerate to 200 mph again.

Case 2. Take a car and accelerate it to 5 mph and brake to a standstill a bunch of times. Now accelerate to 200 mph from standstill.

Do you think the result of the final acceleration from standstill to 200 mph will be any different between the two cases?

My illustration should make this easier to understand, if you can forgive my terrible handwriting skills.
 

Attachments

  • IMG_1197.jpg
    IMG_1197.jpg
    425.3 KB · Views: 55
Last edited:
Top Bottom