Measurement of R2R DAC

[DonH56]

What I think all your DAC tests fail to take into account is the great variation in SQ between the I-V and analog output stages of various DACs. If you think a quick and cheap op-amp will sound as good as say a discrete output stage in an Ayre or Berkeley DAC, well then...

I recognize and accept that many here at ASR believe that your distortion and jitter measurements tell the entire story audio quality. Even setting aside high-end audio manufacturers, I don't think that pro-sound/studio equipment firms would agree.

But anyway, have a great weekend all. We are getting our first low-elevation snow of the year here in the Sierra Foothills. Quite lovely!

Good point. Leaving aside the op-amp comment (my answer would be "it depends"), I agree that the output buffer and filter stages are likely a significant part of the output, but disagree the tests don't account for them. Amir measures at the analog output so they are most certainly accounted for in his tests. What we do not know, and have no way of telling without ripping the boxes apart and running more tests, is how much noise and distortion comes from the output buffers etc. vs. how much is from the DAC itself. For consumers, it's a don't-care, because as consumers we are buying a DAC-in-box and cannot isolate them anyway. How well they perform also depends upon the load, natch, but the line output load is pretty consistent for most audio components. And I've said a number of times that I suspect those analog output stages are what differentiate a lot of DACs, how they respond to glitches and HF noise, etc.

What might be interesting is to set up a few nominal test loads and see if there is any difference in measured performance, e.g. maybe 10k and 100k R's with a couple of different shunt C's (inductance is probably not a big player). Additional IMD and multitone testing would also be neat, but I assume Amir has to sleep sometime. Maybe we need to get the King more coffee?

I think a review of threads would show most of us don't think jitter and distortion are the whole story, but they are a big part of it, and provide a basis for comparison. Why buy a DAC that measures significantly worse on the bench? The bench environment is usually cleaner/better than the real world, so...

Snow, eh, we've had some, and could use more, but I'd be glad to hand you the sub-zero nights and highs in the teens and twenties we've had lately. At least I can broom off a few inches when it's that cold instead of breaking my back shoveling.

 

[DonH56]

Hi Amir,

Good to hear that you're getting the RME. I'm waiting for your measurements before deciding on an RME ADI-2 DAC or Soekris 1541. I have no idea what to choose!

Buy both and send me the one you don't want.

 

[Don Hills]

... Maybe we need to get the King more coffee? ...

Nope. Too much coffee risks his jitter exceeding that of the devices being measured.

 

[Rthomas]

Buy both and send me the one you don't want.

I would if I didn't live in the UK!

 

[March Audio]

What I think all your DAC tests fail to take into account is the great variation in SQ between the I-V and analog output stages of various DACs. If you think a quick and cheap op-amp will sound as good as say a discrete output stage in an Ayre or Berkeley DAC, well then...
!

What do you base this statement on? Sighted uncontrolled comparison?

 

[March Audio]

I think a review of threads would show most of us don't think jitter and distortion are the whole story, but they are a big part of it, and provide a basis for comparison. Why buy a DAC that measures significantly worse on the bench? The bench environment is usually cleaner/better than the real world, so...

Clean well measuring dacs are unlikely to stand out as having a different sound. Audiophiles like "personality"

 

[Superdad]

What do you base this statement on? Sighted uncontrolled comparison?

50 years of listening to live acoustic music performances, being involved in audio equipment design for 40 years, auditions in my custom-built listening studio, and my ears and monkeybone able to know what sounds more like real music. Despite your insistence, it is really not that difficult.

And yes, I have done plenty of unsighted comparisons. At Hovland, we would often be deciding on a small part change in an amplifier--say for a bias level or feedback scheme--and our engineer would bring out a stereo amp where one channel had the change. Would not tell the rest of us which channel was which. We would hook up one speaker listen (single speaker is a great method anyway when doing amp design), then hook up the same speaker to the other channel and listen again. Would go back and forth a couple of times, then argue a bit between us and make a decision. Have done similarly hundreds if not thousands of times over the decades.
One could measure both channels of the above exampled amp, and even if some small differences were spotted they would not correlate at all with how big a difference was heard.

I have asked you a dozen times: Go to an audio show and ask 50 engineers--at firms you think well of and know they have the measurement gear--if they make critical decisions about component design in the studio as well as on the bench. As with race cars, you can model and measure on the computer all year long, but in the end a skilled driver is needed for actual performance feedback.

 

[Frank Dernie]

I would if I didn't live in the UK!

I bought the RME because it has a lot of adjustments to play with too. it has an ADC too, which I haven't tried - I already have several.

 

[Frank Dernie]

As with race cars, you can model and measure on the computer all year long, but in the end a skilled driver is needed for actual performance feedback.

Now that was my actual business, making music recordings has been a hobby for 50 years, designing record players before I went full time into Formula 1. Most of the critical parameters of a race car can NOT be modelled on a computer since they are too complex. The things which make most difference are aero and tyre temperature, neither of which have sufficiently accurate mathematical models to calculate the performance to an acceptable level of accuracy. You do need a skilled and consistent driver to do comparisons but it is mainly the stopwatch data which counts.

Edit apologies for drifting so far off topic...

 

[Don Hills]

....You do need a skilled and consistent driver to do comparisons but it is mainly the stopwatch data which counts. ...

I'll bet you had cases where the driver preferred one setup but the stopwatch said a different one was faster.

 

[Frank Dernie]

I'll bet you had cases where the driver preferred one setup but the stopwatch said a different one was faster.

Yes, but never from the best drivers.

 

[March Audio]

50 years of listening to live acoustic music performances, being involved in audio equipment design for 40 years, auditions in my custom-built listening studio, and my ears and monkeybone able to know what sounds more like real music. Despite your insistence, it is really not that difficult.

And yes, I have done plenty of unsighted comparisons. At Hovland, we would often be deciding on a small part change in an amplifier--say for a bias level or feedback scheme--and our engineer would bring out a stereo amp where one channel had the change. Would not tell the rest of us which channel was which. We would hook up one speaker listen (single speaker is a great method anyway when doing amp design), then hook up the same speaker to the other channel and listen again. Would go back and forth a couple of times, then argue a bit between us and make a decision. Have done similarly hundreds if not thousands of times over the decades.
One could measure both channels of the above exampled amp, and even if some small differences were spotted they would not correlate at all with how big a difference was heard.

I have asked you a dozen times: Go to an audio show and ask 50 engineers--at firms you think well of and know they have the measurement gear--if they make critical decisions about component design in the studio as well as on the bench. As with race cars, you can model and measure on the computer all year long, but in the end a skilled driver is needed for actual performance feedback.

None of which means anything. Recordings are not live experiences. They wont sound the same. So basically, as suspected, for the most part your subjective uncontrolled opinions. So your tests at Hovland were also uncontrolled. You talked about the results with each other. I have watched a room full of audiophiles do this and steered them with comments just to see what happened. Quite enlightening and entertaining.

When you say "involved" with audio equipment design, are you saying you are an EE?

However you actually missed the point I was making which was specific to your example not your general world experience.

 

[Wombat]

"Sounded like real music".

Really, compared to what, unreal music?

Generalities get boring after so many years. Can you be more specific or elaborate and add to knowledge?

 

[rebbiputzmaker]

50 years of listening to live acoustic music performances, being involved in audio equipment design for 40 years, auditions in my custom-built listening studio, and my ears and monkeybone able to know what sounds more like real music. Despite your insistence, it is really not that difficult.

And yes, I have done plenty of unsighted comparisons. At Hovland, we would often be deciding on a small part change in an amplifier--say for a bias level or feedback scheme--and our engineer would bring out a stereo amp where one channel had the change. Would not tell the rest of us which channel was which. We would hook up one speaker listen (single speaker is a great method anyway when doing amp design), then hook up the same speaker to the other channel and listen again. Would go back and forth a couple of times, then argue a bit between us and make a decision. Have done similarly hundreds if not thousands of times over the decades.
One could measure both channels of the above exampled amp, and even if some small differences were spotted they would not correlate at all with how big a difference was heard.

I have asked you a dozen times: Go to an audio show and ask 50 engineers--at firms you think well of and know they have the measurement gear--if they make critical decisions about component design in the studio as well as on the bench. As with race cars, you can model and measure on the computer all year long, but in the end a skilled driver is needed for actual performance feedback.

You are speaking of realities, that is not appreciated by some here.

 

[Wombat]

You are speaking of realities, that is not appreciated by some here.

We all wish to know about the realities, Rabbi. Give Superdad your guidance in this regard.

Avoid the "I can hear it', don't doubt me even though I cannot, generally, support my claim approach.

 

[Sal1950]

You are speaking of realities, that is not appreciated by some here.

Only by the few that come here believing in Santa Claus, The Tooth Fairy, Magic Dust, and Sighted Subjective Listening Sessions.
Unfortunately there are still some that just refuse to grow up.

 

[svart-hvitt]

You are speaking of realities, that is not appreciated by some here.

Ceci n'est pas une pipe.

https://en.m.wikipedia.org/wiki/The_Treachery_of_Images

And finally, «The Listening Room» (La Chambre D'Ecoute) by Magritte:

 

[Wombat]

As we say in Australia, "She'll be apples".

https://www.urbandictionary.com/define.php?term=she'll be apples

 

[Thomas savage]

A fine anecdote is appreciated but can never be a means for a meaningful discussion here because there’s no fortifying knowns establishment, there’s no data.

That’s the crux of the issue with “ I know I herd this” .. might be true might not be but we’ve no way of knowing so can’t draw much from it.

No “avoiding reality “ involved in that just a desire to have meaningful debate.

 

[Thomas savage]

You are speaking of realities, that is not appreciated by some here.

You fundamentally misunderstand why some here object to anecdotal experiences being put forward as facts.

That’s your deficiency not our members