And why do you think AD produces noise and distortion way below audibility and DA doesn't? Again, if that is the case, why making the test at all?
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
AES Paper Digest: Do Audio Op-amps Sound Different?
- Thread starter amirm
- Start date
And why do you think AD produces noise and distortion way below audibility and DA doesn't? Again, if that is the case, why making the test at all?
Because when audience cannot tell the difference you get random results, and as they are random you cannot explain them as there are no logic in them.
svart-hvitt
Major Contributor
- Joined
- Aug 31, 2017
- Messages
- 2,375
- Likes
- 1,253
As I said, in theory you have a point. And the test is not flawless.
Welcome to the world. The Matrix seems to have you.
Let's see...
If one DAC output ultrasonic stuff upsetting the speakers/amp that the other didn't, and the ADC didn't collect that frequency range, then maybe testing with files is no good...
If one DAC output ultrasonic stuff upsetting the speakers/amp that the other didn't, and the ADC didn't collect that frequency range, then maybe testing with files is no good...
They have had detection results for individual devices that reach the 95 percent confidence interval. Still a theoretical possibility that it may all be spurious, of course.
svart-hvitt
Major Contributor
- Joined
- Aug 31, 2017
- Messages
- 2,375
- Likes
- 1,253
The test method seems to be able to discover DACs of colour, but vulnerable to letting some coloured DACs pass as neutral.
PS: This is a common trait of many test procedures, cfr. all documentation that most research (or parts of that research) in science journals is wrong (source: do a Google search on this theme).
Last edited:
But just to emphasize it again: They have not found large differences between dacs (on my interpretation). It all seems very very subtle. Heck, even the smart phones they tested (only did measurements no listening tests) received high grades - "good enough for Hifi". What we're talking about here is whether these dacs are transparent to the human ear in an absolute sense, on all kinds of material. IF their test provides valid results, of course.
My ears aren't transparent, so I don't listen for others...
Ok, putting AD conversion aside, you still believe they were able to distinguish DACs which have THD in the range of o.oooX%?
Even if they were using today's best loudspeakers money can buy (which have THD of app 0.3%), you're actually trying to tell us that there are people who can tell the difference between say 0.0008% of Topping D10 and 0.0005% of Topping DX7s by listening to the speakers which have 0.3%? So the difference between DACs of 0.0003% wasn't masked by speaker's 0.3% and somehow got played clearly enough to be heard? So you basically think somebody can hear a difference of THD between -102dB vs -106dB?
Wow.. I really don't know what to say anymore. Except that I highly recommend to people who can hear such difference to stay away from kryptonite.
Just to be contrary, there can be other things than THD that distinguish different DACs. But I suspect differences among most DACs are heard more by eyes and wallets than by ears.
IMO - Don
IMO - Don
Of course, you always have to be careful about "just THD", to name the obvious:
-across what frequency range
-across what output impedance
-at what output levels
-at what temperature range
-at what input drive level
-across what frequency range
-across what output impedance
-at what output levels
-at what temperature range
-at what input drive level
- Joined
- Feb 23, 2016
- Messages
- 20,772
- Likes
- 37,639
Imaging from using 'audiophile' filters at the DAC output could be near audible levels and yet result in a fine THD result.
To be clear, general THD levels of 0.0008 and 0.0005 are obviously not distinguishable. What gave these dacs away were other things. The Sonos connect had some measurable rolloff in the deep bass. So they could detect it by playing very bass heavy songs (probably something nobody would notice unless knowing about it beforehand). With the oppo sonica, it was the ticking/clicking, which implies something with time. My guess is that it may then have something to do with the implementation of the filter, or perhaps something with jitter, as it seems related to the time domain (again, not THD). The how and why here are beyond my level of expertise.
If you wonder about my own dac, I've been using a Sonos connect for the last years, and been happy with that. Any such bass issues are moot IMO as the room anyway becomes the dominant factor for bass perception in a room like mine. So what is the point of tests like these? Well, if I had to buy a new streaming enabled dac, I would then get the WXC-50 over the Oppo Sonica. I doubt I would hear a difference, but given that the WXC-50 is less than half the price of the Oppo, it just seems more rational to get the better performing and cheaper one.
If you wonder about my own dac, I've been using a Sonos connect for the last years, and been happy with that. Any such bass issues are moot IMO as the room anyway becomes the dominant factor for bass perception in a room like mine. So what is the point of tests like these? Well, if I had to buy a new streaming enabled dac, I would then get the WXC-50 over the Oppo Sonica. I doubt I would hear a difference, but given that the WXC-50 is less than half the price of the Oppo, it just seems more rational to get the better performing and cheaper one.
Last edited:
When I say THD I actually mean both, harmonic and non harmonic distortion (like IM). Worth mentioning beside these is of course SNR and crosstalk, but both of these also cannot really be heard unless you're conducting your tests at 130dB sound level at least otherwise they would be too deep below the signal to hear. But at 130dB I doubt anybody would be able to hear any distortion or other artifact as the ears will start to painfully bleed.
I don't know at which output level they've conducted their test but loudspeaker manufacturers declare THD at the 90dB sound level. If you crank speakers beyond they will start to produce significantly higher THD than 0.3% as stated for 90dB, so chances to spot any distortion from DAC will be zero. Than there are of course distortion and noise artifacts coming from preamp/amp combo which sits between DAC and and speakers also doing their job to mask DAC distortions.
Regarding filters, I was assuming the same filter was sett on DAC which ensures linearity up to 20kHz, but that has yet to be confirmed once we get the text with test conditions.
What other things? Jitter adds to the noise. Take a look at Amir's measurement of the $80 Topping D10 DAC and see for yourself at what dB level is jitter. Good luck hearing jitter at -130dB.
Here is an article with independent measurement of OPPO Sonica and I don't see and rolloff in the deep bass.
https://hometheaterhifi.com/reviews/dac/oppo-digital-sonica-dac-review/
Last edited:
What other things? Jitter adds to the noise. Take a look at the Amir's measurement of the $80 Topping D10 DAC and see for yourself at what dB level is jitter .
Here is an article with independent measurement of OPPO Sonica and I don't see and rolloff in the deep bass.
https://hometheaterhifi.com/reviews/dac/oppo-digital-sonica-dac-review/
You mixed up the Sonos and the oppo. It was the Sonos connect that had bass rolloff according to their measurements, not the Oppo. With the oppo - I am actually not sure what the proper explanation for the clicking anomaly is. I will send a pm to dr Granqvist on the Swedish forum and ask him.
Ok. And how about you finally post their measurements? It's hard for the rest of us to talk about something only you have seen..
Yeah fully agree. Still searching for the back issues... I'm in the process of moving, so temporarily residing at my parents house, which means that things are not at their usual place. If I can't find them (or my mom has thrown them out) I'll try to get hold of them directly from one of the Swedes. After opening up this can or worms I feel responsible for not leaving you guys hanging!
Last edited:
Jakob1863
Addicted to Fun and Learning
@svart-hvitt ,
Exactly.
This experiment like (most probably) every other isn´t flawless (for some reasons humans do not tend to be perfect) and Krunok indeed has a point which has to be considered when trying to draw conclusion from the results.
But his argument isn´t restricted to the use of the ADC but could be expanded to every part of the reproduction chain overall. You need other gear to transform signals into sound and none of this is working in a perfect manner (even if it is considered to be "transparent" ) .
That it is of utmost importance to state the objective of any test as clear as possible, can´t be emphasized enough imo.
@oivavoi ,
As usual, replications are needed to find some more corroboration for the results. I haven´t read the description closely yet, but at a first glance it looks imo like a quite carefully executed experiment.
Just a quick remark; the 95% confidence level does not mean that there is a 95% probability that the confidence interval covers the real result.
As I said, in theory you have a point. And the test is not flawless.
Welcome to the world. The Matrix seems to have you.
Exactly.
This experiment like (most probably) every other isn´t flawless (for some reasons humans do not tend to be perfect) and Krunok indeed has a point which has to be considered when trying to draw conclusion from the results.
But his argument isn´t restricted to the use of the ADC but could be expanded to every part of the reproduction chain overall. You need other gear to transform signals into sound and none of this is working in a perfect manner (even if it is considered to be "transparent" ) .
That it is of utmost importance to state the objective of any test as clear as possible, can´t be emphasized enough imo.
@oivavoi ,
As usual, replications are needed to find some more corroboration for the results. I haven´t read the description closely yet, but at a first glance it looks imo like a quite carefully executed experiment.
Just a quick remark; the 95% confidence level does not mean that there is a 95% probability that the confidence interval covers the real result.
Of course, good clarification!
Similar threads
