• 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). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

The craziest DAC ever #9038AP with distortions analyzer onboard.

JohnYang1997

Major Contributor
Technical Expert
Manufacturer
Joined
Dec 28, 2018
Messages
4,672
Likes
7,217
Location
China
#41
JohnYang1997, why I should care about your thermal environment? You may perform distortions measurement any time you wish, cold either hot whatever, next save the preset with 2nd/3rd harmonics correction and name it "calibration at +35C" or "calibration at -10C" ;)
View attachment 67600
I asked the two questions because first of all this is a technique that's used for long time called pre distortion. Basically you calibrate your system to have low distortion, any measured distortion will be external. But it drifts in the same environment after 10, 30 minutes. It's useful for some cases but has its limit. That's why I asked how you solved this, I assume you know that.
Second question is simply that you had two pictures with two different tones at op. And didn't specify in any post this "only" works at 1khz(which will be basically useless if that's the case). And it's "trivial" to creat presets for different frequencies I assume.
Also I don't know if you have any mechanism to monitor distortion on the fly and correct it on the fly.
Sorry, but if you think this is crazist shit ever you are out of your mind...
 
OP
IVX

IVX

Senior Member
Forum Donor
Joined
Mar 3, 2019
Messages
449
Likes
1,283
Location
South of China, SHZ area
Thread Starter #42
I asked the two questions because first of all this is a technique that's used for long time called pre distortion. Basically you calibrate your system to have low distortion, any measured distortion will be external. But it drifts in the same environment after 10, 30 minutes. It's useful for some cases but has its limit. That's why I asked how you solved this, I assume you know that.
Second question is simply that you had two pictures with two different tones at op. And didn't specify in any post this "only" works at 1khz(which will be basically useless if that's the case). And it's "trivial" to creat presets for different frequencies I assume.
Also I don't know if you have any mechanism to monitor distortion on the fly and correct it on the fly.
Sorry, but if you think this is crazist shit ever you are out of your mind...
Yes, I do, and seems only you don't like that, others look happy to see something out of boring mind ;)
I said a few times, it works like online distortion measurement and distortions compensation at the same time. Any harmonics drift is not my business, I gave you DAC+HPA+THD measurement and control, if you like you can calibrate that every hour, or every 5 years, up to you.
 

JohnYang1997

Major Contributor
Technical Expert
Manufacturer
Joined
Dec 28, 2018
Messages
4,672
Likes
7,217
Location
China
#43
Yes, I do, and seems only you don't like that, others look happy to see something out of boring mind ;)
I said a few times, it works like online distortion measurement and distortions compensation at the same time. Any harmonics drift is not my business, I gave you DAC+HPA+THD measurement and control, if you like you can calibrate that every hour, or every 5 years, up to you.
Actually I like that. The idea is good and making it an product is awesome.
Thing is, I was just asking a question originally. So now it gets answered. Thanks and good luck on the project!
 
OP
IVX

IVX

Senior Member
Forum Donor
Joined
Mar 3, 2019
Messages
449
Likes
1,283
Location
South of China, SHZ area
Thread Starter #44
Actually I got the second time such feedback last 24 hours, that only 1kHz distortions measurement is useless, and need to have a variable-freq like Audio Precision analyzers notch filter(a lot of relays, and additional highly precise parts, about 10kG unit) or at least at 10kHz instead of 1kHz. However, 1k THD is the standard audio test, and all who know about the THD expect 1k test is #1 for sure. So, if I'll offer 10k test instead of the standard one, 99.9% of users will ask me - are you nuts, why 10k?? If I'll going to implement all-frequency THD analyzer, how I'll reduce its cost down to $100 if APx555 wants $30000 for that? I found an inexpensive way to measure distortions of a DAC at a single frequency, and I choose a standard 1k test for that, I see here a logic only, no any madness.
 

JohnYang1997

Major Contributor
Technical Expert
Manufacturer
Joined
Dec 28, 2018
Messages
4,672
Likes
7,217
Location
China
#45
Ok. After went back and read everything a few times. Now I realized the issue. I.think we had misunderstanding here.
My question now should be, how do you measure the distortion? What's the analyzer performance?
How do you separate the distortion of the dac from the analyzer itself? The point of using notch is to reduce the distortion generated in A-D conversion. How do you achieve that as your point is to give a low distortion dac/hp output instead of measuring the distortion of a external analogue device? Predistortion only works in the close loop system in which you calibrated it.
 

scott wurcer

Major Contributor
Audio Luminary
Technical Expert
Joined
Apr 24, 2019
Messages
1,138
Likes
1,806
#46
Ok. After went back and read everything a few times. Now I realized the issue. I.think we had misunderstanding here.
My question now should be, how do you measure the distortion? What's the analyzer performance?
That was my original read, but if there is only a passive notch it is as good as the R's and C's which can be better than -160dB and still be cheap.
There are numerous examples of using a passive notch to measure THD to extraordinary levels using an ordinary sound card.

As for pre-distortion it is important to realize that at exactly 1kHz output and 96kHz sampling rate there are only 96 codes used which is trivial amount of data. For an arbitrary input you would need to save 2**24 codes. But I don't see that this is what we are talking about, it looks like using the internal ESS distortion modifying registers (which I assume modify the DAC's transfer function). AFAIK the ESS DAC only allows adjustment of the 2nd and 3rd harmonic and one set of coefficients are used for all channels.

The AP is a general purpose instrument for measuring amplifiers, pre-amplifiers, etc. Putting another device in the loop as you say eliminates the possibility of separating the sources of distortion.
 
OP
IVX

IVX

Senior Member
Forum Donor
Joined
Mar 3, 2019
Messages
449
Likes
1,283
Location
South of China, SHZ area
Thread Starter #47
Dear comrades(дорогие товарищи)!
I've got one more spectacular idea regarding 9038AP DAC distortions minimizing. I found that during a hot dispute in our discord channel, and I should thank our discord-folk for that.
What's usually stands behind the term "Dual DAC"? The attempt to reduce noise = increase dynamic range, however, do we really need less noise when our dynamic range is already around 126db(A) with a single one ES9038Q2M? At the same time, we may use Dual DAC configuration for the other thing which more important sexy to me, we can use only one channel of each ES9038Q2M to avoid the limitation of THD compensation based on non-ideal Left/Right harmonics symmetry.
Regardless of legendary-stupid ESS secret-policy, nowadays, every school-girl aware that 9038 DACs have 2nd and 3rd harmonics compensation registers affect both Left and Right channels simultaneously, these channels have not exactly the same THD performance, hence, you have to choose a compromising between THD Left either Right channel. I gonna use only one channel of each 9038 DAC, and the best one, to compensate distortion up to the measurement error level. So the idea is using a dual DAC configuration for the lowes THD instead of a higher dynamic range(which is already high enough to me, and actually higher vs competitor's dual 9038Q2M).
 

JohnYang1997

Major Contributor
Technical Expert
Manufacturer
Joined
Dec 28, 2018
Messages
4,672
Likes
7,217
Location
China
#48
Dear comrades(дорогие товарищи)!
I've got one more spectacular idea regarding 9038AP DAC distortions minimizing. I found that during a hot dispute in our discord channel, and I should thank our discord-folk for that.
What's usually stands behind the term "Dual DAC"? The attempt to reduce noise = increase dynamic range, however, do we really need less noise when our dynamic range is already around 126db(A) with a single one ES9038Q2M? At the same time, we may use Dual DAC configuration for the other thing which more important sexy to me, we can use only one channel of each ES9038Q2M to avoid the limitation of THD compensation based on non-ideal Left/Right harmonics symmetry.
Regardless of legendary-stupid ESS secret-policy, nowadays, every school-girl aware that 9038 DACs have 2nd and 3rd harmonics compensation registers affect both Left and Right channels simultaneously, these channels have not exactly the same THD performance, hence, you have to choose a compromising between THD Left either Right channel. I gonna use only one channel of each 9038 DAC, and the best one, to compensate distortion up to the measurement error level. So the idea is using a dual DAC configuration for the lowes THD instead of a higher dynamic range(which is already high enough to me, and actually higher vs competitor's dual 9038Q2M).
Well. The execution has been done on products with 9038q2m 2 years ago. And that does work. Though you may ask that's not the result you have or can get.
 
OP
IVX

IVX

Senior Member
Forum Donor
Joined
Mar 3, 2019
Messages
449
Likes
1,283
Location
South of China, SHZ area
Thread Starter #49
Actually, as I see, I can make switchable modes by iOS/Android app:
1) Dual DAC for lower noise
2) Dual DAC for lower Distortions(-6db max output level)
 
Last edited:

JohnYang1997

Major Contributor
Technical Expert
Manufacturer
Joined
Dec 28, 2018
Messages
4,672
Likes
7,217
Location
China
#50
One thing you will notice once you design it you will see that the chip is pretty consistent. Not the way that left and right are consistent but one side of the chip is consistently different from the other side. I guess it has something to do with the pinout decision or internal layout. You mostly will eventually use the same/very similar compensation settings for both sides.
 

lashto

Addicted to Fun and Learning
Forum Donor
Joined
Mar 8, 2019
Messages
535
Likes
241
#53
While you are at it, was wondering if its possible to make DACs "modular" ?

In that you can add channels to them as needed by "stacking" them and controlling them via a master USB controller.

It has possible applications of multi channel home theater use.

Similarly, it can be the basis of active speaker crossovers that can be adjusted via software.
Like say a 2 way speaker with a 2 channel module each , that are then connected via a "master"module that splits and controls the channels.
A 4 way speaker using 2 x 2 channel modules.

Its just a wild idea but its not something that I have come across on the market. though
If you ever start that modular thing @IVX , looks like you are only a few steps away from a nice AVP. The good DACs are there, the DSP is there, the control app is there, you just need a bigger case :D

Considering the absolute crap AVR/Ps that amirm keeps measuring you'll have a giant market.
 
OP
IVX

IVX

Senior Member
Forum Donor
Joined
Mar 3, 2019
Messages
449
Likes
1,283
Location
South of China, SHZ area
Thread Starter #54
lashto
I tried to google "AVR/Ps" and "AVP" and still don't understand from where I so close? I'm a simple guy and make the ONLY what I want to have, whatever I need that or rather I like to design that.
 

lashto

Addicted to Fun and Learning
Forum Donor
Joined
Mar 8, 2019
Messages
535
Likes
241
#55
lashto
I tried to google "AVR/Ps" and "AVP" and still don't understand from where I so close? I'm a simple guy and make the ONLY what I want to have, whatever I need that or rather I like to design that.
oh sorry, I though everyone knew those catastrophes by now :)
Here's one AVR https://www.audiosciencereview.com/forum/index.php?threads/marantz-sr6014-avr-review.14615/ .. and wiki

AVR = AudioVideoReceiver = all-in-one device for home theater. It takes 10+ channels of encoded bluray audio, does decoding and room-correction in DSP, then comes the multi-DAC board and then 10+ speaker amps at the end.
AVP = AVProcessor = same but without the speaker amps. Basically, just put 8x of your DSP DACs together and you have an AVP.

Those things sell for $2000+ and sound/measure much worse than your wonderful DSP DACs. Actually worse than the cheapest USB audio dongle that you can buy in China nowadays.
 

maxxevv

Major Contributor
Joined
Apr 12, 2018
Messages
1,658
Likes
1,634
#57
As with my original suggestion, a modular USB based DAC that allows users to add channels as needed.
Each module adding 2 channels. With a separate master input box.
Allowing users to configure as needed 2, 2.1, 3, 4, 4.1, 5.1 and 7.1 channels as needed.

Could possibly be used as active DSP for speakers as well.
 
Top Bottom