Hi Paul,
Just saw it now. You are right the Notch increase the skirt. The compensation is not perfect by the file. It looks like the ESS version has better full scale specs.
-
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!
Beta Test: Multitone Loopback Analyzer software
- Thread starter pkane
- Start date
Just saw it now. You are right the Notch increase the skirt. The compensation is not perfect by the file. It looks like the ESS version has better full scale specs.
OP
- Thread Starter
- #3,962
OP
- Thread Starter
- #3,965
0dB play gain, DAC set to -0.0dB volume, 256k FFT with Chebyshev 200 window, measured directly by Cosmos ADC.
Sokel
Master Contributor
- Joined
- Sep 8, 2021
- Messages
- 6,279
- Likes
- 6,406
You can see that at the low left corner,with every measurement.
OP
- Thread Starter
- #3,968
Multitone ensures no digital clipping when generating multitone signals. -8.8dBFS is the measured rms level output by the ADC.
What is the input sensitivity of COSMOS you are using?
What is the full scale voltage of your DAC? 4V or 5V?
I want to repeat the test.
OP
- Thread Starter
- #3,971
4.1v at the output of the DAC, 10v sensitivity at the ADC.
Hi Paul,
It is not clear to me what is the multitone test you are using. It is 31 tones. Anyway, included are some possible tests.
The THD+N is similar to yours, but higher 1/f. I put the cosmos on 10V, mono mode. The DO300EX full scale is 5V, so you see the output impedance
of the DO300EX is really high (compare to yours). In DO300EX the XLR output is taken after the LPF before the conversion to SE (and after blocking capacitors). So, the output impedance is not so low, at least compare to you.
Actually, the RCA output impedance is lower 100ohm.
If you tell me the exact test you made I can repeat.
It is not clear to me what is the multitone test you are using. It is 31 tones. Anyway, included are some possible tests.
The THD+N is similar to yours, but higher 1/f. I put the cosmos on 10V, mono mode. The DO300EX full scale is 5V, so you see the output impedance
of the DO300EX is really high (compare to yours). In DO300EX the XLR output is taken after the LPF before the conversion to SE (and after blocking capacitors). So, the output impedance is not so low, at least compare to you.
Actually, the RCA output impedance is lower 100ohm.
If you tell me the exact test you made I can repeat.
Attachments
OP
- Thread Starter
- #3,973
I used 96k out and 96k in for my testing, otherwise the same. Here are the settings you should reproduce on your side, if you want to repeat this test:
Sokel
Master Contributor
- Joined
- Sep 8, 2021
- Messages
- 6,279
- Likes
- 6,406
Boy,that's a nice one.Yes, it has a really good measured performance. Can't get much better than this
View attachment 357837
Can you please,PLEASE,measure it with DSD64 and DSD256 at 1Khz???
I have a problem. I can't play DSD with MTA. Tried all but not working. I understand I need to choose ASIO driver for the
DAC in order to play DSD. None is working, not DoP or Native. The DO300EX can play native DSD512 or 256 (do not remember)
but not with MTA. I am using Win 11 with the SMSL Thesycon official driver.
Sokel
Master Contributor
- Joined
- Sep 8, 2021
- Messages
- 6,279
- Likes
- 6,406
What does the USB driver panel reports when you try to play it?
Have you tried both of Freq families (44100 and 48000Hz) as most don't play the 48000Hz ones?
Also try 705.6 and 768Khz to see.
Hi Paul,
In your chart (AP32 test) the valid range of measurement is 10hz to 22.4khz and not 48khz as written in the screenshot above.
I changed everything to be like yours except that I put upper limit of 22.4khz.
Now. to optimize the result I first tried to reduce the COSMOS sensitivity to (6,7), that alone increase to 108db. Then, as I told you the output impedance of the DO300EX from XLR output is not low, the LPF filter is un buffered. Hence, I put the COSMOS scaler in 1:1 mode between DAC and ADC and changed ADC sensitivity to 4.5V. Although my DAC full scale is 5V, I trusted the MTA to minimize the CREST or prevent clipping. As you can see in the plot below, this is a new benchmark
. Something that I didn't like is that the scaler introduces some high frequencies. It might comes also from cabling. What I really do not understand in your shot is that you have a tone above 20khz and in my AP32 there is not such a tone, even I put the same setting.
Attachments
I will check. From the DAC driver? I tried flex ASIO which doesn't work or [ASIO DAC] that works but not really DSD. How much time it should take the DSD generator to make AP32, 256K, 256DSD?
Sokel
Master Contributor
- Joined
- Sep 8, 2021
- Messages
- 6,279
- Likes
- 6,406
Depends on the signal and the PC,in my ancient PC takes about 15 secs to generate DSD256 with 8 averages for the first try,after that (on repeat) is instantaneous.
I mean the Thesycon's panel,like this (yellow highlighted) :
OP
- Thread Starter
- #3,980
Guy, in my chart, as I posted, I'm using Multitione 32 signal. That's the signal generated by Multitone, not the AP signal. Multitone will generate this up to the Nyquist frequency, which for this test was 48k. Multitone was set to generate tones between 10 and 44.1k (the two numbers I highlighted in Test Signal settings). So, it is valid for measurement range to be from 0 to 48k, if I really want, I didn't break any lawsHi Paul,
In your chart (AP32 test) the valid range of measurement is 10hz to 22.4khz and not 48khz as written in the screenshot above.
I changed everything to be like yours except that I put upper limit of 22.4khz.
Now. to optimize the result I first tried to reduce the COSMOS sensitivity to (6,7), that alone increase to 108db. Then, as I told you the output impedance of the DO300EX from XLR output is not low, the LPF filter is un buffered. Hence, I put the COSMOS scaler in 1:1 mode between DAC and ADC and changed ADC sensitivity to 4.5V. Although my DAC full scale is 5V, I trusted the MTA to minimize the CREST or prevent clipping. As you can see in the plot below, this is a new benchmark
What I really do not understand in your shot is that you have a tone above 20khz and in my AP32 there is not such a tone, even I put the same setting.
Similar threads
- Poll
