Beta Test: Multitone Loopback Analyzer software

[Guy.Cohen]

OK, it looked a little strange in the lower frequencies. Try increasing FFT size to 512k. It's possible the clock drift is greater between your DAC and ADC than what I get here.

I tried 512k, more or less the same. Clearly the first harmonics looks narrower with 512k, but THD number is about the same plus-minus 0.1db

 

[Music1969]

There's no PCM upsampling in MTA. The test signal is generated for the specified sample rate.

can you add ? (please)

same filters you made for DSD will be good enough

 

[pkane]

can you add ? (please)

same filters you made for DSD will be good enough

Why? What's the point of filtering something that doesn't have any frequencies exceeding Nyquist?

 

[pkane]

I tried 512k, more or less the same. Clearly the first harmonics looks narrower with 512k, but THD number is about the same plus-minus 0.1db

Ok, then it's likely the effect of the wider 1/f skirt that I'm seeing in the noise floor.

 

[Guy.Cohen]

It is clear the 1/f is higher. However, we did not see its impact compare to yours on 0-22.4k. If we want to be sure it is the 1/f, then maybe you will be able to compute your latest MTA32 (on 4.5V) in the range 10-22.4k. Your previous published in that range was with 10V sensitivity.
For 10-22.4k and 4.5V I got 110.1db. I wonder what your DAC is achieving compare to 109.9db in the wider range.

 

[Guy.Cohen]

One more, I believe the two SMSL are very similar (more similar than different). At least power supply are very similar and op amp types (OPA1612). So, I have a feeling that AK4499 (the chip).has little higher 1/f noise. Of course, the I/V resistors, could be of different values, the LPF also but I have a feeling the AK4499 has this drawback compare to ESS.

 

[pkane]

maybe you will be able to compute your latest MTA32 (on 4.5V) in the range 10-22.4k. Your previous published in that range was with 10V sensitivity.
For 10-22.4k and 4.5V I got 110.1db. I wonder what your DAC is achieving compare to 109.9db in the wider range.

Here you go:

 

[Guy.Cohen]

So, probably it is the 1/f that disturb in my DAC. You can see it also in the sfdr that I think is limited by this flicker noise.

 

[Guy.Cohen]

Now it remains to see if it is better in THD vs Level or IMD vs level (CCIF, SMPTE). I don't know if you have a notch and/or scaler as without this there is also hump of the ES9822PRO

 

[Music1969]

Why? What's the point of filtering something that doesn't have any frequencies exceeding Nyquist?

At the moment don't have PCM upsampling though?

Can you consider adding?

 

[pkane]

At the moment don't have PCM upsampling though?

Can you consider adding?

Again, why?

 

[Music1969]

Again, why?

Same reason you have DSD upsampling?

bypass on chip upsampling to see measurable improvements

 

[Sokel]

Same reason you have DSD upsampling?

bypass on chip upsampling to see measurable improvements

You can do the same going straight to the rate you want.
MTA goes 768k and DSD1024,no need for upsampling.

 

[pkane]

Same reason you have DSD upsampling?

bypass on chip upsampling to see measurable improvements

As @Sokel indicated, MTA generates test signals directly at the desired sampling rate for PCM output. There's no reason to upsample these or to filter them: they are generated already band-limited to the desired sample rate.

 

[Music1969]

As @Sokel indicated, MTA generates test signals directly at the desired sampling rate for PCM output. There's no reason to upsample these or to filter them: they are generated already band-limited to the desired sample rate.

Understand

But reason i use the DSD upsampling is to test how DAC output measures at different rates

Would be same goal with PCM (if feature could be added)

 

[pkane]

Understand

But reason i use the DSD upsampling is to test how DAC output measures at different rates

Would be same goal with PCM (if feature could be added)

Sorry, but I still don’t understand what purpose upsampling would serve here since multiple rates are already available without upsampling. Can you give an example of what you think upsampling could do for PCM testing?

 

[Music1969]

Sorry, but I still don’t understand what purpose upsampling would serve here since multiple rates are already available without upsampling. Can you give an example of what you think upsampling could do for PCM testing?

If I want to upsample a 44.1kHz music track to 705kHz

It is not the same as outputting a 705kHz signal in MTA is it? You will have frequency response of that test signal to 353kHz ? Which is not the same as a 44.1k CD track for example

 

[Sokel]

If I want to upsample a 44.1kHz music track to 705kHz

It is not the same as outputting a 705kHz signal in MTA is it? You will have frequency response of that test signal to 353kHz ? Which is not the same as a 44.1k CD track for example

Upsampling would only show the quality of the player's algorithm as the DAC is not involved with this.
Doesn't matter if one plays 8 x 44100Hz or straight 352800Hz on the player or generator or wherever,the DAC's input sees the same sampling.

 

[Music1969]

Upsampling would only show the quality of the player's algorithm as the DAC is not involved with this.

Well if you bypass some DAC chip upsampling, you definitely can see it on the DAC output.

I want to see if such makes a difference for some DACs . I already do this anyway with HQPlayer but would be nice if MTA had it all built in.

Like you can see difference DSD256 makes with ESS DACs with IMD hump, so the upsampling of PCM44.1k to DSD256 is useful for that.

PCM upsampling shows other differences with some DACs.

 

[Sokel]

Well if you bypass some DAC chip upsampling, you definitely can see it on the DAC output.

I want to see if such makes a difference for some DACs . I already do this anyway with HQPlayer but would be nice if MTA had it all built in.

Like you can see difference DSD256 makes with ESS DACs with IMD hump, so the upsampling of PCM44.1k to DSD256 is useful for that.

PCM upsampling shows other differences with some DACs.

I have tried this with foobar and SoX as external generator signal and it had zero difference.
So was upsampling from PCM to DSD256 through DSD processor.
The end result was the same as MTA's DSD256 (a tiny bit worst noise for foobar/SoX)