Beta Test: Multitone Loopback Analyzer software

[Music1969]

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)

RME's ESS based ADI-2's have cleaner output with PCM353kHz (44.1kHz x 8) than original 44.1kHz, in my measurement with Cosmos ADC + REW

Innomaker DAC PRO HAT for RPi4 also (another ESS based).

 

[pkane]

RME's ESS based ADI-2's have cleaner output with PCM353kHz (44.1kHz x 8) than original 44.1kHz, in my measurement with Cosmos ADC + REW

Innomaker DAC PRO HAT for RPi4 also (another ESS based).

Yes, and you can measure that without upscaling, since RME bypasses (edit: some RME devices with AKM chip insert slower roll-off filter at higher sampling rates) changes the digital filter at higher sampling rates.

There is no difference between an upsampled signal and the same signal generated directly at the desired sample rate, except for any artifacts that a poor upsampling algorithm might produce. The DAC doesn't react differently to input that's upsampled or naturally generated (or recorded) at that rate.

I think I know where this comes from (HQPlayer by any chance?) But the difference there is that the comparison is between the DAC doing the upsampling or HQP doing it, with material that's generated or recorded at lower sample rates.

In MTA, the PCM est signal is generated precisely at the desired rate, so there's no need to upsample it You want to test what the DAC does at 768k? Just set the desired output rate to 768k and press the record button. The test signal will be "natively" generated at 768k. There's no upsampling involved or needed at the DAC or in MTA. If your DAC turns off digital filter at 768k, like RME does, then it will be turned off with the "naturally" generated MTA test signal at the input.

 

[Music1969]

Just set the desired output rate to 768k and press the record button. The test signal will be "natively" generated at 768k.

What frequency will J-test be performed at in this example, with current MTA ?

 

[pkane]

What frequency will J-test be performed at in this example, with current MTA ?

768k samplng rate, with the "large" square wave component being generated at half the Nyquist frequency, or 192k.

 

[Music1969]

768k samplng rate, with the "large" square wave component being generated at half the Nyquist frequency, or 192k.

that shows how it is not practical, at present

If I want to see jitter of a red book CD rate but upsampled to PCM 705k, I can't use MTA as it is presently

Jtest at 192kHz is not the same as at 44.1kHz

 

[Music1969]

Whereas with DSD256 upsampling in MTA , it is done correctly

I choose 44.1kHz JTest, output DSD256 and the Jtest is done at correct frequency

I only ask for equivalent with PCM

 

[pkane]

that shows how it is not practical, at present

If I want to see jitter of a red book CD rate but upsampled to PCM 705k, I can't use MTA as it is presently

Jtest at 192kHz is not the same as at 44.1kHz

Sorry, but MTA computes J-Test by AES3 definition, as documented by Julian Dunn. What you're asking for is not a J-test, but something else.

 

[pkane]

Whereas with DSD256 upsampling in MTA , it is done correctly

I choose 44.1kHz JTest, output DSD256 and the Jtest is done at correct frequency

I only ask for equivalent with PCM

DSD test signals are upsampled by necessity, since generating these directly at the desired frequency will take half an hour or longer. PCM does it without upsampling.

 

[MC_RME]

Yes, and you can measure that without upscaling, since RME bypasses the digital filter at higher sampling rates.

There's no upsampling involved or needed at the DAC or in MTA. If your DAC turns off digital filter at 768k, like RME does

I would like to point out that we don't do anything like that. In fact the ES9038Q2M used in the ADI-2 DAC supports all filters in all sample rates. The AKM version (AK4493) does not, it switches to a fixed Slow filter above 192 kHz.

 

[pkane]

I would like to point out that we don't do anything like that. In fact the ES9038Q2M used in the ADI-2 DAC supports all filters in all sample rates. The AKM version (AK4493) does not, it switches to a fixed Slow filter above 192 kHz.

Maybe I misremembered or confused the products. Isn't that what the ADI-2 Pro does at rates higher than 192k -- replace the filter with a slow roll-off one?

 

[MC_RME]

Sure, the ADI-2 Pro uses AKM in and out, and these chips do that. But this discussion is about ESS DAC, and there is no '2-Pro' with these. BTW, your ADI-2/4 has this limitation as well, but only on the ADC side and only at 7xx kHz. It's how the ESS ADC works, we can not change that.

 

[morillon]

I would like to point out that we don't do anything like that. In fact the ES9038Q2M used in the ADI-2 DAC supports all filters in all sample rates. The AKM version (AK4493) does not, it switches to a fixed Slow filter above 192 kHz.

I am taking advantage of the presence of Mr. Rme....
a somewhat specific question :
would it be possible to carry out certains measurements of digital sources without the impact of your powerful steadyclock???
thanks.
;-)

 

[pkane]

Sure, the ADI-2 Pro uses AKM in and out, and these chips do that. But this discussion is about ESS DAC, and there is no '2-Pro' with these. BTW, your ADI-2/4 has this limitation as well, but only on the ADC side and only at 7xx kHz. It's how the ESS ADC works, we can not change that.

Right! I was mentioning RME devices as an example where higher sampling rates may alter the measured DAC behavior by changing filters. I misstated that the filter is removed there, it's just replaced by a slower one.

 

[MC_RME]

I am taking advantage of the presence of Mr. Rme....
a somewhat specific question :
would it be possible to carry out certains measurements of digital sources without the impact of your powerful steadyclock???
thanks.
;-)

No way. For good jitter measurements you will need to spend some money...not to say a ton of money...

 

[morillon]

No way. For good jitter measurements you will need to spend some money...not to say a ton of money...

but is it possible to disengage the steadyclock?
(This could allow some observations...)

 

[pkane]

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

Here's an IMD sweep. As I'm slowly adding features, I've added IMD+N variable in addition to IMD. Compared to APx555, MTA computes actual IMD, while APx555 computes IMD+N. For this reason, you can see that some of the hump behavior is visible in pure IMD line, but is obscured by IMD+N type of measurement. This was captured through E1DA Scaler into Cosmos ADC:

 

[MC_RME]

but is it possible to disengage the steadyclock?
(This could allow some observations...)

I answered that - it's not possible.

 

[morillon]

I answered that - it's not possible.

ok..thank you...

 

[KSTR]

@Guy.Cohen , @pkane
IME, THD vs level is best suited to expose ESS hump as it allows to use a simple notch, helping dynamic range and distortion of the measurement chain. Preferably at a low frequency, like 200Hz or so.
"Too much action" at higher frequencies, be it from the single tone or the second component of a twin-tone, tends to blur the hump as the distortion is then transformed to broadband noise. We want the DAC to very slowly move through its output voltage range.

Actually, the best representation would be absolute level of the distortion residual (with much of the noise factored out) vs generator level.
A characteristic of the ESS hump is that after the onset of the specific distortion it remains almost constant in level... up to the point where the more regular distortion (H2, H3) finally takes over in absolute level.

 

[Guy.Cohen]

Here's an IMD sweep. As I'm slowly adding features, I've added IMD+N variable in addition to IMD. Compared to APx555, MTA computes actual IMD, while APx555 computes IMD+N. For this reason, you can see that some of the hump behavior is visible in pure IMD line, but is obscured by IMD+N type of measurement. This was captured through E1DA Scaler into Cosmos ADC:



View attachment 358588

Hi Paul,

Thanks. So, it has the ESS hump. Here is DO300EX IMD (CCIF, I don't know if you did SMPTE). It done with REW () and you can see scaler/no scaler. The horizontal axes is generator level and of course without noise as one can not see hump with noise. I also have other graphs of SMPTE, THD etc.