Measurements: "ESS Hump" revisited (Khadas Tone Board V1.3)

[mansr]

Btw, "ESS hump" shall be considered a misnomer by now because we can be quite sure it's a problem in the analog output stage (the I/V-converter and subtractor circuit), not ahead of it. The chip itself is not to blame here.

Can we continue blaming the so-called data sheet?

 

[KSTR]

@mansr,
Well, after some time of digesting the known ESS white paper and the ES9038q2m datasheet I would say it's all in there, or at least most of it.

What was a new discovery not explicitly touched in the papers is that the common-mode signal on the I/Vs is quite large and distorted. And it is rf-polluted (as is the differential signal) and the whole concept relies on a stable and deep cancelling. This I could clearly see with the D10B, looking at the single ended signals individually as well as at the differential signal in comparision.

When the indivual I/V's are not up to it and have problems coping with the fast transitions and/or additional glitches and then create asymmetrical errors from demodulation, slew-limiting or whatever, the cancelling falls apart. Additionally we may have less than optimum general operating points here, spoiling the baseline some more.
Same applies for the subtractor itself as some glitch content may just jump over the I/V beyond its bandwidth and create problems in the subtractor as well.

At least that's my pet theory for the momement as to what is happening on the KTB and why the cross-capacitor helped.

 

[Jimbob54]

Oh no. So "Sabre Glare" might be a thing after all. Well this is something of a blow. I wouldn't pretend to comprehend 99% of this thread but its been fascinating tracking the developments. Thanks

 

[mansr]

Oh no. So "Sabre Glare" might be a thing after all. Well this is something of a blow. I wouldn't pretend to comprehend 99% of this thread but its been fascinating tracking the developments. Thanks

I don't see anything approaching evidence that these anomalies are audible.

 

[Jimbob54]

I don't see anything approaching evidence that these anomalies are audible.

Tell that to the folks taking that ball and running with it in the wild :-/

 

[MRC01]

How does all this compare with using a DAC chip having voltage output, obviating the need for external I/V conversion?
The WM8741 comes to mind, but it's several years old. Does anyone make DAC chips like this anymore?
Seems like a lot of work to get an I/V that is clean enough to do justice to the DAC chip output. Is the end result better (cleaner audio)?

 

[staticV3]

How does all this compare with using a DAC chip having voltage output, obviating the need for external I/V conversion?
The WM8741 comes to mind, but it's several years old. Does anyone make DAC chips like this anymore?
Seems like a lot of work to get an I/V that is clean enough to do justice to the DAC chip output. Is the end result better (cleaner audio)?

The CS43198 has voltage output, see Topping D30Pro

 

[gvl]

FWIW I never thought KTB had any glare. Sure, it felt more forward sounding that some other DACs I was using at the time, but if anything I thought it was too clean vs. dirty, more of a wow so much detail in your face. Perhaps this is what some call glare. With time my ears/brain got used to the sound and I didn't have any issues listening to it later on.

 

[KSTR]

How does all this compare with using a DAC chip having voltage output, obviating the need for external I/V conversion?
The WM8741 comes to mind, but it's several years old. Does anyone make DAC chips like this anymore?
Seems like a lot of work to get an I/V that is clean enough to do justice to the DAC chip output. Is the end result better (cleaner audio)?

All the AKM DAC's (except 4499) have voltage output, too.
The advantage of voltage output is that it can be (and is most often) passively RC-filtered before hitting any OpAmp inputs which totally avoids transients and glitches throwing off the performance even with lesser spec'd OpAmps.

 

[levimax]

All the AKM DAC's (except 4499) have voltage output, too.
The advantage of voltage output is that it can be (and is most often) passively RC-filtered before hitting any OpAmp inputs which totally avoids transients and glitches throwing off the performance even with lesser spec'd OpAmps.

Really great work on this and fun to follow, thank you! So does this distortion need a catchy "new name" or is this just a type of the infamous TIM distortion?

 

[KSTR]

I don't see anything approaching evidence that these anomalies are audible.

First listening tests with my emulator show that I need to magnify the distortions by some 30..35dB before they become immediately audible (instantly ABX'able) with selected music like softly recorded solo piano. A gut feeling of increased "fatigue" starts to appear at about 10dB lower but this will need long-term AB testing to solidify.

Assuming something like 10..15dB of safety margin for better skilled listeners than me it looks like there is still a 20dB distance to the actual distortion levels as generated by KTB and other "humpy" DACs... so yes, very likely inaudible.

 

[KSTR]

Really great work on this and fun to follow, thank you! So does this distortion need a catchy new "new name" or is this just a type of the infamous TIM distortion?

Not directly TIM, rather SIM (Slewing-Induced Distortion) and/or plain demodulation.

The interesting thing, which kept me on a false track for a while, is that it manifests as if it were a linearity error in DAC itself, the normal straight slope of input-->output overlaid with a oscillating sinusodial offset:

 

[Pdxwayne]

FWIW I never thought KTB had any glare. Sure, it felt more forward sounding that some other DACs I was using at the time, but if anything I thought it was too clean vs. dirty, more of a wow so much detail in your face. Perhaps this is what some call glare. With time my ears/brain got used to the sound and I didn't have any issues listening to it later on.

Do you still have KTB on hand? Maybe a Gustard X16 or something similar?

If so, would love for you to check post #157 and just clicks on the clips for timing test and tell me your observations. For me, the X16 sounded softer in the hi-hat and KTB stronger/sharper in the hi-hat.

First listening tests with my emulator show that I need to magnify the distortions by some 30..35dB before they become immediately audible (instantly ABX'able) with selected music like softly recorded solo piano. A gut feeling of increased "fatigue" starts to appear at about 10dB lower but this will need long-term AB testing to solidify.

Assuming something like 10..15dB of safety margin for better skilled listeners than me it looks like there is still a 20dB distance to the actual distortion levels as generated by KTB and other "humpy" DACs... so yes, very likely inaudible.

Did you try the online timing test I provided earlier?

See https://www.audiocheck.net/blindtests_timing_2w.php?time=100

Simply play the sample files and don't need to run tests.

Pay attention to the hi hat strength/sharpness. Compare ktb to your other DAC without the issue. At least my Gustard x16 sounded much softer in the high hat than KTB.

 

[KSTR]

With that e-drum "hi-hat" I have to turn up distortions by 50dB(!) with the emulator to note anything different. It's already a noise-like signal so adding more noise doesn't do much. It's tonal stuff that is affected most.
If the stock KTB sounds remarkedly different (will test this as well) this may have a bunch of other reasons not neccesarily connected to the hump-generating distortion.

 

[Pdxwayne]

With that e-drum "hi-hat" I have to turn up distortions by 50dB(!) to note anything different. It's already a noise-like signal so adding more noise doesn't do much. It's tonal stuff that is affected most.

Interesting, may I know your listening chain that you used for comparison? Which DAC did you compared?

Mine is just laptop -> DAC -> amp -> headphones.
Significant difference when comparing Gustard x16 vs KTB in term of hi hats.
Less so when comparing Topping E30 vs KTB.

Thanks!

 

[KSTR]

DAC is RME ADI-2 Pro FS R (Laptop-->DAC-->Headphones) and I used the emulation, not a recording of the actual KTB output (will do that later).

 

[gvl]

It’s interesting that early on I would AB DACs trying to find differences in their sound and I would always find something that I thought was different. These days during similar exercises I mostly try to convince myself the DACs sound the same and I usually succeed too. Given the conflict here I now tend to think these subjective experiments is nothing but a waste of one’s time.

 

[Pdxwayne]

DAC is RME ADI-2 Pro FS R (Laptop-->DAC-->Headphones) and I used the emulation, not a recording of the actual KTB output (will do that later).

I wonder if the difference I heard is due to Windows 10 driver difference between Gustard and KTB....I tried two different type of headphones and the observations are consistent. Gustard sounded soft in hi hats. KTB's hi hat is stronger and sharper.

The amp is Gustard H16. Headphones are k371 and HE400SE. OS is Windows 10. Browser is Chrome.

 

[levimax]

It’s interesting that early on I would AB DACs trying to find differences in their sound and I would always find something that I thought was different. These days during similar exercises I mostly try to convince myself the DACs sound the same and I usually succeed too. Given the conflict here I now tend to think these subjective experiments is nothing but a waste of one’s time.

try blind level matched ABX testing... then you will if there are any actual differences you can reliably hear

 

[gvl]

try blind level matched ABX testing... then you will if there are any actual differences you can reliably hear

That’s exactly my point, there aren’t, or at least at least their magnitude is not big enough to care about.