Best spec ADC Chip currently.... ??

[Rja4000]

OMG, who may want to add the internal DR compressor to the ADC chip, just to reduce the specs for the compression ratio?

Not you, for sure
But some audiophile money maker, who knows ?

 

[IVX]

To be honest, I like to add the compressor to my PowerDAC, and I prefer that, opened-back headphones have poor noise isolation so the vinyl-like compression helps to recover the details, take them out from the ambient room noise.

 

[Rja4000]

Those guys are always looking for a way to differentiate from the competition.
So some slow compression may be a way to achieve that.
It wouldn't show too much in basic measurements, but a few dB level increase here and there would probably be audible enough to justify that they're doing some magic...

 

[KSTR]

I am also interested in I2S output, not just USB from the Cosmos unit. That would really of interest to those of us that want to use the output digitally, either as I2s or convert to S/PDIF, AES/EBU, TOSLINK, etc...

The missing sync options is almost a deal breaker for me for the type of stuff I do. The ADC need some sort of clock output so that a DAC can be synced with it, or vice versa. The simplest solution (for me) would be a TOSLINK/SPDIF Output (with no data) at 1x or 0.5x or 0.25x the ADC sample rate that could be used to sync an ADI-2 Pro perfectly (which can apply the inverse scale factor), both running on USB.

More specifically for the E1DA cosm, Cosmos DAC and ADC should be enabled to run on a single common base clock but arbitrary multiples of a sample rate (say DAC at 96kHz and ADC at 384kHz), that would be a killer feature for killer combeo. Ideally distributed transformer coupled so DAC and ADC may run galvanically isolated... @IVX, mabye for a future revision of these product, are you thinking to implement hard sample sync possibilities?

 

[WolfX-700]

 

[Rja4000]

Thanks !

 

[IVX]

However, 40Hz THD hump is 100% setup-test artifact, I see no anything special at 40Hz in my tests at all. BTW, the ADC is still in the MONO mode, did you read this? https://e1dashz.wixsite.com/index/cosmos-adc

 

[IVX]

KSTR, no time for something like this, if I was a big company but..

 

[bennetng]

For grins, I can also have very good linearity when measuring a Realtek with a super narrowband filter, without the filter, there is no way to identify truncation.

Realtek ALC892 24-bit output, intentionally truncated and dithered to 18-20 bits:

Same as above, with a super narrowband filter:

Just to point out my original post was on 2019, and I've forgotten the detailed filter settings and the frequency of the test tone.

Measurements of Aune S8 DAC

In this thread Followed Aune S8 Actually, I had a bad impression of Aune products before.But since this product quotes ASR data for publicity, and there are some new features that are specifically advertised. I certainly want to buy one for measurement. So ... look at things first This is...

www.audiosciencereview.com

For apple to apple comparison (ADC measurements and same linearity filter settings), here are some other examples from Wolf's website:

Measurements & Review PrismSound Lyra2 - L7Audiolab

台式设备 PrismSound Lyra2 这产品我如果没搞错的话，已经是8年前发布的了，因为这条系列的产品我不是特别熟悉，印象当中比较“接近”的似乎是RME的ADI2-Pro？而且之前（2016年）数码多也测过Lyra2，还被用555的怼过一波。

www.l7audiolab.com

Measurements of Focusrite Clarett 2Pre USB - L7Audiolab

台式设备 这台Focusrite Clarett 2Pre USB是阳子姐姐送测的。对声卡类的不是很熟，查了下这算是比较入门的产品？好了，先看东西：

www.l7audiolab.com

 

[IVX]

bennetng, hence, that test is redundant, right? If we make that test for 20-20k BW, we'll get a kinda weird SNR test with a nonstandard result, that's it. The unweighted noise was reported as -125db level, and nonlinearity+noise reached +.5db at -101db of stimulus. +.5db is the 1+1/17, and 1/17 is -24db ratio, -101db-24 = -125db, so, it is full domination of the ADC noise and nothing about the linearity. I can't believe if the lack of linearity can be located at -100db for any >>16bits ADC.
PS: I just added 24db to the Focusrite Clarett 2Pre USB and got 114db(measured SNR is 116db, seems some tiny portion of nonlinearity is detected). 24db to PrismSound Lyra2 gave exactly what was measured 112db

 

[axismundi]

However, 40Hz THD hump is 100% setup-test artifact, I see no anything special at 40Hz in my tests at all. BTW, the ADC is still in the MONO mode, did you read this? https://e1dashz.wixsite.com/index/cosmos-adc

Indeed, no artifact at 40Hz but your measurements confirm quite a significant deterioration of THD+N at frequencies below 150Hz or so... I mean -110dB THD+N at 20Hz is not bad at all in absolute, but still around 10dB worse than performance at 1Khz.

Is this independent of the compensation parameters or rather a "victim" of optimizing for 1KHz? A limitation of the ADC chip?

This makes me think of AK5394A which exhibits increasing level of noise (and lower THD+N) in the lower frequencies, alleviated to a large extent by placing larger capacitors at VREF pins... Could it be the same cause here?

 

[IVX]

axismundi, maybe. I calibrate each Cosmos ADC at [email protected].

 

[Hee]

Is this independent of the compensation parameters or rather a "victim" of optimizing for 1KHz? A limitation of the ADC chip?

This makes me think of AK5394A which exhibits increasing level of noise (and lower THD+N) in the lower frequencies, alleviated to a large extent by placing larger capacitors at VREF pins... Could it be the same cause here?

View attachment 157975

I agree with this.
If the difference is from the capacitor, It can do better after testing.

Maybe Ivan is already testing.

 

[phoenixsong]

I beg to differ.
You may want to have to look here, as an example.

Yeah, I came for the same reasons (to find good preamp suggestions for the Cosmos ADC), not because I couldn't find any but because most of their prices do not make sense beside the Cosmos lol

 

[Vuki]

The missing sync options is almost a deal breaker for me for the type of stuff I do. The ADC need some sort of clock output so that a DAC can be synced with it, or vice versa. The simplest solution (for me) would be a TOSLINK/SPDIF Output (with no data) at 1x or 0.5x or 0.25x the ADC sample rate that could be used to sync an ADI-2 Pro perfectly (which can apply the inverse scale factor), both running on USB.

More specifically for the E1DA cosm, Cosmos DAC and ADC should be enabled to run on a single common base clock but arbitrary multiples of a sample rate (say DAC at 96kHz and ADC at 384kHz), that would be a killer feature for killer combeo. Ideally distributed transformer coupled so DAC and ADC may run galvanically isolated... @IVX, mabye for a future revision of these product, are you thinking to implement hard sample sync possibilities?

It seems ess adc even has spdif transciever on chip (at least, obsolete devices had it).

 

[Rja4000]

View attachment 157843View attachment 157844

I think you should update your procedure for linearity.

If what you aim is provide comparable results to Amir's, you'd better align to similar process.
He's using 200Hz and narrow band filter around it, as far as I understood.
I also did some measurements, and only got significant results when doing that.

As Ivan said, otherwise, you're just doing a (non really readable) noise test.

 

[Rja4000]

Yeah, I came for the same reasons (to find good preamp suggestions for the Cosmos ADC), not because I couldn't find any but because most of their prices do not make sense beside the Cosmos lol

There are pretty decent interfaces with good Mic preamps out there.
The good ones are noise-limited, at most useful gains and levels. And, well, there is little to be done to improve on thermal noise.
So a better ADC won't help you on that, I'm afraid.

If you want to get a cheaper mic preamp, you may want to play with this.
No clue how well it performs, though.

 

[phoenixsong]

There are pretty decent interfaces with good Mic preamps out there.
The good ones are noise-limited, at most useful gains and levels. And, well, there is little to be done to improve on thermal noise.
So a better ADC won't help you on that, I'm afraid.

If you want to get a cheaper mic preamp, you may want to play with this.
No clue how well it performs, though.

I wanted to ask more about it, but created another thread in order not to spam off topic here https://www.audiosciencereview.com/forum/index.php?threads/search-for-good-mic-preamps.27205/

 

[Grooved]

ell, even for measurements it's interesting.
It's much easier to get a good measurement when the clocks of ADC and DAC are synched.
I usually use the ADC as the clock source.

The missing sync options is almost a deal breaker for me for the type of stuff I do. The ADC need some sort of clock output so that a DAC can be synced with it, or vice versa. The simplest solution (for me) would be a TOSLINK/SPDIF Output (with no data) at 1x or 0.5x or 0.25x the ADC sample rate that could be used to sync an ADI-2 Pro perfectly (which can apply the inverse scale factor), both running on USB.

More specifically for the E1DA cosm, Cosmos DAC and ADC should be enabled to run on a single common base clock but arbitrary multiples of a sample rate (say DAC at 96kHz and ADC at 384kHz), that would be a killer feature for killer combeo. Ideally distributed transformer coupled so DAC and ADC may run galvanically isolated... @IVX, mabye for a future revision of these product, are you thinking to implement hard sample sync possibilities?

KSTR, no time for something like this, if I was a big company but..

Isn't it possible to get a sync signal from the I2S connection to sync a DAC as slave ? Don't know if the LRCK is outputing Word signal or if it can only be used to sync the SDOUT ?
@IVX would it need programming or is the I2S interface already usable ?

 

[IVX]

ES9822