E1DA Cosmos ADCiso Review

[torgeirs]

Agree 95% of users just need USB. And guess just a very few use the I2S out inside the adciso

 

[KSTR]

for what use?

To be able to sync a DAC to it, so that sample-synced record while playback is possible.
The current model can be patched to that functionality (with a special firmware + a driver circuit for a transformer or optical out) but it would be appreciated if it were integrated.
Standalone operation without USB (except for power) is not required for me and would be hard to implement anyway, I think.

 

[torgeirs]

Maybe it is just firmvare that is needed to output spdif from one of the I2S pins on the board

 

[melowman]

MINIDSP Adept ADC & Phono Preamp Review

This is a review and detailed measurements of the MINIDSP Adept analog to digital converter and phono equalization/capture. It was sent to me by the company and costs US $299. The device is rather attractive for its class. User interface though, is unusual. You push the rotary encoder in and...

www.audiosciencereview.com

The issue is that input sensitivity is too low (4.4V max). The 10V input sensitivity of the Cosmos is perfect for me.
Plus, the ADept won’t allow to change the AA filter, as opposed to the Cosmos.
(To answer the question: I would be using it in audio mastering context).

 

[morillon]

The Cosmos was dedicated to measurement... The Adept is for use in the consumer market, with suitable line level and standard impedance... as a complement to their other products... not professional... That's all.
(The quality of your master's degrees really won't depend on these points. ;-) )

 

[IVX]

Maybe it is just firmvare that is needed to output spdif from one of the I2S pins on the board
View attachment 494195

When I added that pin-spdif to the Cosmos ADC FW, no one reported that using it. So, we can see how many users need that.

 

[torgeirs]

quality of your master's degrees

Funny autocorect?

 

[morillon]

Funny autocorect?

Yes
"Audio mastering"..sorry

 

[johny_2000]

ES9823/9843 offers >3db less noise but H3 is -125db, I don't think I can use that chip in the next-gen Cosmos ADC as is. And we are lucky again, all ESS ADC chips have a DSP core onboard, so, if ESS staff are able to implement a phase-accurate THD compensation with the DSP, 9823 MONO will show as THD+N around -129db.

Why not? I think the ES9823PRO chip is a step forward for the Cosmos ADC (No. 2?) to maintain its market leadership.
They revised the datasheet (0.4.1) as of November 25, 2025.

 

[IVX]

Why not? I think the ES9823PRO chip is a step forward for the Cosmos ADC (No. 2?) to maintain its market leadership.
They revised the datasheet (0.4.1) as of November 25, 2025.

H3 -125db is ok? Our folks get used to seeing -136-140db with 80% of Cosmos ADC. Let me test the official samples 9823, and let's see the result.

 

[melowman]

Got an ADCiso grade A but I run through several issues. It's hooked up to an RME ADI-2 DAC FS (AKM version) via custom made 0.6m XLR cables. No APU, no Scaler.
DAC's output is set to +19dBU, ADCiso's sensitivity is set to 4.5V.

In REW (latest beta version), I cannot use the Java Exclusive mode (nor the regular Java mode), I get the following error:

In Windows audio settings, when I try to change the sample rate of the ADCiso to 48kHz, I also have an error:

When I click OK, the sample rate goes back to 44.1kHz. I cannot change the sample rate, no matter what.

I have no idea what is going on.

Using FlexASIO works, but then looking at the levels, why is the right channel 18dB higher in peak than the left channel?

Some info:
- I have installed the Comtrue drivers:

- I'm on Windows 11, LTSC, 24H2:

Has anybody encountered those issues?

---

Besides all that, I am getting better measurements of my RME ADI-2 DAC FS than Amir (with FlexASIO), and for some reason this makes me happy
And that is with ADCiso in 'pseudo MONO' mode, i.e. in MONO mode but with a dual XLR cable -- I know that is not supposed to work like that, but I don't have an XLR splitter at hand, and that gave me nice results --
In STEREO mode though, the Harmonic Distortion is spot on with Amir's results (same harmonic profile and values) but the SNR is about 1dB higher than Amir => 'Pseudo MONO' mode allowed me to have a better noise component.

Look at that:

 

[IVX]

I have never seen that msg from REW. You can try to uninstall the Comtrue driver and try Java EX again(64k buffer looks too small to me, try 256k). Regarding the REW's signal meter, it shows you DC offset levels, one channel -80db(450uV), another -98db(56uV). If it bothers you, check the REW\input options\High pass.

 

[DavidR]

I have seen the first message with the ADCiso and other hardware. I would never use Java. FlexASIO works well once past settings issues that seem to be reappear often. Once REW is working with it during any session I've had no problems at all with my ADCiso. But I have ended up leaving it at the 10V setting for any use.

 

[staticV3]

I have seen the first message with the ADCiso and other hardware. I would never use Java. FlexASIO works well once past settings issues that seem to be reappear often.

Java set to EXCL in the Soundcard settings is bit-exact just like FlexASIO, but without the latter's instability and bugginess.

There is no reason not to use it.

 

[torgeirs]

Also using EXCL in the Soundcard settings and only setting sample rate in REW
Normal windows driver.
Be careful to power the ADC before conecting USB data

 

[DavidR]

Java set to EXCL in the Soundcard settings is bit-exact just like FlexASIO, but without the latter's instability and bugginess.

There is no reason not to use it.

My initial experience had been with the Scarlett 2i2 Gen 4 in loopback, later adding the ADCiso. I just tested the D50-ADCiso using Java. That does work, it provides the various sample rates as options. I also just tested the 2i2 in loopback using Java again. The only sample rate even showing in the option list is 96kHz that was forced on selecting it using the Java driver. Evidently there's an issue with the 2i2 driver. The2i2-ADCiso also had this problem back when I was first trying that pairing. Evidently the issue was(is) strictly with the 2i2 driver.

 

[melowman]

I have never seen that msg from REW. You can try to uninstall the Comtrue driver and try Java EX again(64k buffer looks too small to me, try 256k). Regarding the REW's signal meter, it shows you DC offset levels, one channel -80db(450uV), another -98db(56uV). If it bothers you, check the REW\input options\High pass.

I uninstalled the Comtrue driver, rebooted, the driver was then back to the default Windows driver, and no luck: still the same error message in REW, buffer at 256k or 512k (any buffer size actually). Something is wrong or broken somewhere. I don't know if my unit is faulty.

 

[IVX]

I uninstalled the Comtrue driver, rebooted, the driver was then back to the default Windows driver, and no luck: still the same error message in REW, buffer at 256k or 512k (any buffer size actually). Something is wrong or broken somewhere. I don't know if my unit is faulty.

If your ADC works with FlexASIO, how could it be faulty?

 

[torgeirs]

Thinking back, I have seen those error mesages using ADCiso with my RME babyface. I think it was when using SPDIF and internal samplerate on RME other than the java samplerate

 

[melowman]

If your ADC works with FlexASIO, how could it be faulty?

I want to use ASCiso in other contexts than REW and I am not able to change the sample rates in Windows settings.
Also, I’ve been reading about that REW error message and it suggests that I need to change the sample rate in Windows audio settings - which I cannot.

Windows lists all supported sample rates and bit depths properly but I cannot make any changing, I get an error message and I’m stuck on 44.1kHz. That’s with or without Comtrue driver.
For example if I want to change to 48kHz:

It’s a bit of a bummer to be stuck at 44.1kHz to say the least.