E1DA Cosmos ADCiso Review

[Pow1]

Yes.


Only the native macOS volume control will adjust modes, not (usually) the sliders inside applications.

Where you adjust the macOS volume (MIDI Control Panel, Keyboard Vol Up/Down keys, etc) doesn't matter.

Thank you so much for the reply!

Here are some other questions that I have:

• For mono mode, why do we need a splitter to feed the mono signal into L+R of the ADCiso? From what I understand, Mono mode is L + R and fed out the L channel into PC. Is that how we get the 3dB boost in Mono mode? By doubling the signal and averaging away some of the noise?
• If I don’t have a splitter on hand, can I just feed it into a single ch of the ADC? I’m guessing this will give me same SNR as stereo mode.

I posted on the discord as well but figure to perhaps ask here instead so that it's easier for other people to reach from Google

 

[staticV3]

For mono mode, why do we need a splitter to feed the mono signal into L+R of the ADCiso? From what I understand, Mono mode is L + R and fed out the L channel into PC. Is that how we get the 3dB boost in Mono mode? By doubling the signal and averaging away some of the noise?

You feed the same signal into both ADC channels and it averages the result.

Mono mode correctly fed and read will result in same signal level as Stereo mode.

You only lose 3dB when you feed just one channel in Mono mode and read the left output channel.

If I don’t have a splitter on hand, can I just feed it into a single ch of the ADC? I’m guessing this will give me same SNR as stereo mode.

W/o splitter cable, Mono mode is pointless.

But you can still keep the Cosmos in Mono (if switching modes is a hassle), feed just the Right input channel and read just the right output channel.

The right I/O channel functions just the same in both Mono and Stereo.

 

[Pow1]

Not sure if this is the correct place to ask, but I'm trying to make the following 3 measurements of my DAC using the ADCiso and REW (I linked samples from some of Amir's DAC reviews):

Frequency response

THD+N vs Freq

White noise spectrum

Here are my results for the latter 2, and some questions along with them:

THD+N vs Freq: I obtained this graph using the measurements feature in REW, basically the same thing that you would use to do speaker-room measurements. I just ran a sweep of DAC (Fiio KA15) -> ADCiso (Mono mode). The brown squiggly line at the bottom is the noise floor, and the silver line (sorry for it being hard to see) is the THD+N for the DAC. How is REW able to measure THD+N that is lower than the noise floor?

White noise spectrum: As I understand it, white noise is played beyond the 20kHz range and should reveal the kind of filters used in the DAC. I'm using a Fiio KA15 in NOS mode. I know that Amir does the test with a sample rate of 44.1kHz. But if I try to set that in REW, I can't the white noise to play beyond ~22kHz, which kinda makes sense to me since 44.1 kHz is what you need to accurately reproduce frequencies up to ~20kHz. For me to play white noise up to 30kHz as seen in the graph below, I would need to set my sample rate to 192 kHz, which is fine. But then I get a straight line to 30kHz instead of seeing the drops due to the filters. What is going on here? In the white noise spectrum graph I linked above, we see that the line goes up to 30kHz, but how is this possible with only 44.1kHz sample rate?

 

[staticV3]

How is REW able to measure THD+N that is lower than the noise floor?

FFT gain.

White noise spectrum: As I understand it, white noise is played beyond the 20kHz range and should reveal the kind of filters used in the DAC. I'm using a Fiio KA15 in NOS mode. I know that Amir does the test with a sample rate of 44.1kHz. But if I try to set that in REW, I can't the white noise to play beyond ~22kHz, which kinda makes sense to me since 44.1 kHz is what you need to accurately reproduce frequencies up to ~20kHz. For me to play white noise up to 30kHz as seen in the graph below, I would need to set my sample rate to 192 kHz, which is fine. But then I get a straight line to 30kHz instead of seeing the drops due to the filters. What is going on here? In the white noise spectrum graph I linked above, we see that the line goes up to 30kHz, but how is this possible with only 44.1kHz sample rate?

To measure the filter response in REW to beyong Nyquist, you will have to open the Generator, export a White noise signal (ideally 44.1 or 48kHz) as .WAV file, then play it through the DAC using external playback software like foobar, while you record the output in REW with the Cosmos set to 96+kHz sample rate.

Set playback to Loop and set AVGing to Forever in REW and you can watch the filter response measurement become clearer and cleaner over time.

 

[melowman]

@IVX Look at the mess I did out of your creation
Because yes we need SPDIF output!

 

[IVX]

Hmm, but why isn't optical output? Don't you need an isolation?

 

[melowman]

On my setup I only have coax left, so I use coax. But the connector you see on the picture is not coax SPDIF but just the RAW SPDIF coming from the MCLK pin — which I then connect to an external SPDIF buffer.

But of course, if you decide to put a SPDIF out in the next Cosmos ADC, I think it should be both coax and optical!

Plus, if it had Word Clock capabilities, to make it more suitable in a studio context, that would be absolutely awesome.

—

Btw, maybe it’s too much to ask, but I’d be happy if the Cosmos Tweak app allowed to change the Sample Rate — right now, as reported a few posts ago, I cannot change the SR via the Windows options, I get an error.

The only way for me to change the SR is via ASIO, by using an ASIO software; that’s very cumbersome and quite frustrating.

 

[ObjectiveSubjectivist]

Is cosmos discontinued?

 

[melowman]

Ivan has been talking about an mk3 for a moment now. Maybe they’re ready to launch it! Hopefully with SPDIF out and Word Clock capability. And Sample Rate modification in Cosmos Tweak, when in Master clock mode.

 

[ObjectiveSubjectivist]

I'll wait for the new version then.