Antelope Amari Measurements

[restorer-john]

ADC:

Inputs: XLR
Outputs: ASIO
Fs: 192kHz

Monitor:
Monitor for a 23.9dBu input.

Did you experiment with ADC input levels as -0.128dBFS is very high and most A/Ds will not be performing anywhere near their best at such levels.

 

[Toine]

Did you experiment with ADC input levels

Yes I did SINAD vs level in the comments and THD+N vs level in the post. There's about a 5dB step improvement in SINAD if you go under +8dBu input

 

[capslock]

CS43198 is basically the same thing as CS43131 minus the headphone amp, right? And it was shown that CS43131 cheats with gain switching to attain higher DNR which leads to audible defects with, frankly, atypical signals. Why would one select this DAC type for a reference unit, and even parallel them to further up DNR which does not need it because of the questionable gain switching?

 

[pseudoid]

Thank you for the review, @Toine,

"… adjustable output impedance – from -4.6 to 85.3 Ohm, in 17 discrete steps."

I was a bit about their usage of "Negative Impedance" at their website; perhaps they mean "capacitive" (yet,the values are given in "Ohms").
ADD: Impressive performance at an impressive price.

 

[Toine]

And it was shown that CS43131 cheats with gain switching to attain higher DNR

That's really interesting! Do you know where I can find the measures or threads about that ? I looked around a bit with no success

 

[Toine]

I was a bit about their usage of "Negative Impedance"

Negative resistance is not unheard of in electronics.
I'll open it in the lab one day to check it out. I was suspecting a variable resistor on the output (from maxim integrated or something).
But yeah the mention of negative impedance is intriguing.

 

[lewdish]

Where do you work that just has an AXP laying around lmao

 

[pseudoid]

Where do you work that just has an AXP laying around lmao

How Rude!
I bet I am not the only one who refrained from calling him a lucky stiff!

But yeah the mention of negative impedance is intriguing.

TL&DR >> By using an NIC as a negative resistor, it is possible to let a real generator behave (almost) like an ideal generator, (i.e., the magnitude of the current or of the voltage generated does not depend on the load).

 

[MBO]

CS43198 is basically the same thing as CS43131 minus the headphone amp, right? And it was shown that CS43131 cheats with gain switching to attain higher DNR which leads to audible defects with, frankly, atypical signals. Why would one select this DAC type for a reference unit, and even parallel them to further up DNR which does not need it because of the questionable gain switching?

I did not observe such distortions with Amari and CS43198. But observed in certain cases at CS43131.

 

[DrCWO]

You can kinda make that out on the SINAD vs measured level graph. But I tested it as a +24dBu line in/out, it's less easy to compare with other DAC measurements sorry about that.


As in a special function ? Not really, TRS outputs have +18dBu

Might be interesting to see the SINAD at these outputs. Less amplification may give a better result.

 

[milosz]

Reviews on ASR usually show the price of the unit under test in the beginning of the review....

I don't see the price mentioned anywhere here.

 

[Blumlein 88]

Reviews on ASR usually show the price of the unit under test in the beginning of the review....

I don't see the price mentioned anywhere here.

I think $3195 is list on it. I see Thomann shows it for $2869 plus shipping.

 

[AnalogSteph]

Frequency Response:
Here's a sweep at 0dBFS. Keep in mind the left axis scale has 2dB amplitude.

Still, about 0.02 dB worth of filter ripple is too much. I'd check for the presence of resampling / try at a 48 kHz multiple (I doubt anyone in professional circles would be running their DAC at 44.1 these days).

Speaking of sample rates, you took DAC FR at 44.1 kHz and ADC FR at 192 kHz, was that intentional? It certainly is a good idea to be running an AK55xx at minimum 192, even better 384 kHz, mostly because their filter design for single/quad speed recording is a bit silly (+/-0.03 dB of periodic ripple in an FIR filter even in a flagship ADC, c'mon).

The ADC looks pretty good except that weird plateau in IMD starting at +5 dBu.

Probably the input stage / ADC driver breaking a bit of a sweat. Hard to tell which, almost +24 dBu is quite high after all. A distortion sweep may or may not provide some clues.

One selling point still remaining is the oven controlled clock as I couldn't set up the j-test.

Having an OCXO is more about long-term stability / drift anyway. It's quite hilariously overkill for audio, but I guess if you were to feed its 10 MHz reference input from a GPSDO there would never be any doubt about things being in sync, even without the need for a central master clock (which is all fine and dandy as long as all the equipment physically is within the same studio, but you know how it is these days).

GPSDOs are fairly inexpensive, you can get one for 5-10% of what this converter costs. You do need a spot with GPS reception, and obtaining proper sync can take a few hours.

 

[Blumlein 88]

Still, about 0.02 dB worth of filter ripple is too much. I'd check for the presence of resampling / try at a 48 kHz multiple (I doubt anyone in professional circles would be running their DAC at 44.1 these days).

Speaking of sample rates, you took DAC FR at 44.1 kHz and ADC FR at 192 kHz, was that intentional? It certainly is a good idea to be running an AK55xx at minimum 192, even better 384 kHz, mostly because their filter design for single/quad speed recording is a bit silly (+/-0.03 dB of periodic ripple in an FIR filter even in a flagship ADC, c'mon).

Probably the input stage / ADC driver breaking a bit of a sweat. Hard to tell which, almost +24 dBu is quite high after all. A distortion sweep may or may not provide some clues.

Having an OCXO is more about long-term stability / drift anyway. It's quite hilariously overkill for audio, but I guess if you were to feed its 10 MHz reference input from a GPSDO there would never be any doubt about things being in sync, even without the need for a central master clock (which is all fine and dandy as long as all the equipment physically is within the same studio, but you know how it is these days).

GPSDOs are fairly inexpensive, you can get one for 5-10% of what this converter costs. You do need a spot with GPS reception, and obtaining proper sync can take a few hours.

I have one of their cheaper devices, the Zen Tour. It has just about this amount of filter ripple. No resampling issues for sure. Now I doubt it has the same internals, but maybe the filter is the same.

 

[zhora]

Hi,
...

Thank you very much for testing. Nothing to add except to encourage you to test other pro audio gear that is around when you have some more time.

 

[Zmix]

Thank you for posting this. I've been waiting to see some actual measurements on the Amari.

I wonder if it would be possible for you to test it using the same parameters Amir uses for all of his tests here (4v out, etc) so that we can understand the measurements relative to the hundreds of devices he has tested? Changing the operating level should be possible in the Antelope App under the "Trims" button.

I have been replacing all of my converters in my studio due to forced obsolescence (thanks Apple) and have discovered that all (except two of the manufacturers) go into clipping as the signal approaches 0dBfs, so this result of yours was interesting as most ADC / DAC are worthless at those levels.

I would expect the Amari to have considerably more headroom, and the paralleling of the 8 DAC sections means that each are fed with an attenuated level to make it immune to overs due to inter-sample peaks (at least that what Benchmark do in their DAC 3) .

 

[zhora]

I have been replacing all of my converters in my studio due to forced obsolescence (thanks Apple) and have discovered that all (except two of the manufacturers) go into clipping as the signal approaches 0dBfs, so this result of yours was interesting as most ADC / DAC are worthless at those levels.

Which ones?

 

[Zmix]

Which ones?

All of the professional multichannel converters I have here exhibited this behavior except the SSL XLogix Alpha Link Live-R, which maintained it's ENOB from -102 to 0dBfs

 

[LuminousMoth]

amazing work! really interested to see more pro audio gear measured !!

 

[Jimster480]

Sounds like a problem many of us wouldn’t mind having.

Thanks for the review!

I know right.
I wish I had an APX555... wish there was even budget for one haha