Generic Budget USB to AES Converter Review

[August]

Has the dynamic range decreased as well?

 

[Multicore]

So yeah, "bits are not always bits." See my video on this (around 9 minute mark):

There ought to be a better way of saying this. In the video you have a sentence, "So that proves that bits are not bits." What are they, then? Bananas?

I suppose this is the result of a rhetorical back and forth in which one side uses a strawman to mischaracterize claims of its opponent; quite a clever one in this case, the tautology "bits are bits", inviting the opponent to deny it and declare a nonsense to be the case. Maybe a response could be "Bits are indeed bits but they can be lost or late. Let me show you how." I'm sure some here can come up with even better words.

Debating and education are different activities.

 

[patoulol]

xingcore af200
quloos qu02
matrice-x spdif 3

USB Amanero ?​

Interface Numérique USB Amanero / Bluetooth 5.0 vers AES / SPDIF / I2S 32bit 192kHz DoP128 - Audiophonics

Cete petite interface numérique performante prend en entrée un signal Bluetooth 5.0 ou USB pour ensuite sortir au choix en I2S (via HDMI), AES, Optique Toslink ou Coaxial. La puce Bluetooth utilisée pour la réception est une Qualcomm CSR8675.

www.audiophonics.fr

OR

Topping D10 Balanced and​

​

pro snake AES/EBU SPDIF Cable 3

Câble S/PDIF Certifié numérique 110 Ohm, Fiche XLR femelle vers RCA mâle, Convient pour les interfaces RME, Longueur: 3 mètres, Couleur: noir, Fabriqué en Allemagne

www.thomann.de

 

[patoulol]

Pity there's no listening test. I woulda bin interested in Amir's description of the subjective impression.

Is this jitter actually audible?

I was the one who sent a USB AES Converter to @amirm because I use it with my genelec and I wanted to know if it was good or not to measure! To answer you, I've never heard a problem though, sometimes, my right speaker loses the signal and I have to restart my PC, probably a locking problem on such a bad signal.

 

[unpluggged]

Now that we have this jittery converter, it would be interesting to see how RME's SteadyClock (FS) would deal with it. AFAIR, @amirm has both the ADI-2 DAC FS and the ADI-2 Pro (non-FS), so it would be useful to compare them and to have an independent and real-world test in addition to the RME video that demonstrated their jitter rejection capabilities.

 

[dtaylo1066]

I'm getting a bit jittery reading Amir's review. Waste of dinero.

 

[amirm]

So, all the noise resulting from inserting this gizmo into the digital audio chain is below -120 dB according to your plot.

That's not the actual noise floor. There is 35 dB of FFT gain in there. So actual noise floor is -85 dB or so -- pretty close to computed SINAD.

 

[phoenixdogfan]

If you absolutely need to convert USB to AES/EBU, maybe use something like a Topping D10 to first convert it to Spdif and then a Canare barrel transformer to convert the Spdif to AES/EBU?

 

[patoulol]

Of you absolutely need to convert USB to AES/EBU, maybe use something like a Topping D10 to first convert it to Spdif and then a Canare barrel transformer to convert the Spdif to AES/EBU?

View attachment 417512

On a distance of less than three meters it's not necessary, right?

 

[GXAlan]

So, all the noise resulting from inserting this gizmo into the digital audio chain is below -120 dB according to your plot.

Amir, are you claiming that this will be audible through speakers?

When you do the Fourier analysis, you break it up into smaller chunks. The smaller the chunks, the cleaner the harmonics can be identified, especially in the bass area. Think “resolution” of the graph. But that means for a given noise level, you really break it up into smaller chunks so it makes the line look lower.

Thats the “FFT length 32K” that you see in the right side of the image. The averaging of three runs also makes the noise lower

When hobbyists like me use Multitone to replicate some the APx555 measurements using a E1DA Cosmos or similar consumer ADC, you will see that we have to do two things to “simulate” what you would get from the AP. I will usually use 256K FFT and 8-16 averages.

Multitone is nice because if you look at the legend, it will give you the noise level in dB and dBa to translate the visual line of the noise to the actual noise produced by the device.

 

[pma]

Is this jitter actually audible?

That “jitter” spectrum in post #1 is same as if you analyze 16-bit J-test original wav signal by Julian Dunn. This should have been evident to the reviewer. Same for the resulting THD+N, 16-bit resolution. There is a bug in the test somewhere.

 

[Vive la dissonance!]

Sort of. We still have no idea what gravity, dark matter, and dark energy are. But I get your meaning.

The point is, while mysteries remain, using the scientific method is the only method by which we can hope to arrive reliably at the truth in solving them.

 

[milosz]

That's not the actual noise floor. There is 35 dB of FFT gain in there. So actual noise floor is -85 dB or so -- pretty close to computed SINAD.

I hear what you are saying, but why doesn't the plot show the actual noise level and not something 35 dB lower than what's actually there? Are the plots meaningless and only the SINAD usefully describes what might be audible?

 

[GXAlan]

I hear what you are saying, but why doesn't the plot show the actual noise level and not something 35 dB lower than what's actually there? Are the plots meaningless and only the SINAD usefully describes what might be audible?

It’s tricky because math is math. The peaks for harmonic distortion are real and you need the high FFT numbers to see those peaks. The NOISE is “meaningless” when comparing ASR vs someone else’s measurements (like mine or @VintageFlanker) but the NOISE between two different ASR reviews is accurate for comparison.

 

[amirm]

I hear what you are saying, but why doesn't the plot show the actual noise level and not something 35 dB lower than what's actually there? Are the plots meaningless and only the SINAD usefully describes what might be audible?

Watch this video:

 

[GWolfman]

So yeah, "bits are not always bits."

Bits are bits, but if you change them then… “These aren’t the bits we’re looking for. You can go about your business. Move along.”

 

[NielsMayer]

If you absolutely need to convert USB to AES/EBU, maybe use something like a Topping D10 to first convert it to Spdif and then a Canare barrel transformer to convert the Spdif to AES/EBU?

These passive units might be ok going to 96k but not 192k. At least that's been my experience with the equivalent units from Neutrik. I now use a matrix xspdif-3 for a proper AES/EBU output without needing stupid hacks like these transformers.

 

[NielsMayer]

I was the one who sent a USB AES Converter to @amirm because I use it with my genelec and I wanted to know if it was good or not to measure! To answer you, I've never heard a problem though, sometimes, my right speaker loses the signal and I have to restart my PC, probably a locking problem on such a bad signal.

It is basically an admission of crime to stick this POS in front of your Genelecs. You should give me your Genelecs as you aren't taking proper care of them and I will. And then I won't need to fix the broken amplifier on my 30y.o. 1030As which are gloriously all-analog and DSP-free.

What's interesting here is where are the crystal oscillators? Doesn't look like 'CB1' is an oscillator, but then again, they sanded the chips so you can't read the part #. If lacking a crystal oscillator, I bet that's why this converter produces such horrible results. Wonder if it's timed straight out of the USB->Digital converter? Are there chips on both sides of the board?

And what's with the ridiculous numbering, e.g. "R108" and "C77" -- it's not like there's 107 other resistors, or 76 other capacitors on the board.

 

[patoulol]

These passive units might be ok going to 96k but not 192k. At least that's been my experience with the equivalent units from Neutrik. I now use a matrix xspdif-3 for a proper AES/EBU output without needing stupid hacks like these transformers.

The human ear can't tell the difference after 44kHz, right? unless , you are a mastering engineer who works with EQ files

It is basically an admission of crime to stick this POS in front of your Genelecs. You should give me your Genelecs as you aren't taking proper care of them and I will. And then I won't need to fix the broken amplifier on my 30y.o. 1030As which are gloriously all-analog and DSP-free.

What's interesting here is where are the crystal oscillators? Doesn't look like 'CB1' is an oscillator, but then again, they sanded the chips so you can't read the part #. If lacking a crystal oscillator, I bet that's why this converter produces such horrible results. Wonder if it's timed straight out of the USB->Digital converter? Are there chips on both sides of the board?

And what's with the ridiculous numbering, e.g. "R108" and "C77" -- it's not like there's 107 other resistors, or 76 other capacitors on the board.

Yes, I wanted to try it! there I ordered a Topping D10

But I don't listen to music at more than 44khz or 48khz

 

[darth]

This is the one I got, this is the usb-c/BT version of the above I would guess.