• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Recommendation for a USB Audio-interface DAC/ADC for ADC performance

Buckster

Active Member
Joined
Dec 2, 2019
Messages
110
Likes
54
using USB DAC/ADC/Audio Interface with a Raspberry Pi 4B - as a BirdNET Pi (it detects and automatically identifies birds by their song/calls)

I am using a Clippy EM272 Microphone via TRS-TTRS converter


I was first using a 96/48 Audio interface - Ugreen - Ugreen 94khz/24bitwhich seemed OK but gain seemed lower than I would have hoped

I accidentally broke that one so replaced it with a 384/32bit - Ugreen 384khz/32bit - one which I hoped might give out more voltage on the signal to the mic - but if anything (hard to directly compare) the voltage and gain seems a bit lower

Question I have please is that the CM272 has I think a voltage range of 2V-10V - 2V being about the minimum

I literally can't find the output voltage to the mic listed on any of the USB Audio interfaces/DACs/ADCs, just distortion, power to Headphones etc - so for my usage (I'm not even using the DAC, only the ADC) it is very difficult to find what I am looking for. Any recommendation please ? or should I go for something like the cheapest Focusrite I can get (USB powered - hopefully will be able to be powered from the Pi)

many thanks for any help
 
Last edited:
The EM272 works just fine below 2V Plug-in Power.

Here's how much Plug-in power various ADCs provide to my Primo EM272:
Screenshot_20240525-184735_Sheets.png

As soon as I have some quiet moments, I'll also measure frequency response and noise with each ADC.

I'll post the results here, though I don't expect much correlation to Plug-in Power personally.
 
thank very much for the reply and the chart of the power for various ADCs - very interesting and much appreciated - those are higher readings than I was expecting

unfortunately the 96/24 Ugreen I had - got a bit damp after being left out for a long while and stopped working so I can't test A to B - but subjectively the new USB-C 384khz one seems to have less gain with the EM272. I went for the 384khz version as thought for a few pounds more it might be lower noise etc

TBH I've found the EM272 (it's actually a EM272M as I went for the one with additional protection against interference) with both DACs/ADCs not as sensitive as I was expecting - I have the gain maxed out in alsamixer.

The audio path I have is as follows:

Pi4B - Ugreen USB-A to C converter - Ugreen USB DAC/ADC - Rode TRS to TRRS adapter (TRRS at Ugreen end) - Clippy EM272M

Previous to this I was using a UMIK USB Mic (audio calibration one) - which has higher noise and is meant to be less sensitive but if anything I found it more sensitive. I stopped using that as was hoping for even higher sensitivity and also I didn't want to leave it outside for months on end and get damaged

as a keen birdwatcher and nature enthuiast though - the BirdNET Pi has been a complete revelation to the birds in my neighbourhood

I did find 2 other USB-C DACs/ADCs floating around which I tried testing, one was a Huawei one which came with a P20 Pro, the other unbranded not sure where that one came from, but on plugging both into the Pi 4b neither were detected either for playback or recording (I re-booted went into alsamixer and not even listed as a device)
 
I played a very quiet test tone at constant SPL, then recorded it with my EM272, with each ADC set to maximum gain.

Here's the result:
Screenshot_20240526-031733_Sheets.png

When you normalize the recordings in post, here's how I would rank the ADCs in terms of cleanliness/SNR:
ADC Comparison 1.png
ADC Comparison 2.png
ADC Comparison 3.png
ADC Comparison 4.png
ADC Comparison 5.png
ADC Comparison 6.png
The SXFI is not shown, as it didn't play nicely with my recording software (USB Audio Recorder Pro).

If you're able to apply a volume boost in software, then I'd primarily focus on the FFT graphs.
 
Last edited:
thanks very much indeed for those graphs - much appreciated.

I can apply a boost in software to manually listen/check identification correct (I use Audacity) - but in the chain between the recording and it being analysed I've not yet found a way to boost

looking at all of your graphs, although not the best cleanliness/SNR I might give one of the USB-C Apple dongles a go (I had forgotten there was a USB-C version) - for £8.99 might be worth a comparison/have as a spare

I will have a look through some of the cleaner/better SNR ones on your graph too

many thanks and much appreciated for the info
 
looking at all of your graphs, although not the best cleanliness/SNR I might give one of the USB-C Apple dongles a go (I had forgotten there was a USB-C version) - for £8.99 might be worth a comparison/have as a spare

I will have a look through some of the cleaner/better SNR ones on your graph too
The Apple dongle had one of the cleanest recordings in my test (graphs are sorted worst to best).

Only the UMC22+Antlion adapter was slightly cleaner below 100Hz.
 
thankyou and oops I was reading the graphs the wrong way round !

quick question please, as I live in the EU, presumably the EU output limit would only apply to the headphone out and not any voltage sent to a mic ?

many thanks
 
quick question please, as I live in the EU, presumably the EU output limit would only apply to the headphone out and not any voltage sent to a mic ?
the voltage sent to the mic from both Apple dongle versions (A2049 and A2155) can be seen in post #2.
 
the voltage sent to the mic from both Apple dongle versions (A2049 and A2155) can be seen in post #2.

ah sorry didn't realise the versions indicated US vs EU I thought there were US/EU versions of each - thankyou

ref graphs and distortion - oops I was reading them the wrong way round !

have ordered the Apple USB-C to 3.5mm I'll let you know how it compares - should arrive tomorrow

many thanks for the help
 
have ordered the Apple USB-C to 3.5mm I'll let you know how it compares - should arrive tomorrow

many thanks for the help
Keep in mind that you may have to buy a CTIA Headset splitter to plug the EM272 into the dongle.

If your mic uses a TRS 3.5mm plug, then it's can't be plugged directly into the dongle.

I believe Micbooster also make "Smart" versions of the Clippy/Pluggy, and these can be plugged into the Apple dongle without splitter.
 
thanks for the heads-up

I use a TRS to TRRS converter (Rode SC4) - I will look-up to see if I also need a splitter thankyou
 
thanks very much - can confirm works fine with just the SC4

have only just plugged in - but the gain/sensitivity seems to be much much greater than with the other adapter !

and alsamixer only set to 48/100 before I was maxed out at 100

I am very surprised by how huge a difference there is

I think the other adapter must have not had enough volts to drive the EM272

in fact I seem to be getting much more gain than the 96/24 Ugreen adapter I was using too
 
Awesome! I'm glad that you found an inexpensive solution.

I've been using the BirdNET App for a few weeks now, it's a fantastic piece of software IMO.
 
the BirdNET app is fantastic - via the BirdNET Pi put at the bottom of garden it has completely revolutionised my view of the birds visiting and flying over the garden (I'm a very keen birder and the results have really surprised me)

even just raptors, already have recordings of Kestrel, Sparrowhawk, Barn Owl, Tawny Owl, Red Kite and Marsh Harrier - I've never even heard a Marsh Harrier call (with my own ears)

I've had to bring in for now as although the mic has a Rode WS9 Windshield on it which does protect it from some rain - the rain is heavy enough tonight for me to be cautious so I've brought it in - but from initial tests the gain/sensitivity of the Apple Dongle seems so much higher than either of the Ugreen devices and is giving me the sort of sensitivity I was expecting from the EM272 (was a bit disappointed when I first used the EM272)

thanks very much for the help, I'd have never even have thought about an Apple USB-C to 3.5mm
 
happy to say detection sensitivity much much better with the Apple dongle thankyou

one question though please have you ever seen this pattern of "noise" - any any idea where from ? Its odd - I have seen horizontal lines (but completely straight lines) before with the Ugreens but not this pattern - some aren't even straight lines (as in don't start and end at quite same frequency - for instance one starts just below 12khz but ends at 12khz. Some have "waves" which makes it look more like interreference pattern

I tried a USB extender to move the Apple Dongle away from the Pi but made no difference

doesn't make a notable difference to the bird recordings - I was just curious what could be causing it

example-2.PNG


House-Sparrow.PNG
 
one question though please have you ever seen this pattern of "noise" - any any idea where from ?
Yes, the Apple dongle has some inherent interference noise on its input, which you can also see in the graphs I posted before:
ADC Comparison 5.png

I don't know where it comes from.
 
Yes, the Apple dongle has some inherent interference noise on its input, which you can also see in the graphs I posted before:
View attachment 371791

I don't know where it comes from.

ah yes that makes sense thanks for the reply

so far the results have been very good with the dongle (thankyou) - reference the noise it seems variable - I was thinking gain related but I am now wondering if it is also temperature dependant as it seems (not sure yet) it might be worse in the day than at night

gain was higher in the "Early" screenshot than the "Afternoon" one - but temperature also much higher in afternoon one

or it is signal related but interesting such a marked difference
 

Attachments

  • Afternoon.PNG
    Afternoon.PNG
    443.6 KB · Views: 64
  • Early.PNG
    Early.PNG
    481.2 KB · Views: 62
10 minutes apart, maybe the Audio Interface is picking up some external interference - find it hard to believe it would cool down in 10 minutes


Lines.PNG
 
Back
Top Bottom