Suggestions needed for software, interface etc.

[Tubetec]

The SSL2 controler/interface accepts levels up to +24dbu on the line level inputs ,
192 euros .so you'd need to add something small to the order to get free shipping from the .DE supplier .
The package gets you the use of a couple of channel strip plugins on the host pc as well as recording and production software .

 

[john12ax7]

RMAA is not capable of using the ASIO driver, so measurements at 192k are not possible.
Audiotester works with the ASIO driver and gives good results, after a lot of tweaking, that it looses when quitting. 

Have you tried ARTA? Works with ASIO at 192k.

 

[abbey road d enfer]

Have you tried ARTA? Works with ASIO at 192k.

I did. I get no useful results. Freq resp is +/-10dB and I can't get it to go beyond 20kHz.

 

[john12ax7]

I did. I get no useful results. Freq resp is +/-10dB and I can't get it to go beyond 20kHz.

Maybe a setup issue somewhere, perhaps it's reading an errant cal file? If you right click on the graph you can change the axis.  Here's an FR1 sweep of my setup at 96k and 192k

 

[abbey road d enfer]

Maybe a setup issue somewhere, perhaps it's reading an errant cal file? If you right click on the graph you can change the axis.  Here's an FR1 sweep of my setup at 96k and 192k

OK, got it to work all right with ARTA. Shows -3dB exactly at 60kHz.

 

[ruffrecords]

Mine is due to be delivered this morning.

Cheers

Ian

 

[abbey road d enfer]

Mine is due to be delivered this morning.

Cheers

Ian

Have fun!

 

[pvision]

Could be wrong, but I thought all modern interfaces worked with Mac OS via Core Audio with no third-party drivers needed?

 

[ruffrecords]

Have fun!

Got it set up last night and tested at 192K with REW. I can confirm the bandwidth is 3dB down at about 63KHz. Thanks for the tip Abbey. This is now a permanent part of my test set up.

Cheers

Ian

 

[Tubetec]

For music were not too bothered about much above 20khz , but in the case of measurement a flat response out to 90khz or more would be an advantage.  When we see 63khz -3db @192khz , is that the effect of a LP filter in the input stage ?
If there's a set filter it might not allow maximum bandwidth at 192khz as it might be optimised for a lower sample rate.
Would there be any sense to trying to change filter components in an A/D stage to better suit  measurement  app's  where we want as much  bandwidth as we can get ?

 

[abbey road d enfer]

For music were not too bothered about much above 20khz , but in the case of measurement a flat response out to 90khz or more would be an advantage.  When we see 63khz -3db @192khz , is that the effect of a LP filter in the input stage ?
If there's a set filter it might not allow maximum bandwidth at 192khz as it might be optimised for a lower sample rate.
Would there be any sense to trying to change filter components in an A/D stage to better suit  measurement  app's  where we want as much  bandwidth as we can get ?

After reading a number of converter datasheets I came to the conclusion that the HF response is limited by one of the digital filters.For example CS5361 specifies the -0.1dB point at 0.24Fs @ Quad Speed, i.e. 46kHz @ 192kHz SR.

 

[Gold]

Has anyone considered an evaluation PCB? They come with connectivity and I think  you can set most parameters with them. You don’t need the latest and greatest so I’d imagine an A/D and D/A could be had for the same or less than a budget interface.

You usually have to hard set the parameters via software but for a dedicated measurement rig it would be okay.

 

[Tubetec]

A calibrated audio probe with usb conectivity is something Ive been thinking about a while ,
wouldn't it need a dedicated Asio driver ? 
Could dsp processing/eq  on the incoming signal be used to tweak out extra resolution ?

 

[john12ax7]

One option could be a switch to bypass the filter altogether.  Very little content typically above 96k to worry about.

 

[abbey road d enfer]

One option could be a switch to bypass the filter altogether.  Very little content typically above 96k to worry about.

You can't switch off an embedded digital filter.
Some chips allow a choice of filters, though. One of the most recent Cirrus/Wolfson D/A offers a choice between 5 different filters, only one extends the response close enough to the Nyquist frequency. All the others cut at about half of it.

 

[abbey road d enfer]

Has anyone considered an evaluation PCB? They come with connectivity and I think  you can set most parameters with them. You don’t need the latest and greatest so I’d imagine an A/D and D/A could be had for the same or less than a budget interface.

You usually have to hard set the parameters via software but for a dedicated measurement rig it would be okay.

Apart from the inconvenience of having to initialize the converters, one has to manage the digital flux from and to the S/PDIF or AES/EBU connections to the computer. The nominal audio levels from and to the converters need to be addressed too; it's a common need with most bus-powered soundcards though.
In terms of economics, an eval board costs about $300, to which one has to add a PSU, an S/PDIF to USB converter, that would exceed the cost of a budget interface.

 

[john12ax7]

You can't switch off an embedded digital filter.
Some chips allow a choice of filters, though. One of the most recent Cirrus/Wolfson D/A offers a choice between 5 different filters, only one extends the response close enough to the Nyquist frequency. All the others cut at about half of it.

I don't think this applies to all chips,  I remember reading some data sheets that offered this provision to allow for designers to design their own digital filters.

Edit: just checked a pcm1792 datasheet I had,  there  are control signals for digital filter enable / disable.

 

[Tubetec]

There is a cheap USB C dac's with mic input , ALC4042 ,

https://www.ebay.com/itm/USB-C-Connector-With-ALC4042-Chip-312bit-384khz-Dynamic-Armature-Type-C-Male/153843432679?hash=item23d1c87ce7:m:mYI6C7Gi20s3x0xrVUrlaNw

It has a headphone amp and mic input with convienient solder pads ,
I couldnt seem to dig up much about the chip itself , there must be software that lets you tweak the parameters ,                        Could the mic input be reconfigured as measurement input with appropriate bal/unbal level conditioning as required ?
ALC4042 seems to have a predecessor 4040 still not much info on it though.

 

[Tubetec]

While looking around at USB C  headset DACs I found a few interesting products with superb performance for small money.

https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-e1da-9038s-bal-portable-dac-amp.8424/

Noise 122 db down , 85khz -0.2db, thd&N 0.0002%  ES9038Q2M/AD8397 chipsets

https://e1dashz.wixsite.com/index/9038s              makers site .

It has a pair of balanced and floating outputs on a 4 pole mini jack , it will deliver 0.5w into your phones .
That basically means a ground is not shared by both amps neg signal terminals like headphones with 3 pole stereo mini jack . It seems likely to me driving a balanced line input should work fine with it , it does also say not to ever use it single ended without a transformer !

Next up is the Meizu Dac , it utilises Cirrus Logic CS43131 ,the pro version adds  dedicated TI OPA16op amps ,

https://www.audiosciencereview.com/forum/index.php?threads/meizu-hifi-pro-usb-c-phone-dongle-dac-review.10423/

The spec doesnt change much between the pro and the standard version ,in fact the standard wins in some respects like noise and crosstalk.

Another unit with slightly less good spec is BGVP-T01 ,it sounds like a virus I know 
It has a Quallcom WHS9410 which  has an ASIO driver for it ,


So the price,
The E1DA                $80
Meisu Pro                $50 
Meisu                        $25
BGVP-T01              $30

Compatibility with legacy os and hardware  does seem to vary , but some do work fine over Usb2 adapter.
Some of the higher end chipsets incorporate dsp processing for things like multi mic noise cancelling or microphone beamforming arrays. The Quallcom WHS9420 has three  inputs with programable gain and dsp  processing , noise levels are a little high at around -100db on the inputs .
Each manufacturer must have ,for in house use at least a way to flash or program the dsp . Maybe the stupidly expensive  development kit for a particular chipset comes with the programing tool  , or maybe a copy exists in the 'wild '.

The other great thing about these 'soc'  dac/adc/dsp/usb chips is size , you could easily solder one of the smaller boards into a jack/xlr/rca plug shell ,or virtually any other electronic project you wanted to hook to a computer . The 29.99 Meizu looks like a great bargain for someone who already has their own  output stage chipsets or config the like to use , the E1DA's 80$ pricetag for a fully balanced outputs with that spec is unbelievable , Maybe seeing as Quallcom offer an Asio driver for their chips the software is more acessable too , maybe some of the more software orientated will let us know more about programming these.

 

[Tubetec]

A new version of Ivan's  E1DE 9038S  is due in any time , it will include an android app thats able to tweek the dac settings via usb, it also has improved performance over generation 2 ,lower noise , lower distortion and wider bandwidth , at 384khz he quotes 5-160000hz +0.1/-3db .