Suggestions needed for software, interface etc.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Winston O'Boogie said:
The M2?  Scan in the UK says it's 96K but the MOTU site says 192K

Yup. Just saw that myself.
Sweetwater needs to update their site.
Still not seeing anything about frequency response.
 
Motu M2 User Manual
https://cdn-data.motu.com/manuals/usb-c-audio/M_Series_User_Guide.pdf

Frequency resonse and other info on page 25.
 
lsburden said:
Motu M2 User Manual
https://cdn-data.motu.com/manuals/usb-c-audio/M_Series_User_Guide.pdf

Frequency resonse and other info on page 25.

OK thanks for hunting that down.
It seems that the M4 has direct to converter line ins with less deviation than the combi inputs on the M2 but nothing is stated beyond 20KHz so who knows?  😜

 
Any practical probe that allows the interface input to connect to a DUT with more than a few volts of signal requires the probe (eg. 10:1 or 100:1) to be compensated for the interface input impedance. 

I found the EMU0404 interface has about an 82kHz corner with 10:1 and about 40kHz (I recall) with 100:1 cheap scope probes and using the 1Meg un-balanced input.  But that is raw bandwidth, and using a calibration file can extend to eg. REW limits of 2Hz and Fs/2, but with an awareness of how much the calibration file is modifying the raw signal, and that twiddles will typically occur in those shoulder regions.  And then there is diy tweaking that could be done - eg. to extend LF bandwidth - which seems to happen for cheaper interfaces where diyers want to enhance what they have.
 
I was wondering about MLSSA , its a bit of an old timer , I remember in the good old days of 'Studio Sound' magazine , no equipment review was complete without MLSSA graphs.

http://www.mlssa.com/

Its still around by the looks , and the later hardware seems quite capable ,even by modern standards.
 
Tubetec said:
I was wondering about MLSSA
I don't know any company that still uses MLSSA as its principal measurement tool.  It's been replaced since long by systems that don't need internal PC cards such as CLIO, Klippel and EASERA, or even REW. The fact that their website has not been updated since 2009 seems to testify so.
 
Thanks for the extra info Abbey, Id never even heard of those other softwares before , REW is ticking all the boxes for me just now most especially at  the price.
 
When I'm testing with  REW analysis suite, I use an interface that has 4x  'line input' and 'line output' balanced connections ...

One pair of I-O is for REW  analysis  ..  ie.  a main  'DUT' send and return, and  also  the  'Loop Back Reference' send and return.

The other pair of I-O I use to 'audition' stereo audio, to an amp and speakers  ..  could be a media player -  most of the time [long, long hours at the bench :) ] -    or a favourite intrument channel or so  ..  which I can 'software' patch into the 'measurement' chain when I want  .. for those 'spot' measures.

It helps nicely to have 'a pair' for testing DUT  and 'another pair'  to listen on the speakers ..  while you go .. but completely seperate from the 'signal generator'  at all times  8)

And, apart from that, I do like to use an SPDIF pair too, for digital perspectives and to drive my 'Finalizer' :)

BAsically a 4x4 bal analog with 2x2 digital  ....  if they have a 'mic input' and/or a 'DI' on front, all the better  [for test and measure].

...

As far as the input and output range goes, plenty of cheap m-audio  do balanced  I-O  up to around 10Vpp sine  with really high fidelity as 'a loopback'.    Which is fairly decent, I think. 

Much, much higher fidelity than almost any reasonable cost hardware signal generator box  out there.

For anything I need to drive higher than around 10Vpp, I use a step up 'test bench' transformer  *and/or* a  solid-state amp module  ...  I'm talking 100 dollars for a bunch of  Edcor  [good performance and value for matching transformers] and a few ebay diy 'op-amp' modules,    in the budget [as well as].

-> ie. for various kinds of testing of intermediate level and distribution  amps and so on especially for bal/unbal bs.

I also like the Behringer Headphone Amp 1ru box for this role, in conjunction with the afore-mentioned [said] Edcor{s}  for similar overall money.

They [ebay op amp modules] give you  'balance-debalance'  ..  and 'straight wire gain with Alps fader [quiet and crackle free] ' level control.

These modules let me measure some of  the finer details of things quickly [I have a limited attention span]  and easily [low threshold for the ditchin'].  Again, some Behringer gear fits this 'test bench' role nicely for similar dols.

...

For anything I need to receive higher than around 10Vpp, I use a diy  'psuedo balanced attenuator' box  for the 'single ended' stuff and a basic attenuator for the balanced.

-> ie.  mostly for tube biz of all kinds,  and for the probing [with a suitable safe probe  :mad:]  of course.

If the power is still higher, then I can link in a 'load box' into the above  scenarios  and a solid state power amp module with still very high levels of 'measurement channel integrity'

..

Covers most audio ranges with quite amazing fidelity and a good launching point for the more 'hi resolution' endevours.

My 100 dollar m-audio box  Profire 610 does all the above, in a way that if you accidently 'bork it', it's not too big of a deal. :)
Admittedly, it is old so one needs to consider their hardware platform.

I think I get something like 0.0015% at least, in most testing scenarios  for my best 'measurement channel'  in real life situations.

[firewire in my case - up to fs 96KHz with no probs]

For more modern money, I'd love to try some of the nice ones out there.

They do do an order of magnitude better  for 10x the investment - but mind you don't fry it  with tubes!  :)



 
alexc said:
When I'm testing with  REW analysis suite, I use an interface that has 4x  'line input' and 'line output' balanced connections ...

One pair of I-O is for REW  analysis  ..  ie.  a main  'DUT' send and return, and  also  the  'Loop Back' send and return.

The other pair of I-O I use to 'audition' stereo audio, to an amp and speakers  ..  could be a media player -  most of the time -    or a favourite intrument channel or so.

It helps nicely to have 'a pair' for testing DUT  and 'another pair'  to listen on the speakers ..  while you go .. but completely seperate from the 'signal generator'  at all times  8)

SO - apart from that, I do like to use an SPDIF pair too, for digital perspectives.

...

As far as the input and output range goes, plenty of cheap m-audio  do balanced  I-O  up to around 10Vpp sine  with really high fidelity as 'a loopback'.    Which is fairly decent, I think. 

Much, much higher fidelity than almost any reasonable cost hardware signal generator box  out there.

For anything I need to drive higher than around 10Vpp, I use a step up 'test bench' transformer  *and/or* a  solid-state amp module
-> ie. for various kinds of testing of intermediate level and distribtion  amps and so on

Some solid state 'balance-debalance'  ..  and 'straight wire gain'  .. modules let me get to some of the fine detail for less than 100 dollars.

For anything I need to receive higher than around 10Vpp, I use a 'psuedo balanced attenuator' box 
-> ie. for tube stuff of all kinds

If the power is higher, then I can link in a 'load box' into the above  scenarios  and a solid state power amp module with still very high levels of 'measurement channel integrity'

..

Covers most audio ranges with quite amazing fidelity and a good launching point for the more 'hi resolution' endevours.

My 100 dollar m-audio box  Profire 610 does all the above, in a way that if you accidently 'bork it', it's not too big of a deal. :)
Admittedly, it is old so one needs to consider their hardware platform.

I think I get something like 0.0015% at least, in most testing scenarios  for the 'testing channel'  in real life situations.

[firewire in my case - up to fs 96KHz with no probs]

For more modern money, I'd love to try some of the nice ones out there.
There's no doubt that many current and pastinterfaces provide adequate performance over the "traditional" 20-20k bandwidth. The big issue is finding one that extends to about 80kHz; that would mandatorily be capable of Quad Speed operation. It turns out that several interface boxes capable of 192kHz SR have a frequency response that falls significantly short of the Nyquist frequency. Very often frequency response vs. SR is not specified in mfgrs litt.
Response beyond 20kHz is not required for acoustic measurements, so soundcards recommended e.g. for REW are no indication.
Response beyond 20kHz is needed for electronic measurements because many transformer resonances and circuit stability issues happen in the 30-70kHz zone.
 
I measure audio up to 48Khz in REW flat to less than 0.5dB at around +4dBu with around 0.001% 'headline' thd figures on a sine wave on my REW + M-audio [100 dollar rig] on an old PC laptop no problem, worst case so to speak.

I think it does do up to    fs 192KHz  - but I have to say I've never tried it :)

In REW there are very minor differences amongst the different modes of operation ..  fs 44KHz, 48KHz, 96KHz etc  ..  but very small.  Can't say I've noticed anything more or less than 0.5dB max [at the outside] in REW loopback freq sweeps under any sensible  tests so far. Generally, it's much less like +0.1 / -0.2 dB or so over the specified audio band.

 
Read into it but not a builder  - interesting but I think one could do a more useful set of facilities  imho .

Nice concept - could be updated to be really useful in the post- REW  age.  :)
 
Murdock said:
Has anyone tried this little Thing?

http://www.pmillett.com/ATEST.htm

Sounds quite usefull.

This looks like an ideal tool for my personal needs. Thank you for sharing.

We are building out (at some point in time) a Beis AD/DA convertor. Uwe Beis has been a good friend and very informative source on conversion. The DIY Kit he provides is a about $800, so my wife needs to sign off on this at a later date. But the pmillet design is green lite!

https://www.beis.de/Elektronik/ADDA24QS/ADDA24QS.html

Thanks!
 
Hey Wall,
Yep I've seen the Beis stuff and it looks pretty decent.  At quad sampling (192KHz) he states that frequency response is flat up to 80KHz which is nice.  At lower sampling, the response is of course F/2. 
Can this be bettered using an off-the-shelf built device?  I don't know as accurate specs seem to be thin on the ground.
It should surely be possible that the purchasing power and clout of a big company could get you to the same place at lower cost but then, with cost X 4 being the usual markup for commercial stuff, dunno?

 
Winston O'Boogie said:
Hey Wall,
Yep I've seen the Beis stuff and it looks pretty decent.  At quad sampling (192KHz) he states that frequency response is flat up to 80KHz which is nice.  At lower sampling, the response is of course F/2. 
Can this be bettered using an off-the-shelf built device?  I don't know as accurate specs seem to be thin on the ground.
It should surely be possible that the purchasing power and clout of a big company could get you to the same place at lower cost but then, with cost X 4 being the usual markup for commercial stuff, dunno?
That is a very interesting question. There are loads of interfaces out there, both cheap and expensive ones, with a quoted 192KHz sampling rate but by far and away the vast majority do not quote specific analogue bandwidths at each sampling frequency. In fact in general their analogue specs are woefully inadequate. All the ones I have seen simply quoted a blanket 20Hz to 20KHz..

I am lucky in that I have a test oscillator than go go up to 40KHz. So far I have used this with a current generation Focusrite 2i2 and also a cheapo Behringer UMC202HD. So far I have not been able to get either of them to convert a 40KHz sine wave at any  sample rate.

Cheers

Ian
 
Back
Top