looking at some MOTU interfaces in Spectrafoo Complete

I guess this is the right place to put this.

New MOTU 16A compared to the old PCI line 2408mkIII.

With the adjustable inputs and outputs on the 16A, we have a question around 'unity'.

16A inputs are referred to as digitally controlled analog trim.
16A outputs are referred to vaguely as probable digital trim before DA.
16A outputs are roughly 4 dB hotter than 2408mkIII when set to 0.
Not sure about the inputs yet.
Spectrafoo balanced loop test, 16A output has to be set 0 and input has to be set +2 to get +0.26dBr difference WHEN COMPARING INTERNAL SIGNAL GENERATOR TO RETURNING LINE.
As a result FS output from internal signal generator will clip input at +2 input setting, not at +1.
I still need to measure ext sig gen AD DA with all set at 0, see what results. EDIT: see below....

I have for now set the outputs at -4 so I see something comparable to what I got with the PCI interfaces.
With -4 output, you'd need +5 input to get close to unity loop-through without clipping a FS source.

From Audio/MIDI on a Mac, the 16A is described as 32 bit floating, whatever that's worth. I haven't looked at an actual 32 bit source in Spectrafoo to examine the bitstream yet.  EDIT: see below....

Noise floor comparisons of MOTU 16A and MOTU 2408mkIII, scale is 0dBFS relative. 
192K/176K/96K/44K. Lower group is the 16A.

This one more clearly shows the PSU related noise in the 2408mkIII.

Phase comparisons. Upward curves are 2408mkIII, downward are 16A.

16A responses and phase at 192K/176K/96K/44K.

Upper frequency response comparisons of MOTU 16A and MOTU 2408mkIII. 176K isn’t shown.

Lower frequency response comparisons of MOTU 16A and MOTU 2408mkIII. Oranges/Yellow/Red are the 16A, white is the 16A at 192K, Blues/Green/Purple are the 2408mkIII.

If I were picking the best rates based on phase with the 16A, I'd choose 88.2K or 192K. 96K and 176.4K aren't as phase linear, but still clearly beat the snot out of 44.1 and 48.  I tend to work at 88.2K most of the time already. 

Some more numbers:

600 ohm output signal gen sending +24dBu reads 23.7dBm at outputs with multimeter, -0.25 dBFS incoming to 16A set input trim 0.
16A output set 0 (max), output reads +21.7dBm with multimeter.

Impossible to get same analog output level as analog input level without DSP gain. Restated, the input has greater headroom than the output. This agrees somewhat with the manual specs, I'm wondering if there's a typo there since I am seeing higher output levels than the listed max.

16A noise numbers as seen on the plots above:
60Hz (AC line) -147dBFS
125Hz -154dBFS
250Hz -151dBFS
715Hz spike about -140dBFS  (certainly looks higher on the plot, but zoom in confirms)
1kHz -147dBFS (note the manual states in or out THD+N of -110 dB, -1dBFS, unweighted, 1kHz)
4kHz -143dBFS
16kHz -137dBFS

Spectrafoo code stream appears to be 24 bit, with Audio/MIDI describing 32bit float and DP8 set for 32bit float recording.

That's very interesting emrr - thanks for those measurements.

I have some 2408Mk3 and see similar (out to 44KHz) noise profile - my setup is calibrated for the dBu measurement, albeit a little crudely given the resolution of my mmeter and cro.

I also get a difference in 'received in software' dBu level for a 'loop back' test.

ie. I calibrate the 'ins', I calibrate the 'outs' - each is verifyable externally.
 
    then on the 'loop back', the 'received in software' dBu for a given output is around 1.4dB difference

So, it's very interesting to see your measurements and so on.

ps - my max 2408MkIII input level at the onset of distortion is around +10dBu I think ( I have it all in my notes).

CHeers

I recall my 2408mkIII loopbacks being closer to unity, but not perfect.  Here I was paying more attention to the 16A as a new product. 

Doug,

How are you finding the 16A?  Sonically, stability, latency?

I'm trying to help a friend with a very limited budget update his very convoluted and aging Pro Tools setup (888.24 on a G4 synced with a Mac Mini for VIs!).

I'm hoping I can get him into something like a new Mac Mini, 16A and Pro Tools 11. He has a 32 input Audient and being old skool prefers to mix there so the aim would be to run two 16a side by side as soon as possible.

I'm looking at lots of options for him - Rednet etc.  His audio needs are 8i, 24o at 48k max but he is a heavy virtual instrument user so getting him to 64 bit Pro Tools 11 would be great.

Cheers,
Ruairi

Sonics are great compared to the old line. I don't have anything else to compare to.  Much clearer and deeper into the low level well, a more highly polished magnifying lens. 

Stability is great, I can't make it fart.  I've had audio streaming between multiple computers tied together with word clock, and can unplug TB or USB to the 16A if it's solely routing control with no disturbance, likewise I can re-patch live routing on the fly while playing audio with no glitching. 

I've always monitored using 'direct hardware playthrough' in DP8, so I'm unaccustomed to any detectable latency.  That mode does not currently function in DP8, which I'm unhappy about.  I don't know what PT does with monitoring.  I'm trying to drive the control app like 'direct hardware play through' with analog monitoring, and it's a pain to switch back and forth, overdubs are a bit of a pain in the ass for the way I work.  These are likely problems others are accustomed to dealing with, but it's new to me.  The control app with built in mixer was borderline beta, and they've issued two updates already, more additions to come hopefully.  With my typical session template I can't comfortably run buffers any lower than 256, on a late 2012 3.4GHz i7 iMac with DP8 in 64bit.  I don't really know what all factors into buffer requirements for a given system.  I'm not anywhere close to hammering CPU or RAM. 

I could monitor with the built in mixer, but then I'd be building the same mix twice in parallel.  I'm always building mix elements while tracking, and printing live rough mixes with analog processing while tracking.    The latest update claims:

round-trip latency—now down to 135 samples at 32 sample buffers over Thunderbolt.

Click to expand...

There's probably a few more inventive ways I could route monitoring I haven't thought through yet, combining the built in mixer for live OD tracks with discrete multitrack playback to analog world. 

Great info Doug, thank you for taking the time.  Sounds promising for my friends set up.

Hi Doug,
Thanks for sharing this info.

Is this all on a Mac? I'm hoping the 16A will work on a Win7 machine in the near future so I can replace my pair of 896HD mkIIs.

Mac.  No telling when and what Windows systems. 

Thanks for confirming.

I have been hoping that a Windows driver will be offered soon. I'd like to go with a pair of 16As.