Holo Cyan 2 + xDuoo TA-66 system distortion surfaces

It would be a reasonable assumption to consider combining individual component distortion measurements for a system result. In this particular system the TA-66 distortion might be assumed to be dominant with Cyan 2 having negligible contribution. However the following measurements demonstrate system performance can differ from that assumption.

Individual measurement setup and conditions for the two components discussed may be found here:
Holo Audio Cyan 2 technical measurements
xDuoo TA-66 technical measurements

Measurements for the specific tube used may be found here:
Five 6AS7g type tube comparison

System specific setup:
1) TA-66 volume adjusted to output +10 dBu for Cyan 2 at 0 dBFS
2) 300R load

The following graphs will be presented in the same order for each distortion surface:
1) Cyan 2
2) TA-66 with Thomson-CSF 6080WA output tube
3) System measurement
4) animation of 1, 2 and 3 above


Very low distortion.


It would be tempting to dismiss any contribution of the Cyan 2 THD+N to the system result.


However the system result demonstrates a difference, though small on THD+N. D2, D3 and 4+HD+N below will show larger differences.


animation for convenient visualization

Pay particular attention to the surfaces between -30 and -15 dBu where the system represents approximately 70 to 85 dB SPL for HD6x0 and HD800 headphones.






D2 is definitely impacted at low frequencies.

 

D3 shows the largest system difference from individual component measurements.









4+HD+N is less impacted, closer to that of THD+N.



The above differences suggest the importance of considering system level measurements over assuming component level measurements will simply combine.

It also suggests that further investigation may provide insights into system synergies.

 

Humble request...

Could you make static visualizations for configs 3 (entire system) minus configs 2 (TA-66 alone) for each distortion component as well? Probably best just as 2D colour maps so the surfaces don't get in the way and obscure the dBu vs frequency relationships, with an added colour-bar (which could be re-scaled to highlight the small differences).

 

This is a one-off as I don't have automated processes for surface differences, yet. All manual today.
These are the TA-66 surfaces subtracted from the Cyan2+TA-66 System surfaces to produce the differences.




As the D2 differences were greater than my default scale it was necessary to change Distortion dB scales for a better view as shown below:



D3 impacts are greatest in the range of -30 to -15 dBu

 

I think this looks great. Can someone please translate this if I'm 5 y.o.? Sorry just need FAQ or something...

 

Essentially, even though it's "common sense" that DAC differences should not make for a very obvious change in distortion (this is the extra noise you're adding to the stuff that's already there! [You did specify age 5 okay back to normal speech:]) because of how most DACs just measure ruler flat with negligible added harmonics (unless something is VERY broken), there are changes that might actually be easily audible at the transducer level because the magnitude of difference in 2nd and 3rd harmonics are gross enough.

Also, because these are changes to the harmonics closest to the fundamental signal, it can explain why there's so much character change when playing with different systems. Still need to investigate this as it could go a ways towards explaining why some gear might pair poorly or pair exceptionally well, which so far is only ever fairly evaluated on a qualitative basis.

 

@atomicbob This is a VERY low priority request, but I think it'd be really cool to see what the result may be with other DACs into the TA-66, or any other amp. Would be very fun I think to compare the degree of changes between known-nuanced amplification and others that follow the high global NFB school of design, see whether that might explain in part how the latter is often more insipid and disengaging to listen to.

Very rudimentary but I got some upper treble differences at the transducer level using Magni Unity and HD600 on MiniDSP EARS (comparing Modi Multibit 2 OS and NOS mode, and the internal ES9018 DAC in the Unity). Doubt this fully encapsulates all the audible differences between them but it was intriguing that the change was that gross.

I am well aware that there's already a lot of stuff queued up at the lab and you'd (rightfully) much rather be spending time with the cats and nature, cannot emphasise enough how low priority the request is, but might be fun if there's time. I'd try to do more of this myself if I had the equipment and a quiet enough space for it, bah.