Thanks again!
First I will test it with a longer input length parameter. Let's see what it gives.
Then, depending on the results, I will try to follow jwg1749's new method. I don't know whether the usb input can be used to output other digital inputs - while it is clear that it can be used for analog.
Other point is that I seem to remember that every input was upsampled to maximum before being treated in the system - therefore you're losing the bit-perfect feature and the results can be all over the place. Let's see.
Jwg1749's suggestion is a good one too and the USB input 'registers' input for both digital and analog. However I wouldn't be sure that the results give you a real conclusive conclusion as I think using the USB input you're not using all the parts of the amp normally involved in playback.
Regarding the length setting of REW I also don't believe it'll give you very different results since your toslink measurement were OK. It's more a "general" suggestion.
(02-Oct-2014, 16:41)Antoine Wrote: [ -> ]Jwg1749's suggestion is a good one too and the USB input 'registers' input for both digital and analog. However I wouldn't be sure that the results give you a real conclusive conclusion as I think using the USB input you're not using all the parts of the amp normally involved in playback.
That's true. But you would be using the input sections of the Devialet, including the mysterious DUET card that has been the subject of some speculation on this thread.
The idea is just to try to narrow down somewhat the list of candidates that might be generating the strange phenomena.
Matt
No changes extending the length of the test to REW's maximum. Still same problem.
Emailed again Devialet to check if they have any suggestion.
As a IT user with some idea, the typical support response would be "reinstall the firmware". Surprising it has not been done yet. Mine is 7.1.1
I'm new here but have found the same with a Mac Mini and the Devialet - it sounds poor. Slightly edgy, coloured and un engaging.
I have a Tascam DA 3000 recorder which accepts USB flash cards and connects via SPDif. Vastly preferable despite playing back exactly the same files.
No idea why as I've used the Mac via Firewire into a Dac before to great and enjoyable effect.
"Downgraded" to firmware version 7.1.0; the problem (ragged frequency response for USB and Air (Ethernet and Wifi)) persists.
Still no answer from Devialet. I have updated them with the new graphs.
The comb filter effect can be see with all types of averaging - here it is shwon with 1/24 octave averaging
More tests: I have configured my windows PC for REW and Devialet and have found that, in Windows, there is not any comb-filtering. So, it is something coming from the Mac. The issue is that is not about a single defective mac - there are at least two (a 2011 mac mini with 16 Gb RAM and a Mac book air with 4 gB RAM)
I am more inclined to think that it'll have to do with how the Devialet interacts with Mavericks. Toslink is not affected, but USB and Air are - so it is not Air alone.
Can any Mac user try to check these results?
(09-Oct-2014, 20:59)Divasson Wrote: [ -> ]Can any Mac user try to check these results?
Not sure whether it is answering your request, but in my tests with Dirac and FuzzMeasure playing tone sweeps through AIR (Ethernet) I didn't see anything untoward in the measured amplitude response. This is on a MacBook Pro with Mountain Lion (specifically 10.8.5).
Ian
It may have to do with Mac OS version. I remember the system to sound better with the previous OSX version.
I have moved 100% to Optical with good results - but know it can be better!!!