I’m not trying to mock anybody, but if you want to hear anything beyond what’s there in CD resolution, you’ll have to wait until our auditory systems evolves enough to hear above 20kHz and/or not get permanently damaged by levels of 120dB SPL and above. Assuming that’s the direction of evolution… If a remaster sounds differently, it’s not because of increased resolution.
With all due respect I think you are missing the point. The point of high resolution audio is not the reproduction of ultrasonic information, but avoiding the effects of low pass brick wall filters in the time domain in the audio band. You are welcome to continue enjoying CD resolution audio and wait for your auditory system to evolve, and I will continue to enjoy high resolution reproduction with my decidedly average middle-aged hearing.
So @streamy68 (I think you know what I am going to say! ), as I reported over at devialetchat I tested the new 2.14.4 firmware release with your test files and my initial findings indicate that nothing has changed. In other words the signal in both the 24/96 and 24/192 files appears to have been low pass filtered somewhere between 44 kS/s and 50 kS/s (presumably 48 kS/s).
However Roon is reporting in “Signal path” that the 24/96 audio is being reproduced losslessly which would appear to be incorrect. I am guessing that the Phantoms are not reporting this subsequent filtering back to Roon.
@danny, what’s your view on this? Is correct reporting of downstream processing not part of the RAAT certification? Any other light you can shed?
If I am wrong and there is some error in my method hopefully somebody will point it out. Nobody will be happier than me!
If you’re talking about the frequency domain where there are very simple ‘measurements that show what people can and cannot hear’ then I (still) respectfully suggest you are missing the point.
If you are talking about the time domain where there are no measurements I am aware of ‘that show what people can and cannot hear’ (at least as far as phase shifts are concerned, as opposed to jitter - if you have some please provide links) then I continue to suspect we are religiously divided. My own measuring instruments (my ears) can hear the effects, and that is enough for me.
Indeed! According to your measurements made using my files a 25kHz sine signal is not transformed into an acoustic signal that your microphone would record. Obviously, the Phantom speaker internally filters or down samples (I tend to believe latter) before it drives the tweeter. Before the update the Roon signal path showed this as far as I can recall. The question is where RAAT ends to report the signal path.
According to this post RAAT only displays what the device reports.
If down sampling to 48kHz still takes place Devialet has simply stopped informing RAAT about it.
However, I believe that there might be more to this than we know presently so a conclusion is difficult to make. One thing seems certain; 25kHz is not reproduced by the Phantom.
Just to close out the original topic of this thread. I opened a ticket with Devialet on this and they confirmed that the reporting of lossless playback of streamed material >24/48 in the Roon Signal Path since fw 2.14.4 is a regression bug. While Phantoms do accept material up to 24/96 on the streaming inputs (RAAT and UPnP), the Devialet Operating System “DOS 2” is still filtering and downsampling such material to 24/48 before DAC. This incorrect reporting will be fixed in an upcoming fw release.
The DAC in current Phantoms is 24/96 capable, however this can only be exploited via Toslink in Direct Optical mode and not via the streaming inputs.