So counterintuitive. Happily the NEO iDSD volume knob is truly hardware/analog.
1 Like
As is my iFi Zen… no software volume control, just a nice smooth analog control.
Hi @DonaldM,
I’m sorry about this situation, I couldn’t do more testing on my side. I will look at this and come back to you toady on this manner.
I leave the volume setting on my iFi NEO iDSD fixed, but when I turn on the headphones it automatically changes to variable.
Using the remote control, it is very precise, but using the device’s volume button, it is not as accurate.
When I turn the knob it sometimes jumps a few DB both up and down.
From what I saw on the new NEO iDSD2 they changed the button type to a better one.
@DonaldM , I know understand what is going on and why I couldn’t reproduce this issue on the first place. When I made the initial setup of the Neo that were linked to a Neo iDSD via USB, I may changed it’s settings and put the Mixer Type to None. This would explain why during my tests for the certification of the unit, I have been able to pay DSD 512 without an issue.
I can confirm you, and for all Neo Stream users, you need to but the Mixer Type for the Volume Options to None to play DSD 512 tracks, otherwise you will be limited to DSD 128.
1 Like
@Antoine
Were you able to play DSD512 via Neo iDSD via USB using Windows?
Because for mcOS I only get DSD256.
If you use the ASIO driver from IFI in Audirvāna, you will be able to do so.
Yes, but I only use macOS.
Then it will be limited to DSD256.
Thanks.
2 Likes
I have no idea why after all this time Apple hasn’t bothered to put a native DSD driver in Core Audio.
It is probably too niche. All the OS’es I know don’t have native DSD drivers. The ASIO drivers for Windows are not from Microsoft either (the protocol is originally from Steinberg for latency purposes in professional audio recording). The ASIO drivers for Windows provided by DAC manufacturers are all third party made.
Apple (just like Microsoft) will probably never be interested at all in providing such a driver for such a ‘unimportant’ niche group like ‘audiophiles’. Maybe something for a third party to develop a native driver for MacOS (if possible)?
1 Like
This because your DAC supports up to 768kHz PCM, where 705.6kHz is the DoP 1.1 PCM carrier sample-rate for 11.2MHz (DSD256)…
Otherwise:
When the audio group at Apple Inc. see a competitive reason to support *native *1-bit PDM via USB or TB, they will… Otherwise DoP serves the audiophile market.
You’re right about niche. Linux has had a native DSD driver in the kernel for years, but for consumers that is a niche OS. (I wonder if Damien’s Direct Mode driver was capable of native DSD?) The Linux driver is yet another reason I’m using a DIY mini-PC streaming solution with Audirvana. (BTW, if anyone is curious about doing this themselves and knows a little Linux or doesn’t mind learning it, I’m happy to help.)
I was hoping Apple would do something along the lines of what happened with the USB 2 driver, where Apple actually had it before Linux, and Microsoft brought up the rear a decade later. But whether it’s a deemphasis of audiophile concerns since Steve Jobs passed away or some other cause, that’s not happening.
2 Likes
I know this, it’s just that Antonie’s answer seemed to have something new.
1 Like
I think that today with Qobuz and Tidal most users end up listening to 16/44Khz.
DSD is really something for few users.
For a long time I used Audirvana with Windows 10 exclusively for audio. I optimized using Fidelizer and AudiophileOptimizer, in my opinion even though I could do native DSD, I didn’t like the quality.
I don’t know today with Windows 11.
What was very good, practically the same quality as macOS, was using Windows Server 2019 optimized for audio.
1 Like
DSD64 and DSD128 will not provide distortion and noise measurements as good as PCM705.6 or 768, but once you get to DSD256 and above distortion and noise will very likely be better than feeding high PCM rates to the DAC.
I used optimizers in the past, but turning off services, which is what the optimizers do, always runs the risk of turning off something your player software expects to see. Not wanting to make things any more complicated than normal, I stopped using the optimizers.
My DIY streamer runs on minimized Ubuntu Server, which is more stripped down than Windows Server while having all necessary services running.
1 Like
We’ve been through this before… the Nyquist of a 2.8MHz (DSD64) signal is 1.4MHz… The Nyquist of a 5.6MHz (DSD128) is 2.8MHz… You cannot compare 1-bit PDM to PCM at any bit-depth… The Nyquist of 768kHz is 384kHz… You cannot compare noise-shaping in 1-PDM signals to bit-depth in PCM A/D… The bit-depth and sample-rate of PCM A/D defines the character of any PCMxxx recording… What we can compare is the resolution of a 2.8MHz PDM A/D to a 5.6MHz or 11.2MHz A/D or the resolution of a 16 bit/44.14kHz PCM A/D to a 24 bit/ 96kHz A/D or a 24 bit/192kHz A/D or any of the PCM bit-depth/sample-rate A/D to 24 bit /352.8kHz A/D.
I am enjoying the world of upsampling now for about a week. I like the sound quality, even with an older preamp.
What I notice that sometimes the music stops suddenly. I looked before a stop at the activity monitor, but nothing special there: ca 66% of cpu unused and memory usage is also very low despite upsampling to dsd512. Is this welcome to the world of upsampling, this can happen at these high sample rates? Are there perhaps certain precautions to be taken? Apart from the Audirvana sysoptimiser settings I mean.
Studio doesn’t crash, the music just stops and is released. Also if I look at the Ifi app there seems to be nothing wrong. There is no other software actively running apart from screen sharing on my Mac mini.
There is no error-correction of digital-audio signals in transmission, and any loss of synchronization information will have influence on the playback behavior in your scenario… How do you eliminate or mitigate these anomalies? You have been given insights on system design, from those that generally have reasonable QoS in their network design configurations here in this thread… A 22.4MHz PDM (DSD512) file is huge and the transfer-rate is high-speed, so more susceptible to RF, EMF and power/ground noise along the signal path, that can induce corruption of the digital-audio data… Bi-directional communication must be unencumbered and timely in nature (very low latency).
Outside of these potentials, it is well known that UPnP protocol In application on any given network amalgamation of components is far from perfect…
Thanks for your explanation!
I hoped that reading large chunks in memory would solve this…