Matrix Audio X-SPDIF 2 (USB - I2S interface)

The X-SPDIF 2 does need a reference in the SonicOrbiter kernel to expose the native DSD capability to Sonore devices and that will be included it in a future update. Thanks to @Jesus_Rodriguez and Harry (@spockfish ) for your help in sorting that out.

I’ve been listening to Classical piano PCM upsampled to 705.6 kHz in HQP with an IIR filter and NS5 dither and greatly enjoying the dynamics and transients.

I haven’t listened long enough yet to be sure but there seem to be fewer issues with power. The LPS-1 is powering the mR as usual, and passing power on the 5v line to the X-SPDIF 2. The other change I’ve made is to update HQP to 3.17 (involving a hunt for my licence keys, found them :sweat_smile:). If everything keeps stable for a day or so then this would seem to be resolved.

Haven’t noticed the static issue again either. Rebooting the DAC may have fixed this.

So things are looking good !

Edit: I’m still getting sporadic pops and clicks at 768/24. 705.6/24 is pop free.

Edit 2: The pops and clicks seem to be associated with playing 44.1 family material at 768. I’ve enabled Auto-rate Family in HQP and it’s stable so far.

Edit 3: leaving the LPS-1 feeding the mR and X-SPDIF2 while away at work resulted in a quite hot to the touch LPS-1 and a crashed mR. I think it must be hitting the hard current limit of 1.2 amps.

So I turned the LPS-1 down from 7v to 5v. The mR is stated to require 6-9v, but it booted with 5v and the X-SPDIF lit up and I’m listening to music now. Will report more after leaving it running.

If you try this, be sure to wait while the LPS-1 adjusts from 7v to 5v. It’s light goes from red to amber to green until it’s ready.


Ok. Latest update to SonicOrbiter software from Sonare includes kernel patches that allow the microRendu to send native DSD 512 to the XSPDIF2. Listening to Diana Krall with a white light showing on the XSPDIF2, indicating native DSD.

I’ve separately powered the XSPDIF2 with a 5v iFi as I was still experiencing sporadic excess current shutdowns by the LPS-1 even switched to 5v. I’ve shifted it back to 7v and it’s solely powering the microRendu again. Everything seems much happier (no crackles or artifacts as yet).

Now to settle in and do some comparative listening.

Sorry to report that, in last night’s listening, I was getting infrequent drop-outs with TIDAL and local material at both DSD512 and 705k PCM. I’m wondering if it’s related to the new Ropieee build which self-installed itself on my Pi. However, it could also be a processing issue; I need to check that.

Hmmm. Sorry to hear that. The only thing that’s changed that could potentially influence that is an updated kernel. Still puzzled to what that could be though. Maybe it’s performance?

I’m getting myself a DSD512 capable DAC later this week, so let’s see…

Hey Harry, IIRC I was close to the edge with DSD up-sampling (1.0 to 1.1x), but PCM up-sampling had a lot more headroom. I can’t get to processing speed at the moment as it’s broken in my current Roon build!

looks around furtively
I … I’ve been listening to PCM again …


Only because you now have 768 and 768 > 512 :wink:


Do it again and you’ll go blind! :smiley:


what dac are you using again? the halo spring?

Yep. Holo Audio Spring (Stage 2) after an mR into a Matrix Audio X-SPDIF2. The IIS/HDMI input into the Spring supports 768.

Holo Spring doesn’t support 768 even using IIS/HDMI input. It maxed out to 384. Go to page 18 below:

I’m listening to it as I type this. HQP is outputting 705.6, the Spring display is showing 705.6. I think @joel confirmed with Holo Audio that the Spring will do 705.6/768 over the I2s input.

The interface must support 705k to support Native DSD512.

Yup. Holo Spring is limited to 384k PCM over USB, but supports 705k and 768k PCM and DSD512 native over IIS/HDMI.

Singxer SU-1 limitation?

This is a bit confusing, Jeff Zhu told me 705.6/768k is not supported at all (USB and IIS/HDMI) and he has no intention to update Spring. While this come as surprise that @andybob manage to make it works! In CA, many have asked this, Ted_b and Tim from Kitsune HiFi gave the answer ‘No’.

I also got an email from Jeff Zhu…

705K and 768K via usb is not supported in current firmware. But it can be supported by I2S. We may release some firmware update to support 705K and 768K via USB later. But it seems not too much user need that. So the current plan is to support this at Spring2.

No, native DSD transmission is not dependent on PCM sampling rate unlike DoP. DSD256 DoP requires 705.6k PCM

Hi @andybob

The X-SPDIF 2 looks interesting. Did you experience any nasty pop sound when playing back DSD? Did you try on ROCK? Thanks

I’ve always had the mRendu in front of the X-SPDIF2.

There are two pops when rebooting the mRendu with DSD selected. They are moderate in volume and only occur on reboot. This occurs both with and without the X-SPDIF2 so it is not attributable to it.

There used to be a loud pop when playing each new track using native DSD through the mRendu. This was found to be due to an error in the ALSA code running on the mRendu where PCM silence was sent instead of DSD silence. Sonore fixed this in the Sonicorbitor 2.5 software upgrade and it has never recurred for me.

Before the X-SPDIF2 arrived I was experiencing some crackling playing back DSD 512 through the mRendu. I found that reducing the buffer in HQP to the minimum 5ms fixed this.

After hooking up the X-SPDIF 2 I have experienced a crackle when playing back 48k family material in PCM 768 using Roon/HQP/Windows10/mRendu/XSPDIF2/Spring. If I upsample to DSD 512 (8x48k) I occasionally get a slower replay as if the 48k material was being played back at 44.1k. This slow replay issue predated arrival of the X-SPDIF2 and I haven’t yet isolated what is causing these issues. I suspect I may have a damaged circuit in the Spring resulting in it sticking on the 44.1kHz oscillator instead of shifting to the 48k oscillator. For the moment when I want to listen undisturbed I use the poly-sinc filters in HQP and upsample everything to 706.5kHz.

I am enjoying listening to poly-sinc-xtr-mp upsampled to 706.5 kHz PCM with the NS5 noise shaper.

I don’t think any of the above is attributable to the X-SPDIF2. I would recommend using a separate power supply rather than attempting to power it through the USB connection. I think it is an impressively clean and clear component.

I haven’t used the X-SPDIF2 with ROCK.

1 Like

Thanks @andybob

May I know where and how much you got the Matrix X-SPIDF-2? Thanks