Roon core decode 16 bit MQA to what?

What should a 16 bit 44.1 MQA that is encoded from 44.1 display under signal path and MQA Core Decoder ?

88.2kHz 24-bit

Yes the MQA core decoder currently displays it transforms a 16/44.1 to 24/88.2.

But the question use “should”, as I’m very convinced you need minimum Authentication in Roon to display MQA 88.2 or higher in order make anything transform/decode. So I’m notifying @support that there is an error in how Roon displays this in the Signal Path.

In Bob talks section at MQA website, which for me as a non native English speaking sometimes can be hard to follow, there maybe is hidden an explanation to this. To me Bob is very unclear on 16 bit MQA. He talk a lot about encoding and the MQA CD format.

He says about 16 bit MQA streaming:
Recently, as we have rolled out many more MQA files sourced from 44.1 kHz 16b masters – where we expect the MQA file to be 16b”.

Well then I also expect 16 bit to be displayed.

If Lumin HW where to the decoding (no MQA decoding in Roon), what will be displaying (if possible) on such a MQA track/album.

It’s not an error. It’s by design. I say this based on my experience from MQA integration. I cannot, however, explain the article you quoted.

Only the master sample rate - 44.1kHz here. We do not show internally what happens, other than Roon signal path.

Hello @R1200CL,

If Roon detects that the device does not have MQA Core Decoding capabilities, it will engage its built-in MQA Core Decoder to do the “first unfold”.

Here’s MQA’s explanation of what’s happening during the “first unfold”:

The MQA Core Decoder always outputs either 24bit/88.2kHz or 24bit/96kHz, regardless of the original bit-rate or sample rate of the content.



Is this information correct when you’re playing a 16 bit 44.1kHz MQA file ?
What is the actual format that is being outputted ?

You’ve been told 3-4 times already. It outputs 24/88.2 for 16/44.1 MQA files.

88.2/24. If this is passed on to an MQA rendering DAC then it will most likely come back down to 44.1 in my experience but they all vary. My Hiby R5 renders all MQA to the highest sample rate of its dac regardless of the source sample rate. But the core first decode is always to 88.2/24 or 96/24 by its design.

No, it will not be sample rate converted back to 44.1 kHz. MQA at the encoding stage may apply downsampling to facilitate folding ultrasonic information. But MQA decoding does not apply any downsampling, only upsampling.


Well I am afraid your wrong here when I played 44.1/24 mqa master via Roon to my mqa capable DAC the rendering stage reported it’s 44.1 in Roons signal path not 88.2 . The core decode was to 88.2. so explain what its doing. If it’s a higher res MQA mastering say 88.2 then it would render to this master sample rate. Although it could render to the max the DAC does if it’s setup that way.

Here is an example at 48khz same principle applies.

No. It only tells you it is rendering a MQA 44.1kHz music, it does not tell how at what sample rate the rendering occurs. It differs among different hardware, but it’s never 44.1kHz. The display of a MQA DAC shows the sample rate of the original (master), not the actual rendering rate. I think the rendering is usually a secret, unless the manufacturer tells you or use it as a marketing feature.

In an early MQA video on youtube, Bob Stuart says something about playing at a high sample rate to improve SQ during rendering.

The renderer is showing you the original frequency sample rate which is 48K
This does not meant it’s down sampled, it’s just re stating the original.

Ok, that’s just confusing then. Why bother show it again?


My Hiby dap shows what its doing and they confirmed it samples all mqa to max pcm rate it shows.

Fair enough still confusing though

This is so totally wrong and misleading.
There is absolutely no upsampling done by a MQA decoder. The decoder is unfolding encoded original rates of 1 and further up to to whatever the original is.

So in order to unfold anything, the original file must be at least 88,2 or 96 kHz.
A decoder do not add 8 bits to an original 16 bit file and output 24 bits.

I would like to see any documentation showing my understanding is wrong.

If it did, it would be of no consequence to the music file as they would be LSB 0’s

This was covered in the Bob Stuart post on 16b MQA

You seem to have a comprehension issue. You post your own multiple screenshots of 44.1 kHz ORFS being input to the MQA Core Decoder and being output as 88.2 kHz (or 48 kHz ORFS -> 96 kHz). And numerous others have explained that is how the Core Decoder operates – it always outputs 88.2/96 kHz. For 44.1/48 kHz ORFS, that entails upsampling. Bar none.

Yet, you persist.


This may bring some confusion. At least to me.

What is your definition of rendering rate ?

I would assume this is a process that take place after the MQA decoding is done ?

When you use the term sample rate of the original, I suppose you mean rate of the encoded MQA file ? (Which is displayed under authentication in signal path in Roon).