I use Audirvana to convert files to DSD64. Just because I can, I reckon.
But I’m puzzled by some curious identification of the DS DAC input files.
If I start a new listening session with Audirvana, DS DAC shows the input as “DOP” and rate/bit as DSD64/1.
If I start a listening session with Apple Music, the input correctly displays “PCM” and 44.1/24 (which are the parameters set up on my mac mini).
However, if I now go directly to Audirvana (after Apple Music), the input changes to “DSD” (and the usual “DSD64/1”).
What causes the input file to magically change to DSD in Audirvana ONLY after listening through Apple Music?
My audio chain: mac mini>AQ Carbon usb>Matrix X-SPDIF 2>AQ Vodka hdmi>DS DAC.
There’s no audio change for the good/bad/worse that these ears can detect. Just curious.
The DAC is telling you truthfully what it receives. It’s not a file at that point, it’s just a stream of audio data over I2S. In both cases the framing is PCM framing but it’s switching between 44.1/24 actual PCM audio (which implies you are not sending bit-perfect CD rips but perhaps doing some software volume or EQ) and 176.4/24 with the DSD64 data embedded according to the DoP standard.
You’ll need to seek answers from people who know Audirvana.
I do believe that DOP and DSD are essentially the same stream. I’m just puzzled as to why the “labels” change in the above scenario.
What I’m saying is that the stream does change. The “label” is accurate. First you’re getting actual PCM at 44.1. Then you’re getting DSD via DoP. The reason why will be something on the Mac side.
The music signal encoded in a DSD stream and a DoP stream are the same, in fact the DS internaly repackages DSD as DoP
I don’t believe this to be true. It should be the other way where everything gets converted to DSD before being outputted to analog. Can you share your reference?
Edit: I stand corrected, see below.
Any incoming native DSD is wrapped into DoP at the inputs in the DS so that there doesn’t have to be a parallel path for DSD to the already available PCM path through any input select and buffering. At the final upsampler any DoP is converted to native DSD and processed identically to PCM. Finally the upsampled signal is converted to single bit quad rate DSD for final FPGA output to the analog filtering.
Whatever the screen of the DS DAC shows you at any point in time, that’s the digital audio signal format that it is receiving in that moment. If it says DOP DSD64 then that’s exactly what you have on the wire. Your original question is about why does it sometimes say PCM 44.1/24, and what I’m trying to tell you is that when it says it’s getting PCM 44.1/24 then it really truly is getting that.
The DS DAC I2S inputs only make use of two signal lines instead of the usual three: the serial data line and the word clock line aka left-right clock. The word clock switches high or low to indicate the start of each new sample appearing on the data line, alternating between a sample for the right channel and one for the left channel. In the intervening time some number of bits will be signalled on the the data line – most often 16 or 24 or them, but other sample depths are possible.
When the incoming audio is sampled at 44.1kHz the word clock will toggle high and low 44,100 times per second. The DAC can directly observe this, and won’t be confused about it. If it says 44.1 on the screen, the signal being received is sampled at 44.1.
DOP is a clever scheme to pack 16 bits of DSD data inside of each “sample” in a 24-bit PCM transmission system in a way that makes it obvious to a DOP-compliant receiver that the real audio data is DSD not PCM. To match the data transfer rate of DSD64, it is necessary to pack the bits inside a 24/176.4kHz PCM system. When coming over I2S this means the word clock will be switching 176,400 times per second. Again there is no opportunity for confusion at the DAC end. If it sees this sample rate, and recognises the pattern embedded in the first 8 bits of each “sample” as the one which signals DOP encoding, it will process the audio as DOP and let you know on screen.
So what I’m trying to tell you is that when the DAC says it’s getting 24/44.1 PCM that’s what the computer is sending it. And if you don’t want it to be doing that, you’ll need to talk with people who understand how Audirvana interacts with macOS to try and figure out how to ensure you get the DOP transmission from the computer that you’re looking for.
I understand the PCM part. My question was: Why do DSD-converted files from Audirvana sometimes show as DOP>DSD…and other times DSD>DSD.
Tho I completely trust the FPGA’s measuring of sample rate, sample width, DoP’nes and DSD’nes as well as the UI’s display of sample rate, sample width and either DSD or DoP, I have just a little less certainty that the display code always gets DSD vs DoP correct, simply because I’ve never seen that UI code. Still I’ve never seen it wrong on my screen and I always know what I’m sending.
So, as dvorak has been saying, the most likely explanation is that your source is sending exactly what the DS is displaying. That is, it’s sending DoP when the DS displays DoP and DSD when the DS displays DSD. Why it’s doing that is a question for them.
In any case it’s completely benign: the same exact DSD bits ultimately arrive in the DAC even if sometimes they are sent as DSD and other times as DoP or conversely the display of DSD vs DoP is sometimes broken.
Thanks, Ted and Dvorak. That’s what I wanted to hear. Because I don’t hear a difference lol.
Ah, I now see I misunderstood your question. I used to run Audirvana in iTunes integrated mode, where it would watch what you selected to play in iTunes and hijack the actual track reading and output to the DAC. So I thought you were describing selecting music in Apple Music which would actually be played by Audirvana, and a situation where clicking “play” in one or the other resulted in PCM or DOP coming from Audirvana to the DAC – same song, same track position, different encoding.
Instead it seems your question is why Audirvana, in standalone mode, sometimes sends your DSD files as DOP and sometimes and DSD native. Naturally playing straight from Apple Music sends PCM, because that’s all it’s capable of doing.
My answer is basically the same, barring Ted’s comment above that the display might not be reliably showing you what the FPGA is doing (which would be truly annoying ). On the technical side, native DSD via I2S is done by using the word-clock line as a second serial data line. You just stream DSD bits in parallel down the two lines because there are no words to start and finish. One line is for left channel and the other for right. Again the DAC will pick up on this within microseconds and start decoding appropriately. That’s the condition which should show “DSD” on the display rather than “DOP”.
IIRC the Matrix has an option to always decode DOP to native DSD before relaying down I2S. Could be wrong. Doesn’t really matter either way.