bsneed
New Head-Fier
- Joined
- Nov 19, 2012
- Posts
- 18
- Likes
- 10
I'm going to mainly speak in the context of the Mac in this post.
So, once upon a time (maybe a year ago?) I downloaded the then-GPL source to Audirvana. It just so happens that I still have it handy. I wanted to add integer-mode to my own player this evening since I see it discussed here so often, so I dropped into the Audirvana code to see what exactly it entailed to *them*.
To my surprise, to Audirvana as it was then, it simply meant selecting it as the device's Physical Format in CoreAudio. Many DACs support various configurations of Xch-Xbit-float vs. Xch-Xbit-integer. Is it really this simple? If it is, I might conclude its a misnomer for some of the discussions I see happen here. Namely those which say something to the effect of "X player is the only one that supports integer mode". If what I'm looking at is true, it becomes:
A) does the player support sample rate switching per file? If so, then so long as it selects a matching format for the sample rate its going to. I have yet to come across a case where integer mode is available on some sample rates and not others on a given DAC, though I've only tried 3 so far and on all of those, integer mode was the default.
or
B) if it doesn't support sample rate switching and resamples, then its entirely up to the user as to whether integer mode is enabled or not.
Take these screenshots as examples:
1. My Retina MBPs Built-In Output on toslink: http://d.pr/i/6r9z
2. My Benchmark DAC1 HDR on USB: http://d.pr/i/4Z1O
Where it seems like it may actually make a lick of difference is with regards to the Virtual Format. For those who don't know what Virtual vs. Physical formats are here you go:
Physical Format = The format in which the device expects audio to be delivered in.
Virtual Format = The format in which the audio graph will do its audio processing.
However as you can see here for my Built-In Output device (http://d.pr/i/rTg5) only 32bit float is actually supported for a Virtual Format. This is also true for the Benchmark HDR. In case you're wondering the application that's telling me that is Apple's HALLab app.
Surely I'm missing something and hoping anyone here in-the-know can help.
So, once upon a time (maybe a year ago?) I downloaded the then-GPL source to Audirvana. It just so happens that I still have it handy. I wanted to add integer-mode to my own player this evening since I see it discussed here so often, so I dropped into the Audirvana code to see what exactly it entailed to *them*.
To my surprise, to Audirvana as it was then, it simply meant selecting it as the device's Physical Format in CoreAudio. Many DACs support various configurations of Xch-Xbit-float vs. Xch-Xbit-integer. Is it really this simple? If it is, I might conclude its a misnomer for some of the discussions I see happen here. Namely those which say something to the effect of "X player is the only one that supports integer mode". If what I'm looking at is true, it becomes:
A) does the player support sample rate switching per file? If so, then so long as it selects a matching format for the sample rate its going to. I have yet to come across a case where integer mode is available on some sample rates and not others on a given DAC, though I've only tried 3 so far and on all of those, integer mode was the default.
or
B) if it doesn't support sample rate switching and resamples, then its entirely up to the user as to whether integer mode is enabled or not.
Take these screenshots as examples:
1. My Retina MBPs Built-In Output on toslink: http://d.pr/i/6r9z
2. My Benchmark DAC1 HDR on USB: http://d.pr/i/4Z1O
Where it seems like it may actually make a lick of difference is with regards to the Virtual Format. For those who don't know what Virtual vs. Physical formats are here you go:
Physical Format = The format in which the device expects audio to be delivered in.
Virtual Format = The format in which the audio graph will do its audio processing.
However as you can see here for my Built-In Output device (http://d.pr/i/rTg5) only 32bit float is actually supported for a Virtual Format. This is also true for the Benchmark HDR. In case you're wondering the application that's telling me that is Apple's HALLab app.
Surely I'm missing something and hoping anyone here in-the-know can help.
