Problem playing 24/96 filewith 9y-3a

[boleary]

Well, I get an INVALIDARG error when I play a homemade 24/96 file with the 9y series. No problem playing the file with 9x-7. Its a home made file from a cassette tape that I converted to a 24/96 wave file using Audacity. It's a rather huge file. I looked everywhere but  couldn't find the cd for sale. ( Dvorak's Violin Concerto in A Minor, with Ruggeriero Ricci on violin with the London Symphony; I have other versions with other players but they pale in comparison. Anyone know where this can be gotten digitally?)

Anyway, I will  send the file to you if that would help (You really should hear it!). I'm itching to play it with 9y-3a, but don't need a quick fix; it's not like I'm playing Dvorak every day........ 

[PeterSt]

Interesting ... I mean, if this played with 0.9x and not with 0.9y I REALLY want to know why. Hopefully it is related to your other problem (format switching). And, hopefully it doesn't play here as well ...

So if you can, yes please, send me that file !

[boleary]

Thanks! It will be on its way shortly.

[PeterSt]

Well ... it looks like we are on to something ... (I incorporate the format switching here) ...

Here too I get the INVALIDARG, *but*, the data in the picture below just is consistent. Now :

I found that setting the DAC Is to 16/96 makes it play (at this moment I really can't guess why this can make a difference, but anyway it does), but it doesn't seem to play right. It is too fast, and I thought Dvorak was more a slowish guy.
Maybe you can confirm this (it being too fast) ? -> Set DAC Is to 16/96 and I guess it will play at your side just the same.

So, if all is right, this is a 16/96 file (below message tells this, and the Info tab in XXHighEnd tells that too), which is not what you expected (obviously, see your first post). Can there be a mistake ?

Don't get fooled by the fact that 0.9x played this, because 0.9x did not incorporate the the file itself at initializing the device, while 0.9y does. I don't say it should, but for some situations it is necessary, and, it is (formally) better. Thus, now these things come out (something wrong with the file (??)) while otherwise this may go unnoticed. This is very tricky to understand because otherwise it may also workout for the better, but depending on the situation.

Right now I can't decide what is actually wrong, apart from "something" strange going on at that message appearing with "DAC Is" set to (as in my case) 24/192. Afaik this shouldn't matter a thing, but the fact it does may be important at finding your switching problems in the first place (and this is what I hoped for !).

Before I proceed, let me know whether indeed the speed is wrong. Also, it doesn't sound good to my ears, but that may be denoising artifacts. However, it just as well may be the cut off of least significant bytes (as in one byte missing during playback, the file being 24 bits (3 bytes) afterall, but just putting out 2 bytes because of "the" issue going on here). This may combine with the perceived wrong speed, which should come down to (2/3) * 2 (or even *3) I think.
Never saw anything like this.

Btw FYI, I am using 0.9y-3q (so, your latest).

Let me know (the sooner the better).

[boleary]

  You are correct; it is a 16/96 file and it plays here when the 9y DAC setting is set to 16/96. Sorry about that. As for the speed it is correct here; it is mostly a very fast piece. As for the sound quality, it doesn't sound too good, artifacts and all; however, the way it is played is very good. Because it played in 9x at the 24/96 dac setting, I too thought there might be a link to the switching problem. Back to the drawing board......

[PeterSt]

Back to the drawing board......

Not quite, because if you can "admit" this is meant to be a 16/96 (and not 24/96) it should be something in my software not covering for 16/96 (which I by now think can be so).

Not to cross our by now ever so long email thread about the switching problem ... it won't be so that at your testing the switching you used more of these 16/96 files without knowing ? if so, it may invalidate some of the tests. Not that you used the wrong files, but merely you used flawed software in an area *I* didn't know about until today ...

[boleary]

Just checked and all the files used for testing for the switching problem were, indeed, 24/96.  :heat:I would be REALLY embarrassed if they weren't!

[PeterSt]

Ah good. Thanks.

Well, I just solved this one. A couple of hours searching for the culprit learned that actually there wasn't support for 96 (and 88.2) 16 bit files padded to 32 bits (or to 24). So that's why a 24 bit DAC wouldn't play it but the 16 bit DAC would.