"Device Allocated But Cannot Play" error

[charliemb]

Background:

• Vista 32 w/ 4GB
• XXHE 0.9z-5-2
• Wyred4Sound DAC2 ->proprietary USB 24/192
• The only device option that shows a USB device is KS:
• Yes, I set the data to be 32bits
• max rate of 176 or 192 selected either manually or through fx, confirmed

Behaviour:

• Plays fine for a while at any rate
• I eventually get the dreaded "Device Allocated But Cannot Play" error which required a reboot
• I also get the "Engine #4 did not start within the expected time" error, but not always.
• I never saw the "Device Allocated But Cannot Play" error until I started 4x upsampling while choosing Arc Prediction and Peak Extend.
• with 4x upsample, I let xxHE auto-change the sample rates and had fx checked
• After I get the "Device Allocated But Cannot Play" error, not even JRiver can play music until after a reboot.  (The software path to the DAC is dead.)

Please Help!!

[PeterSt]

Hi Charlie,

Let me first notic that you are not using the latest XXHighEnd version (that would be 0.9z-6-1c).
Then, I don't assume this occurs when another format is encountered in the Playlist, right ? Notice that this already will happen when an HDCD album is encountered (during playback, from one album to the other).
But careful, this "format change allowed" does *not* work with the latest version (I think with the version you use it should, but I really can't tell).

The above all not assumed to cause the problems ... this is a tough one. I too played with the W4S DAC2, and it was without problems. But maybe I didn't play long enough with it to encounter the problem. Anyway, it should be a driver issue. That is, I can't think about anything else. Something like the driver doesn't shut down the internal audio playback properly and then it stays engaged. But, I don't think such a thing should or can happen when playback is going on, so you'd have to use Stop first, and *then* suddenly things won't work anymore.

To this latter regard, the latest XXHighEnd version works quite differently, and theoretically even more encouraging for the problem. I still advise you to try this, because we can learn a few things from the behaviour which may occur then.
Notice that with 0.9z-6-1c the Audio Engine keeps on running always (in a certain "state") and to that matter it just may avoid your problem. But also, the fact hat you can throw out the Audio Engine (XXEngine3.exe) - which goes by means of the Off button in there, may change a few things.

Outside the above, I would try to lower the SFS a bit. Something like : when you now can play with 120, make it 90 and see whether this helps. So, this would be a very first thing to try.

I eventually get the dreaded "Device Allocated But Cannot Play" error

If this is with Unattended and just in the middle of playback somewhere (but could be at the boundary of tracks), lowering the SFS should help (you are consuming too much memory) and otherwise we must see further (but see above options of course).

I hope this is getting you somewhere ...
Peter

[charliemb]

Let me first notic that you are not using the latest XXHighEnd version (that would be 0.9z-6-1c).

Peter, the reason I don't use the latest version is that I am using my TV as my monitor from afar and I really need a 10ft interface ala XMB or JRiver Theater mode.   With this setup, I can easily hit the minimize OS button by mistake and I cannot afford that because I have no OS re-install media (I lost the DVD).  I'd be without music for two weeks if I had to start from shopping for and buying an new OS, getting all the drives, etc.   NOT Going to happen.
--Perhaps if you have a way to make "Minimize OS" impossible from the UI of the latest version, then I would reconsider.--

(since this is its own topic, will respond to the other points in a separate post)

[charliemb]

Hi Peter

Then, I don't assume this occurs when another format is encountered in the Playlist, right ? Notice that this already will happen when an HDCD album is encountered (during playback, from one album to the other).

I want to say I agree.  The first time this happened was immediately after xxHE timed out and closed down; it happened on the first time I hit Play after bringing it up again.

And now that you mention it, one of the earlier instances did include an error message about HDCD.

But careful, this "format change allowed" does *not* work with the latest version (I think with the version you use it should, but I really can't tell).

It is possible that this is the problem then.  I play with this and report back.  Right now I'm attemting 2x for awhile.

...W4S DAC2...Something like the driver doesn't shut down the internal audio playback properly and then it stays engaged. But, I don't think such a thing should or can happen when playback is going on, so you'd have to use Stop first, and *then* suddenly things won't work anymore.

That's correct.  It never happens in the middle of playing.  So far it only happens on a fresh start of hitting the play button _{(which is too small)}.   Now we are getting somewhere.  (Note the only option I have is engine #4.)

Otherwise, I'll try the SFS trick.  Thanks.

I eventually get the dreaded "Device Allocated But Cannot Play" error

If this is with Unattended and just in the middle of playback somewhere

This is not with Unattended.

Thanks, 
Charlie

[PeterSt]

Hey Charlie,

The first time this happened was immediately after xxHE timed out and closed down

I think this tells all ...
So, you are using the Demo version. Well, in that case indeed the Audio Engine is just removed from memory and nothing much is offficially closed down. Precisely the situation I described/expected. I'm not even sure whether I changed this for the upcoming 0.9z-7, but I can. But ...
This is not how a driver should behave ...

You can easily test this at the cost of one reboot (most probably ). Just start playback and with TaskManager kill XXEngine3.exe and see what happens. Need a reboot ? then indeed this is the culprit.

Your remark about MinimizeOS is appreciated. But, what you may not know is that there's always a mesage "Are you sure" after the (by accident or not) attempt. But then with Demo Mode this is moot of course, because it won't go anyway with that. Haha.

Btw Charlie, I must thank you for your preciseness in your elaborations. Really helpful !

Ok, let me know ...
Peter

[charliemb]

I think this tells all ...
So, you are using the Demo version. Well, in that case indeed the Audio Engine is just removed from memory and nothing much is offficially closed down. ... This is not how a driver should behave ...

You can easily test this at the cost of one reboot (most probably ). Just start playback and with TaskManager kill XXEngine3.exe and see what happens. Need a reboot ? then indeed this is the culprit.

With all these reboots, why not one more?

Okay, so, yes, in the middle of a song, I killed xxEngine3.exe and that reproduced the problem.

Congratulations! 

No what? 

[PeterSt]

Now what ? Now I better create something which officially "terminates" playback. But this won't help you when for other reasons playback gets terminated. So it really is a diver problem ... (which btw not has been reported from other drivers, yet). So you may shoot an email to W4S as well.

Btw, I am not sure how to tackle this for "format change" in a decent fashion, because as far as I'm concerned this is all okay already. I mean, I may be able to approach this more decently, but I can't test it really (it needs your W4S).

For now, don't allow "format change" is the least you should set. That the problem is incurred for by the Demo version is unintentional of course ...

Peter

[charliemb]

But this won't help you when for other reasons playback gets terminated.

I hit the stop button all the time to change tracks; and this, by itself, does not seem to cause the problem. 
Besides the demo cutting off 3.exe, what "other reasons" could there be for terminating playback?

So it really is a diver problem ... (which btw not has been reported from other drivers, yet). So you may shoot an email to W4S as well.

What do I tell W4S?  _{(This problem does not happen with any other software.)}

For now, don't allow "format change" is the least you should set.

If I go back to manual format change, i.e., manually using the slider for 1 2 4 8 upsampling, and this problem does not reoccur, then I think this proves that the problem is with the auto format change code.

[PeterSt]

Hi again Charlie,

Besides the demo cutting off 3.exe, what "other reasons" could there be for terminating playback?

Numerous, and if you never encountered one you are just "lucky". But to name a few things :

• For whatever reason your PC is stressed and you must kill a few things. Amongst that the audio player (still playing) or a (without you  knowledge about it) related process otherwise.
• Your PC hangs temporarily. This will incur for the same situation.
• The software player is out of order and stalls. Yes, this happens with JRiver too (my own experience).
• You do something to the device, like unplugging it, shut off the power (during playback).

If I go back to manual format change, i.e., manually using the slider for 1 2 4 8 upsampling, and this problem does not reoccur, then I think this proves that the problem is with the auto format change code.

You can call it that if you want (far down it will be true). But still this is not really the case because as I said, what I do is fairly decent. The only thing which does not happen at a format change is that the audio device is re-"instantiated". But it is normally stopped and everything. So, your driver just can't cope with the device changing format "underway" in normally (pin) stopped state.
But of course we could also look for others having such a problem. Well, none (that I know of, but see the forum). Btw, you are the first I see here with a W4S (DAC2) (which doesn't tell all of course because I may not know), plus it needs the more or less coincidental format change. And funnily enough, only two weeks back someone noticed that the format change doesn't work for 0.9z-6-1c, while this version is out since last October ! IOW, it *also* needs running into a format change.
That this doesn't happen with many is undoubtedly related to relatively few people using Hires of some sort, which by (not) accident will be related to XXHighEnd focusing on Redbook fairly explicitly. This is a longer story, but at least I never play HiRes, so I won't run into a format change as well. Btw, my earlier remark about HDCD will count only when Redbook is played natively (16/44.1) and *then* a HDCD comes along (which will end up in 24/44.1). That too will be a most rare situation because most if not all will use Arc Prediction and output 24/176.4 (etc.) all the time anyway, and for HDCD this will then be the same (format).

What do I tell W4S?

Well, first of all that it (the device is not freed) will happen with any software (contrary to your subscript ). So, start JR and kill it via TaskManager. Notice though that we talk Kernel Streaming here, and that WASAPI may not show the problem at all (you could try it if you want). However, you must understand that XXHighEnd is special to the sense of how the audio engine is separated from the GUI program (which is XXHighEnd.exe) and what happens (needs to happen) when this is so. Thus, XXEngine3.exe runs in memory with the device "instantiated" and can't do this for every track coming along, while it doesn't even know about tracks (all is in memory and stuff). But, it easily can stop the device (the Pin) (most officially), change the format, and start it again. So, this is what happens, and this is what your driver can't cope with. Similar counts for not stopping it (like when XXEngine3.exe is killed), where the driver should take care of stopping all. In your case it doesn't. Now when a next playback session comes along it will still be in use. By what ? nothing. And this is why the driver MUST take care of these things.
Think a bit like plugging in a USB disc (etc.), switching it on, and next pull the USB cable out again without switching the device off first. It obviously is not allowed that you need a reboot now before you can use the USB port or that same device again. So, the driver must take care of that it's officially vanished, ready for a next task.

With the above it still will be so that I can solve this for you (not the format change !!), and 95% chance that in 0.9z-6-1c this already is solved. But I will take care it for sure is in 0.9z-7 (I mean, why not). But remember, the driver is still not OK.

One last thing : When you press Stop and then start another format, the problem does not exhibit, because after Stop XXEngine3.exe is removed from memory (this time *after* the device is nicely stopped) which frees the device (this may seem confusing to you). Now when you press Play, XXEngine3 is started again, and all is fresh and new. Now, to this regard 0.9z-6 is different because there XXEngine3 always keeps running, no matter you press Stop. So, can you feel the difference (going either direction for better or worse) ?

Ok, I'll stop, because I'm sure I'm making it only more confusing for you.
Peter


PS: But please keep in mind the difference between the format change (which any other player may deal with OK) and the device not being freed by the driver when the playing program has gone unexpectedly (which thus will happen "to" any player). Both however, are incurred for by the same driver problem.

[charliemb]

...what I do is fairly decent. The only thing which does not happen at a format change is that the audio device is re-"instantiated". But it is normally stopped and everything.

I thought you said that you just kill 3.exe as part of the demo.
Anyway, last night I started having this problem more often than not, even though I did these three things, I still ran into the problem:

• Turned off auto rate change, and unchecked fx
• Manually selected upsampling
• Gave up on upsampling altogether, kept it on 1x _{still failed}

So, your driver just can't cope with the device changing format "underway" in normally (pin) stopped state.

There might be a false assumption here somewhere:  I get this error even when all files are of the same type, and none HDCD.  There's probably something else going on here.

Btw, my earlier remark about HDCD will count only when Redbook is played natively (16/44.1) and *then* a HDCD comes along (which will end up in 24/44.1).

I just want to emphasize that, although I did once see a warning about HDCD.exe, I was not playing any HDCD.  I got that error when playing the Indigo Girls CD (original release) which must have predated the invention of HDCD.  So I really think that this has nothing to do with HDCD _{unless there is a false positive detected by hdcd.exe}.

I'm playing mostly aif 44.1/16, and flac for high rez, and hhxe does not yet play DSD (.dsf).  _{btw there is google code for DSFIFF .dff files to PCM, please consider.  I think it is called DSD2PCM.}

...95% chance that in 0.9z-6-1c this already is solved. But I will take care it for sure is in 0.9z-7 (I mean, why not). But remember, the driver is still not OK.

Because the demo disables the Minimize OS, thus making the demo "impossible" to allow mistake click by me, and because .9z-5-2 was so horrific with my DAC, and because the Minimize button is no where near the play and stop buttons (I've already accidentally hit the unattended button by accident and wrongly answered the popup message and lost control of the system), I installed the latest version .9z-6-1c.    Sure enough , I have yet to have this problem.  It may come, but I did not see it in about 1 hour of very intermittent use late last night.
BTW, please, no more RAR.  I am joining the chorus.

However - please note that the problem with no network play still exists in the latest .9z-6-1c.

I will email W4S and let you know.
Hopefully with .9z-6-1c I will finally be able to at least evaluate the audio engine. 

So I imagine the "won't play over the network" thread will be more active going forward, since it appears you already know how to solve this problem (at least so far) as per the latest version.   With .9z-6-1c I changed format, etc, several times

P.S. As suggested, I may kill JRiver in the middle of play to see if indeed it nukes the driver.

P.P.S.  Tonight I will also change format several times with .9z-6-1c and see (I think I did but don't specifically recall).

[PeterSt]

Okay, short answer for now Brian :
What about me not knowing anymore what was going on with the version you used ? I'm just honest here ...
Still the version you used is way too young to ever know about a problem regarding this (?? ) in there. This, while some sort of Format Change problem sure was known. 3-4 years ago and with exactly one DAC (an EMM Labs).

Btw, please notice my PS in the previous post ... Format Change (if ever) is a different thing (for XXHE !) from the device keeping allocated (though caused by the same Driver problem).

But hey, if the latest version solves it, why bother ... right ?
Let me know how things proceed (network thing), and again thanks for your elaborations,
Peter

[charliemb]

This problem is solved from at least the latest version .9z-6-1c . 

Perhaps tomorrow I will try with WASAPI on the "Minimize OS" free version 0.9z-5-02 .