Special mode stability

[Marcin_gps]

Hi Peter,

I've always had problems with special mode and that was with 3 machines and 3 different interfaces (M-Audio Transit, Cantatis Overture and RME-AES32). In each case it 'looks' the same, it plays fine for a while (few minutes, few tracks) and suddenly it goes wrong - distortions and noise, like if I chose to small buffer size. I have to start playback again until distortions appear and do this over and over again. Is there anything I could do to fix this? Is this driver's fault or what? It never happens in Adaptive mode. It never combines with 'To many buffor errors' massage. Playback doesn't stop after distortions appear. Distortions show up even with highest buffer size and high Q1 values. I'd appreciate your help on this, because special mode is better now IMHO - much cleaner, detailed sound, better bass compared to Adaptive which is somehow 'dirty'.

Best regards,
Marcin

[Calibrator]

All things being equal I would be putting money on the card/driver combination being the culprit. I have a ESI Juli@ card in both my music server and video server (Vista X64 and Vista x86 respectively ), and use Special mode in both with no stability issues at all. The drivers and Juli@ card are rock solid and I don't think I've read any problems with anyone else here that has them, apart from when KS was first introduced and we had to find the appropriate version of driver to use.

If you can borrow a Juli@ card from somewhere to try you will certainly enjoy the low latency it can provide

Cheers,

Russ

[PeterSt]

Hi Marcin,

This looks very similar to the problems I had with my new mobo (I started a topic about it), although I had it in any mode.
Btw, notice that Special Mode can't be accompanied by "buffer errors" just because the check is not - and can not be there.

The whole phenomenon of the distortion getting worse and worse (up to no music can be heard at all) is beyond my understanding. And of course I tried hard to understand, just because it would give me an opening in solving it, theoretically.
It took me days and days of changing BIOS settings, but in the end it was solved. How ? I still don't know because everything seems to influence. I use it for a few months now, and it's really gone, while before there was no way it would stay away (longest time would be 4 minutes IIRC).
Remember, I had it with all modes.

One thing to keep in mind (I think this *is* important) : Vista is just prone to this. I mean, with Vista it still can happen, as it happened (say) two years back on my old system. Especially when right before playback much happened (conversions) - while "Start playback after conversions" is WORSE than "Start Playback during conversions". This never changed apparently, and I noticed it at going back from W7.

I guess what you have is not different from what I had, while in my case there were no buffer errors from Adaptive. This is key, because the check I apply makes sure the buffer wasn't lost. Stil it gets lost, but where. I always thought it was PCI bus related, and then at the sample level. Also important I think : player settings didn't matter at all, which tells me it isn't buffer related. Also driver settings didn't matter, which tells me it's also not driver buffer related. It is "in the PC" and that's why PCI comes to mind. On the other hand each and every BIOS change mattered. It mattered for the severety, the time it stayed away, the distortion which would be there in the first place (a single playing piano was always a good thing to check for minor distortions, because a small tick at playing a piano note is very well audible (while e.g. drums wouldn't show it). This really helped me at quick checking, instead of wait whether it stayed away (which, as said, could take 4 minutes max).

Obviously I thought of what could go out of sync, because it really looks like that. Like the "system" telling the buffer runs empty, while it really was not. That would explain the severety getting worse and worse (things getting more and more out of line).

Lastly, think of the possibility that you have it with Adaptive Mode just the same, but it may be 100 times later because of the buffers really being that amount larger. And so you may never notice it ... (like in 100 times 4 minutes is not a realistic playback session).

I can't help more. But try to change BIOS settings of which you think they may be relevant, and at least try to find whether it matters. If it does, you may be able to find this one combination which lets it stay away.
And oh, I was talking about Vista being prone to it, but W7 and Vista really didn't matter for my case.

Peter

[PeterSt]

PS: Of course I won't go against anything from the Man Down Under, but I used a Juli@ ...

[Marcin_gps]

Calibrator, I don't have to get Juli@ to check how ESI drivers rock! I've just flashed my Cantatis card to Prodigy 7.1 hifi and, as before, I'm able to play with 1 or 2 samples of latency, while RME or Cantatis wouldn't allow stable 1024 samples playback in special mode. So, like Peter said, PC/BIOS configuration matter, especially anything that influence PCI Bus (sound card's IRQ, PCI Latency timer, IRQ affinity masks etc), but it looks that driver's stability is on top of everything.

Cheers,
Marcin

[Calibrator]

PS: Of course I won't go against anything from the Man Down Under, but I used a Juli@ ...

Indeed you do Peter, and I'm aware of that, but your initial teeth gnashing when putting your new system together appears to be due to immature drivers/BIOS for that uber motherboard you got rather than a limitation of the Juli@ card itself. When Marcin writes that 3 combinations of audio card and system configuration fail to remain stable, logic suggests ( well, my logic anyway  ) that trialling a known stable audio card would be the next thing to try.

Marcin, from my reading of your posts, you come across as having a good handle on PC matters so I presume you've been down the path of ensuring you have the current drivers/BIOS installed. Did your instability problem arise after you applied all the "tweaks" , or was it evident even on a fresh unmolested install?

Perhaps you could try unplugging the external audiocard and enabling the internal device for testing, assumming there is one available, which is most likely the case these days. Whether you can get KS Spec mode working at all though on those inbuilt devices is a bit of a coin toss.

I wish you the best in hunting down the underlaying cause of your stability issue.

Cheers,

Russ

[Marcin_gps]

PS: Of course I won't go against anything from the Man Down Under, but I used a Juli@ ...

Marcin, from my reading of your posts, you come across as having a good handle on PC matters so I presume you've been down the path of ensuring you have the current drivers/BIOS installed. Did your instability problem arise after you applied all the "tweaks" , or was it evident even on a fresh unmolested install?

Perhaps you could try unplugging the external audiocard and enabling the internal device for testing, assumming there is one available, which is most likely the case these days. Whether you can get KS Spec mode working at all though on those inbuilt devices is a bit of a coin toss.

I wish you the best in hunting down the underlaying cause of your stability issue.

Cheers,

Russ

Russ, newest BIOS/drivers anythign I could think of. I even contacted Asus support, because I thought sth is wrong with my MB. Right now I have 2 PCs and 3 different OSes (on one machine). Special mode instability or whatever we could call that happens on all the systems and that is before any tweaks was applied in the first place. But here's sth interesting. As I stated before, I flashed my Cantatis card in order to use ESI drivers and right now I'm able to play using buffer size at 32 samples and Q1=-4 with QAP with some small random clicks (with Q1=-3 completely glitchless) and the sound is unbelieveable, much better than RME-AES32 (slave) + dCS Elgar Plus DAC (master) in dual AES mode. How crazy is that? Amazing extensions of low and high frequencies, the sound is so powerful that it could crash walls and yet silk, fluid, amazing... I'm not selling Cantatis, no way. If Peter's DAC (and PC interface) is/will be able to play with 1 sample latency and I'm sure it beats the cr*p out of Cantatis, then I'll get one. For now, I'm staying with Cantatis and I believe that I can get even more out of it by just finetuning Windows/XXHE.

Cheers,
Marcin