kalimba

Sorry to ask a stupid question, but where can I find that? I only see releases up to 3535...

Also, is this to get rid of the beeping problem? Or for some of the other issues addressed on this thread...

Thanks so much!

RIV@NVX

Test releases are usually linked to in General section, check this thread 3536a development release is out

__________________
-- Vedran

rubber_glove

hi:
i poseted on the original thread, so i thought i'd come back with an update.

here's my info:

Windows build - v.5.0.2195.2.Service Pack 4
DirectX build - v.5.3.0000000.900 built by: DIRECTX
OpenGL renderer - (n/a)

System board
------------------------------------------------------------------------------
CPU speed - 1497 MHz
Type - Intel
AGP aperture - 64 MB
AGP transfer mechanism - Disabled
AGP non-local memory - (n/a)
AGP revision - 2.00
AGP transfer rates supported - 1x, 2x, 4x
Current AGP transfer rate - (n/a)
Sideband addressing - hardware support, but currently disabled
Fast write protocol - hardware support, but currently disabled
AGP texturing - (n/a)

Graphics card #1
------------------------------------------------------------------------------
Identity - NVIDIA RIVA TNT2 Model 64
Memory clock - 149.79 MHz
Engine clock - 125.01 MHz
IRQ - 9, shared
AGP revision - 2.00
AGP transfer rates supported - 1x, 2x, 4x
Current AGP transfer rate - (n/a)
Sideband addressing - (n/a)
Display driver - nv4_disp.dll, v.5.00.2195.0522
DirectX driver - nv4_disp.dll, v.5.00.2195.0522
Attached monitor - Korea Data Systems Visual Sensation VS-5/VS-51/VSx-5 (Microsoft)
Monitor caps (1) - 1280x1024, 70kHz, 120Hz

Device enumeration
------------------------------------------------------------------------------
D-Link Ethernet controller (30431106h) - using IRQ9
Creative Labs Audio device (00021102h) - using IRQ9
Creative Labs Input controller (70021102h)
Creative Labs Audio device (00021102h) - using IRQ9
Creative Labs Input controller (70021102h)
ADS Tech Firewire bus (8020104Ch) - using IRQ9
NVIDIA RIVA TNT2 Model 64 (002D10DEh) - using IRQ9
Intel CPU-to-PCI/AGP bridge (25308086h)
Intel PCI-to-PCI/AGP bridge (25328086h)
Intel PCI-to-PCI/AGP bridge (244E8086h)
Intel PCI-to-ISA bridge (24408086h)
Intel IDE controller (244B8086h)
Intel Universal serial bus (USB) (24428086h) - using IRQ9
Intel SMBus (24438086h) - using IRQ10
Intel Universal serial bus (USB) (24448086h) - using IRQ9

ha... perhaps too much information,.

as for pci latency, well. the first time i changed the pci latency, it made the bug (and another....unrelated issue) disappear. not for too long, however. so now all my cards are set to 32....

what else... it seems that the bug is getting more and more frequent as of late. it doesn't seem to make it very far through an mp3 playlist before silence or 'beeping' occurs. i am going to install the latest drivers and try some observing and post what i notice...

thanks,
martin.

rubber_glove

well. the kxcrtl -rv feature does not seem to stop the beeping. the only thing that gets rid of it is teh usual 32 clicks.

Eugene Gavrilov

>> kxcrtl -rv feature does not seem to stop the beeping

does it change the order the voices are allocated (in the analyzer window?)

/E

Rox271

>> does it change the order the voices are allocated (in the analyzer window?)

No Eugene. This version is the same as the previous one in all aspects.
/R

Eugene Gavrilov

>> No Eugene. This version is the same as the previous one in all aspects.

that is, when you execute 'kxctrl -rv' -- does it re-start allocating voices from the very beginning?...

//E

rubber_glove

>>that is, when you execute 'kxctrl -rv' -- does it re-start allocating voices from the very >>beginning?...

yes. the next voice allocated is at the beginning, but it is still necessary to start/stop up to the voice which started the beeping.
on another note, executing 'kxctrl -rv' just now during normal playback caused some strange behaviour. playback stopped for a second then became choppy - playing in short bursts with some silence in between...

-martin.

would it be possible to alter the -rv feature to reallocate to the current voice instead of the beginning? maybe this would at least save some clicking....

Eugene Gavrilov

>> playback stopped for a second then became choppy - playing in short bursts with some silence in between...

you should not run '-rv' when playing back audio -- this is a bad idea
ok, now we know that the problem is definitely unrelated to incorrect register programming since '-rv' resets -all- the per-channel registers...

/E

Rox271

Yes it starts from the right beginning but the beep remains.
Weird. So far the only fastest and most efficient way to cancel the beep is to play a MID file.
Is this to be considered as a clue?

Eugene Gavrilov

>> Is this to be considered as a clue?

no, since MIDI playback uses the same audio resources
it simply allocates audio voices faster than clicking 'play/pause' in winamp

so, we have at least two main reasons of the bug:
- a hardware issue (PCI bus, arbitration, latency...)
- a software issue: invalid OS logical/physical mapping

/E

Rox271

I just wanted to add something.
When the beep happens, I play a MID and the beep is cancelled WHILE still playing...

Sammydad1

Hi,

Are we talking about the "Hearing Test Tones" I get sometimes instead of normal audio playback ??

If so, I have had this happen every so often, for me it usually occurs after being in the AOL Tech Support chat room for a while. Everytime I should get some type of audio, I hear a "test Tone" like used in hearing tests.

Some times i can correct this by going into Windows Sounds Control Panel (XP Pro sp1) and resetting the audio Theme. Other times I have to reboot the system to get normal audio playback.


KX Rocks !!


Dave

Eugene Gavrilov

>> When the beep happens, I play a MID and the beep is cancelled WHILE still playing

that's very interesting
and does it happen when playing back regular wavs?
(that is, start WinAmp, reproduce the bug -- don't click 'pause' -- keep the audio playing, open a second instance of WinAmp (you may need to enable this in WinAmp setup), and now start/stop audio playback (in the second instance) & watch the analyzer)

/E

Rox271

I will try this when the beep occurs.
I'm not using winamp, only winmedia player and I think winmedia is only using DXSound. Also the beep would quite often occur in game. I'm not 100% sure but I can't remember the beep happening with MME or ASIO. So could this be a channel issue (I mean DXSound vs ASIO vs MME)?

Followup: ok. I had the beep bug from Netscape playing some sounds. I quitted Netscape, went in Soundforge and played a WAV that goes to "kx Wave CT4780 10k1 [ece0] 0/1". I assume this is a MME voice isn't it? Right now I can only see two red strips and the WAV is beeping.
I leave the WAV on and go in Win Media Player to play a mid. I can see nice green and blue strips dancing around the two red strips from left to right. The sound of both MID and WAV is corrupted. After a while, everything goes to normal all the sudden.

Anything particular you would like to know?

Eugene Gavrilov

>> So could this be a channel issue (I mean DXSound vs ASIO vs MME)?

winmm playback doesn't re-allocate voices on each start/stop, while dsound usually does
that's the only reason for the differences

>> Right now I can only see two red strips and the WAV is beeping.

then, definitely, this is winmm (since dsound would use 4 strips: 2+2)

>> The sound of both MID and WAV is corrupted. After a while, everything goes to normal all the sudden

have you noticed any logic in the abnormal<->normal transition? was there any 'cycle' of allocations (visible in the 'analyzer' window?)
was the wav file large enough to keep playing continuously? (or was it looped?)
>> "I leave the WAV'
-- you mean 'keep it playing'?

>> I can see nice green and blue strips dancing around the two red strips

I'm glad you find them nice I like watching MIDI playback sometimes

/E

Rox271

The WAV was large and didn't have a time to loop during the whole experimentation. I cannot see any logic. Sometime the beep is canceled immediately when I start playing MIDI. Sometime it would allocate all available voices before the beep is cancelled. But definitely playing a MIDI file sooner or later cancels the beep.

Eugene, there are some very weird things happening on voice allocations (or may be there is some logic that I cannot understand?).
At the right beginning, when playing a MIDI file, the driver would allocate all voices sequencially from left to right.
After some experimentation (hard to reproduce) the driver tends to save voices and use repeatedly only a few of them. This applies for all kind of voices even MIDI.

Also after playing MIDI file, the top of all voices is dark blue and if I now play something with WinMediaPlayer, there are only 4 voices allocated and each time I start/stop, the driver would always reuse the same voices.
I then stop all applications and do a "kxctrl -rv" to reset the whole thing. I notice that all dark blue square are still there and the voice allocation strange logic remains.

It looks like there is some wrong tagging on voices which remains and I never can go back to the initial clean situation.
This of course doesn't prevent the driver from playing correctly but since we're looking for weird things, I wanted to mention that one.

Yes.

There are many other good things in the driver

Eugene Gavrilov

>> there are some very weird things happening on voice allocations

well, they aren't 'weird'
there's very complex logic that is used for voice allocation in case of MIDI:
some voices can be looped, some are already stopped, some are 'released' or 'releasing' etc, etc, etc
the driver checks note status, state (released / stopped), its volume, its looping state etc.
that's why playing back MIDI files causes 'strange' voice allocations sometimes

so, there's nothing 'strange' or 'weird' -- except 'beeping'
btw, as I've already mentioned, Creative drivers use different voice allocation strategy, so, there's a possibility they've never met such a bug

/E

Salival

Hey there, this bug has been driving me crazy for a while now. Especially in games like max payne 2 where any sound that plays can possibly cause this bug to occur. Very annoying when characters are speaking and it sounds like they are cursing up a storm. Anyway, I'm here to try and help.

It seems that the discussion of this bug operates on the assumption that this bug happens when you start a new sound, or between two sounds? Now, I have no idea if this will help or not (which it won't if that was not your assumption), but based on what I have observed this bug occurs when a sound ends. I've come to this conclusion because while listening to music if another sound occurs during playback, such as a windows error sound, the beep sound does not occur until the error sound finishes playing and the music starts to beep. Hopefully that is useful.

My specs are:
P4 1.4Ghz, Intel i850 Rev. A2 , 768Mb RD-RAM, SBLive Value through SPDIF, Windows XP Pro

Let me know if you need anymore info on my specs or my experience with this bug. Thanks for the good work.

Eugene Gavrilov

has anybody tried to upgrade nVidia nForce1/2/3 drivers (the new one released in November)? did it affect the issue (if applicable)?

/E

DrGOA

Anyway, I am almost 100% sure that this bug is related to nVidia nForce chipset. Before 1 month I upgraded my motherboard with nForce one /the old was with VIA chipset/.
With VIA I had NO bugs and problems with all versions of kX.
Now I have.

Upgraded with latest nVidia nForce drivers /posting date: 03.11.2003/ - the bug still occurres

Rox271

Some followup...
I just upgraded my motherboard (Asus P4P800/ chipset 865 / FSB 800) and processor P4 2.4GHz overclocked 3.2.
There is now a PCI lattency setting in the BIOS set to 32. I also turned off multi threading which seems to cause many crashes.
So far, I haven't experienced the bug since that.
So it could be due to the 850 chipset, FSB 400 or just the processor speed, or questionable Dell motherboard?

Caithleann

I have been lurking here for a while in hopes that this problem would be resolved but apparenmtly it is a lot harder than I thought. Following Rox271 in his/her footsteps I run an ASUS motherboard as well:

ASUS P4C800 DeLuxe
Intel 875P Chipset
FSB 800 MHz
DDR400 2×256MB memory modules.
P4 - 2.4 GHz originally overclocked to 2.6 GHz but set back to normal. (No difference)
PCI Latency was 64, changed it to 160 after reading the Linux article on PCI Latency... no difference.
I have not yet turned off hyperthreading, I shall try that next and report accordingly.

I would like to add that I have noticed in this thread that most people seem to experience a single tone beep for either a protracted time or in regular intervals. In my case it sounds very different, as if sound samples loaded in soundbanks are played randomly, and loudly, at more than 8 times the original pitch/speed and this occurs completely at random, sometimes even when NO sound is being played at all. I shall attempt to capture a sound sample of this behaviour although to date I have not actually been able to do so through the normal recording procedure with Cool Edit Pro/Adobe Audition. I suggest that others who experience this malfunction do the same so all can hear what we are talking about.
The screeching seems to be more prevalent when a game has been loaded and ready to play than at other times. Also heard it once or twice during playback of mp3 data through winamp but never through Windows Media Player.
Hope this helps. I will report my findings on the hyperthreading issue later.

__________________
One cannot lie unless one knows the truth!

Rox271

Interesting.

>> PCI Latency was 64, changed it to 160 after reading the Linux article on PCI Latency... no difference.
Can you try to decrease the value down to 32 rather than increasing it? With my previous PC I noticed more stability and less beep issues with this value for all PCI devices.

>>P4 - 2.4 GHz originally overclocked to 2.6 GHz but set back to normal. (No difference)
I'm back to 2.4 too cos the CPU was a little bit too hot in my opinion for a normal long term use (2.4 :34°C/93.2°F and 3.2: 44°C/111.2°F)

>>I have not yet turned off hyperthreading, I shall try that next and report accordingly.
I think I read somewhere that kx drivers were not compatible with hyperthreading.
Also I could read that hyperthreading is not recommanded for gaming. Anyway, on my PC, I got blue screens...

>>I would like to add that I have noticed in this thread that most people seem to experience a single tone beep for either a protracted time or in regular intervals. In my case it sounds very different, as if sound samples loaded in soundbanks are played randomly, and loudly, at more than 8 times the original pitch/speed and this occurs completely at random, sometimes even when NO sound is being played at all.<<
I agree. The Soundfont beep is very different from the WAV/MP3 beep. I was under the impression the soundfont had a amplitude modulation while WAV/MP3 really sounds like short sample loops.

>>The screeching seems to be more prevalent when a game has been loaded and ready to play than at other times.
I agree too and I suspect it has something to see with DirectX.

Caithleann

OK.

I disabled hyperthreading and set PCI latency to 32 and have not heard any screeches coming from the sound card since. Not even during my games (Visual Pinball)

It is too early to hang out flags or be too exuberant, so let's see what happens in the next few days but if nothing happens, I would venture to say that hyperthreading and PCI latency are definitely something to look at, Eugene.

If the noise comes back I will be too.

Thanks for the trigger post, Rox271.

Caithleann

__________________
One cannot lie unless one knows the truth!

Rox271

Glad if I could be of some use Caithleann ;-)

>>It is too early to hang out flags or be too exuberant, so let's see what happens in the next few days but if nothing happens
You're wiser than I

I've seen that Eugene already adressed some hyperthreading issues in the v3536c but I'm not anxious to try this since many other applications will crash.

After a few days, my system is incredibly stable and I haven't experienced the bug once since I got the new PC.
The only thing I could hear is the sound crackling some time (mainly in games) but all the rest is damned stable.

After one year of beeps, what is now the conclusion? Well hard to say. A fast driver requires a fast PC and a properly balanced PCI setting?

Caithleann

As I stated in my last post, it was too early to pass any judgement on the quick work-around of disabling hyperthreading and reducing the PCI latency to 32.

Now I can say that killing the hyperthreading does indeed have adverse effects on some programs which run at less than half their original speeds notably the Eudora Email program which is reduced to a sorry 10% of its original speed. Whether sending or receiving; one is put in a task queue for more than 20 seconds before anything useful happens and downloading 20 emails takes almost 5 minutes. Naturally that is unacceptable.

Other programs, notably video codecs appear to have been negatively impacted by disabling hyperthreading. I am not sure why hyperthreading was included, but I know that Pentium 3 processors without hyperthreading perform beter than a Pentium 4 with hyperthreading disabled. So, it seems hyperthreading is more hyperbole than a real improvement. Nevertheless, I need to re-enable it or this Pentium 4 behaves too much like a 500 MHz Pentium II. Even opening a windows Explorer takes more time now.

If anyone can explain what hyperthreading really is I would love to hear it. Thanks.


Caithleann

__________________
One cannot lie unless one knows the truth!

Rox271

Well, I'm not an expert but this is my understanding (and personal opinion).

As far as I understand, hyperthreading is like emulating two CPUs. So the computer is capable of executing two things at the same time. There is plenty of information here Intel

Before this new technology, CPUs were already prediciting what would be the next execution steps and executing programs in adavance. Of course, if the program was then branching, we were loosing the benefit of this in advance execution. Basically the improvement was only effective in some very special cases and can only be measured statistically Intel is certainly doing a lot of marketing; I think that hyperthreading is only available in some special occasions, this is not actually like having two CPUs. Therefore the gain is probably statistical, we are not talking about raw performance.

I noticed that my L1, L2 and L3 cache memory amount is very poor compared with AMD. That could explain poor performances as well. Have you tried overclocking? My PC is very stable with a 30% overclock.

You can also try to reactivate Hyperthreading and see if the beep comes back.

Caithleann

Ok,

I re-enabled hyper threading and the noise is back in full force. so We may now say with certainty that (at least in my case) the trigger is hyperthreading.

Eudora is back at full speed too as are other programs developed after the Pentium 4 release.
Not only is eudora faster in retrieving mail, it also sends it much faster again.

As I have stated before, I would have include a sound sample of the horrid noise as soon as I was able to capture it but notice that despite this forum's slick appearance the upload capability has been disabled. That puts a severe handicap on the trading of such data, Very sad and unconducive to quick solutions to problems

A copy of the sound sample may be downloaded from my own site

__________________
One cannot lie unless one knows the truth!

Rox271

Your beep is very different from mine.
On my side it was more like a continuous sound.
I'm not sure this is the same issue that was described in this thread.