View Full Version : LATENCY . . . . !!??!!
mihmar
07-16-2000, 08:10 AM
...\"GigaStudio consistently delivers latencies in the 3-10 millisecond range for all voices, not just the first ten or so.\"...
Yes, this is from the NemeSys WEB page where the GigaStudio features are presented.
So I bought the upgrade to GST160, I bought the WaveCenter/PCI card (is on the list!), a very fast dedicated computer (as specified) and I made all the Windows tunings and.... guess what?
I got latency of 17..20 ms!! It is UNACCEPTABLE!!
Any comments, anybody?
My system: P3/700MHZ, 256MB, Maxtor7200/2MB dedicated drive, WaveCenterPCI card, Winman 1x1 midi interface, Windows95.
I did the measurements triggering from a separate PC the same note once to GSt (loaded with a minimal drum kit) and then to a hardware module. Then I recorded the two audio tracks to Cubase VST (panned one to L and other to R) and used Wavelab to measure the gap between the to signals.
mihmar
mihmar
07-16-2000, 10:15 PM
Thank you very much PaPaChalk for your response.
I am pretty much introduced into the computers field (PCs), so I can say for sure that everything is well installed, clean, no conflict at all. All the optimizations are also done.
I am not using Cubase (or any other sequencer) at all on this computer. Only the GSt. I have a second PC running Cubase/Wavelab etc
There is a setting at WaveCenter/PCI panel, it\'s called \"ASIO buffer size\". This one definately affects the GSt behaviour. But I thought that ASIO was for the Cubase (etc) and not for the GSt, which uses the GSIF interface driver.
So the GSt only work well with buffer=256 samples (that coresponds to latency=11ms, but I am getting near 20 ms!!). If I put buffer=128 or even 64 samples, then yes, the latency goes down but lots of pops/click occures.
I have the latest WC/PCI driver (under Win95), I don\'t know what\'s happening.
One more thing: running the GSt diagnosis I noticed that ALL WC/PCI outputs \"...did not map...\". What does it mean? Because except for the big latency, the system runs well.
mihmar
[This message has been edited by mihmar (edited 07-17-2000).]
PaPa Chalk
07-16-2000, 11:53 PM
You probably have a driver problem. I do not get the same results at all. Check your audio properties in cubase I would suspect there is a problem there. What is the latency of the card you are recording too?
in cubase. Is it set to MME or Asio. there are many factors which may contribute to your
problem.
I have a similar system setup. PIII 733MHZ(550 coppermine O/C) 256Mb ram, cubase etc.
soundforge is our editor of choice.
Restart your computer and go into safe mode by pressing F8.
Then go into control panel then click on system. Check to see if there are any double midi drivers in your device manager.
We do not get what you\'ve described on any system we have here.
http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif
PaPa Chalk
[This message has been edited by PaPa Chalk
[This message has been edited by PaPa Chalk (edited 07-16-2000).]
SteveMitchell
07-17-2000, 11:19 PM
Hmmmm. Maybe I\'m missing something in the discussion, because 17ms, that is 17/1000\'th of a second doesn\'t seem to be a significant error in terms of MIDI. I introduce variances far greater than that intentionally, just to keep the track from sounding mechanical. What am I missing here?
Steve Mitchell
The Classical MIDI Resource (\"http://www.classicalmidiresource.com\")
mihmar
07-17-2000, 11:41 PM
Yes, Steve, it is very audible!
I agree with you that sometimes music needs to be unquantized but if you are doing for example a simple drum track (a kick drum) you can hear the timing gap between the metronome and the drums!
So again: 17ms (or ocazionaly even 20ms) is not (always) acceptable!
mihmar
tinerelu
07-18-2000, 04:59 PM
Oh, I am having about the same problem.
I can notice a latency while playing the piano.
Is anyone else encounting the same trouble?
tinerelu
PaPa Chalk
07-19-2000, 01:00 AM
<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by mihmar:
I have the latest WC/PCI driver (under Win95), I don\'t know what\'s happening.
One more thing: running the GSt diagnosis I noticed that ALL WC/PCI outputs \"...did not map...\". What does it mean? Because except for the big latency, the system runs well.
mihmar
[This message has been edited by mihmar (edited 07-17-2000).]<HR></BLOCKQUOTE>
From what I see in your quote here there is a problem with the sound card configuration I would try to reinstall gigastudio. You should not get high latency at all.
http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif
PaPa Chalk
[This message has been edited by PaPa Chalk (edited 07-19-2000).]
NAZARU
07-20-2000, 04:11 PM
Papa Chalk,
Now that the price of PIII is more reasonable I am more apt to try overclocking at the risk of frying my CPU. What are the prospects of overclocking my PIII600 and what does overclocking do to the life expectancy of the CPU? I have and ABIT BE6II motherboard and 256 Megs of ram. Would I need additional fans for my ATX case? What can a PIII600 be overclocked to? Peace.
Simon Ravn
07-20-2000, 05:02 PM
If it\'s a regular P3 600 (that is, not a Socket 370, P3 600e or EB) it probably can\'t be overclocked at all. However my P3 500e is running 700Mhz without any trouble with the out-of-the-box cooler. Yes it will shorten the life of the CPU - maybe my P3 will only last 5 years instead of 10. But since I probably have to upgrade the system in 1-2 years time it is no big deal.
And there\'s always signs that can tell you if you\'re overclocking too much: crashing while loading Windows, instability in general etc... The biggest problem is the overclocking also affects the AGP and PCI speeds - if you have a mobo with a VIA Apollo chipset it is not so much of a problem since it has both 2/3 and 1/2 dividers for the AGP clock frequency - AGP is supposed to run 66mhz, and I am running 140Mhz FSB to get my P3-500 to 700Mhz - that means the AGP is now running 140/2=70 mhz - not exactly a problem for any AGP video card. But if you have a BX or other Intel chipset it would be running 140Mhz*0.66=92 Mhz which could cause trouble for the card. Another thing is that your RAM will be running 140Mhz on an Intel chipset too, but on the VIA Apollo you can set the RAM speeed to FSB-33=107Mhz which should work for all PC100 RAM.
PaPa Chalk
07-20-2000, 05:07 PM
We, also nemesys does not endorse overclocking.
I have done it myself but I wouldn\'t recommend it to anyone as with all things there is risk you could fry your board. You really have to know what you are doing. Here are sites that have great info on overclocking.
http://www.ugeek.com/procspec/overclk.htm (\"http://www.ugeek.com/procspec/overclk.htm\") http://www5.tomshardware.com/guides/overclocking/index.html (\"http://www5.tomshardware.com/guides/overclocking/index.html\")
http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif
PaPa Chalk
[This message has been edited by PaPa Chalk (edited 07-21-2000).]
djuth
07-21-2000, 07:36 AM
mihmar wrote:
One more thing: running the GSt diagnosis I noticed that ALL WC/PCI outputs \"...did not map...\". What does it mean? Because except for the big latency, the system runs well.
-----
I have this same thing showing for my Gina. I sent the diagnostics to Nemesys but have not yet heard anything. Gigasampler worked fantastically on this same machine.
Does anyone have any idea why this happens?
PaPa Chalk
07-21-2000, 08:43 AM
Check your control panel setting for the Gina most likely something is set incorrectly there.
http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif
PaPa Chalk
Cool7s_Dad
07-21-2000, 02:09 PM
Hey Papa:
I also show the same problem on my Gina. In addition, in the Harware/Routine screen it says \"GSIF: Yes\" but in the diagnostics text it says something like:
GSIF Compatible Cards Found: 0
I too have sent this off to Nemesys and have yet to receive a response. But I think it\'s more of a problem with the diag routine than it is with the Gina, drivers, or GSt itself.
Peace,
Tim
http://www.elithic.com (\"http://www.elithic.com\")
PaPa Chalk
07-21-2000, 05:05 PM
Try upgrading to 2.01 . from what you describe here. This sounds like a configuration issue. check your gina control panel. If your Sequencer is using asio make sure you don\'t have it set to directx for the card. try setting it to MME.
PaPa Chalk
<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by Cool7s_Dad:
Hey Papa:
I also show the same problem on my Gina. In addition, in the Harware/Routine screen it says \"GSIF: Yes\" but in the diagnostics text it says something like:
GSIF Compatible Cards Found: 0
I too have sent this off to Nemesys and have yet to receive a response. But I think it\'s more of a problem with the diag routine than it is with the Gina, drivers, or GSt itself.
Peace,
Tim
http://www.elithic.com (\"http://www.elithic.com\") <HR></BLOCKQUOTE>
Tim,
I\'m using a Gina but am totally unfamiliar with all that is being said in this discussion.
Can you point me to some reading material about GSIF...and where is \"Hardware/Routene\"
screne that you\'re talking about?
Thank you,
Steve
Cool7s_Dad
07-24-2000, 11:04 PM
Sorry about the typo, it was supposed to be \"Hardware/Routing\" not \"Hardware/Routine.\" I guess I\'ve been programming too long http://www.northernsounds.com/ubb/NonCGI/images/icons/wink.gif
Papa, I\'m pretty sure the Gina\'s configured correctly. I have DirectX support turned off. I\'m using the \"professional\" not \"consumer\" mode. The only output pair I enable in GSt is 9,10, which is the SP/DIF output.
I\'m not sure this is the source of my \"sluggishness\" problem. I rather doubt it. Since that problem is not accompanied by any clicks and pops, but just sluggish response to MIDI (even at very low polyphony), it\'s probably something else. This just happens when I use certain instruments... the AO string samples in particular. I\'ve got some incredibly complex scores that don\'t use AO very much and it doesn\'t show up with them.
Anyway, the GSIF thing is baffling, but as I said, I suspect it has more to do with the diag routine and not the recognition of the GSIF capabilities of the Gina Card. I\'m getting 150 voices of polyphony on some scores without any problems. I haven\'t thoroughly tested Nemesys\' suggestion of disabling all the MIDI on the SB Live! card that\'s also in the machine with the Gina. I\'ll let you know in a couple of weeks (since we\'re going on vacation to Ouray, CO for 10 days http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif ). If it doesn\'t help, then I\'ll just disable the whole SB Live! card. I may even take the dang thing out of my machine, since this machine is dedicated purely to GSt.
Thanks (as usual you\'re a gem),
Tim
http://www.elithic.com (\"http://www.elithic.com\")
Cool7s_Dad
07-24-2000, 11:07 PM
Steve:
The main screen we\'re talking about is under \"Settings\" in GigaStudio. Then click the \"Hardware/Routing\" tab. We\'re also referring to the \"Diagnostics\" procedure you can run that creates a complete report.
Peace,
Tim
http://www.elithic.com (\"http://www.elithic.com\")
djuth
07-26-2000, 07:13 AM
<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Anyway, the GSIF thing is baffling, but as I said, I suspect it has more to do with the diag routine and not the recognition of the GSIF capabilities of the Gina Card. <HR></BLOCKQUOTE>
I believe you\'re right about the problem being the diagnostics routine. My latency issues are nil except when using the 1gb piano, and then they are sporadic (more RAM beyone my current 128mb and replacing my Celeron 300a with a PIII ought to do it; everything else on my machine (two IBM 7200 rpm drives, etc.) seems OK). Also, in Gigastudio, it says right at the top of one of the screens \"Even Gina - GSIF compatible,\" or something like that. I mean, it works as well as Gigasampler worked on this same machine.
Of course, I\'ve got to rail on Nemesys tech support a little. I did in fact get a response from them in a reasonable amount of time, but it basically said \"disable DirectX for you Gina and your Montego, and get the latest drivers.\" Already did that on day one http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif
BTW, anyone know what\'s up with http://www.echoaudio.com/ (\"http://www.echoaudio.com/\") ???
<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by tinerelu:
Oh, I am having about the same problem.
I can notice a latency while playing the piano.
Is anyone else encounting the same trouble?
tinerelu<HR></BLOCKQUOTE>
Using just 1 piano note...
Yes...the latency...although greatly improved
over other potential situations, is still much more noticable than triggering a hardware based sampler...but...I can compensate for the slight echo,
...but then again, I\'m a guitarest that \"programs\" my key parts, not a keyboardist.
Steve
AMD-k6-2-400, 128 mb ram,
IBM DTTA, 9ms, UDMA 33.
djuth
07-28-2000, 09:46 AM
Alright now I\'ve discovered something that some of you may already know but others may find helpful. This sporadic latency issue with my system (Gina soundcard, DirectX disabled, blah blah blah) that occurred only on large samples like the Gigapiano led me to believe that I needed more RAM and a PIII. However, last night I replaced my basic IDE cable with an Ultra ATA/66 cable for my Gig drive (second drive, IBM 75GXP). Even though both of my 7200 rpm drives are running UATA 33, this cable COMPLETELY SOLVED all latency problems. I can\'t believe it! I loaded the Gigapiano and the Trachtman Steinway together and I could play either one flawlessly. Diskbench confirms the improvement -- from sustained 19mb/s to almost 30mb/s. That\'s amazing.
So, give this a try if you\'re system is up to spec in every other way.
[This message has been edited by djuth (edited 07-28-2000).]
[This message has been edited by djuth (edited 07-28-2000).]
Powered by vBulletin® Version 4.1.10 Copyright © 2012 vBulletin Solutions, Inc. All rights reserved.