2013-11-18 19:26 UTCThe mixer report does not appear to be complete. Please get the mixer report immediately following an overrun and email the results to nbsweb at neurobs dot com.Some things you could do to try and reduce the occurrence of buffer overrun/underrun is to increase the 'Rewind Distance' to 30ms. This is in Settings Audio Advanced Settings Rewind Distance.
Bu tip programlar 'kapasite' icin yazilirlar, problemler veri tabanina daha hizli erisim, sayfa hizli yukleme gibi seyler etrafinda doner. Son zamanin Internet Java paketleri, sanki siteleri bir kisi icin yaziyormus gibi yardim eder size; ondan sonra kopya kagidindan cikarmis gibi 10,000 kisi icin ayni kodu kullanabilirsiniz. Hammerhead Rhythm Station latest version: A free Audio program for Windows. Part of the category Audio software and has been created by Threechords.
You should then set the minimum start time of your sound events to 30 or higher. This will allow Presentation time to write the audio to the sound stream far ahead of where the stream is playing from in memory. You should read the section of our help file for some information regarding scheduling delays and other playback timing considerations.We recommend against DirectX Software mode in Windows 7 because it is very laggy and jittery. You should instead try Presentation Mixer Shared mode to increase reliability, which offers the same consistent and accurate timing as Exclusive mode while only adding some scheduling delay. Scheduling delay can be mitigated by using a 'time' 'Rewind Distance' for sound events in your trials.If you want multiple sounds to play at exactly the same time, if possible, you should mix them into one single audio file using a program like Audacity (free). If you can't do that, you should at least check the logs to make sure they played at the correct intervals relative to each other.
Don't trust the logs if you are using DirectX Software mode.Above all, you should empirically test the latency and jitter of your audio device in each of the audio modes, if audio timing is important to you. Information on how to test sound cards can be found atFleur Bouwer. 2013-11-19 20:09 UTCOk, so I tried changing the Rewind Distance to 30 ms.
The first rhythm then started at 50 ms and subsequent rhythms even later. The program still crashed. I tried this both in the Presentation Mixer Exclusive and Shared modes.
In both modes the program crashed. Now this may not be very surprising given that it mentions in your own documentation about Rewind Distance: 'This parameter is not used on Windows Vista/7/8 where a minimum distance is always used.' Just to clarify, my questions are not about which mode to use to get good latencies.
I do know that the latencies are crappy with DirectX Software, but as the program seems to crash in any other mode, I do not seem to have a choice. So I would like to know why the program crashes for two reasons:1. Is it indeed the case that this bug will not appear (or be less likely to appear) in DirectX mode? As I mentioned this seems to be the case, but as the program does not always crash in other modes, I cannot be certain that this indeed solves the issue.2.
Is there any other way (as the Rewind Distances did not help) to solve this bug and still use the Presentation Mixer?Thanks for your help!Peter Pebler. 2013-11-20 13:32 UTCWith the Direct X Software mixer, you will never get an error of this type. The sound production latency will just be really bad. For example, the sound might actually start 20-40ms later than the Presentation event, though I can't say exactly for a particular system. If that kind of delay doesn't matter for your experiment, then you could use this mode and never worry about these errors.I apologize for the confusion about the rewind distance. The documentation is correct.When you use the Presentation Mixer, Presentation mixes audio data into a buffer from which the data is sent to the operating system at the requested time.
Most of the time, Presentation mixes data well ahead of the data that is playing so that even large system interruptions do not cause an 'overrun', where the play position reaches a point where data hasn't been mixed yet. However, when the mixer is told to play something as soon as possible, it goes back as far back as it can and mixes in the new data. At that point, it is more susceptible to large system interruptions.The mixer report indicates a couple of troubling problems. First, there are for some reason very large interruptions, some on the order of 100-200ms. Second, there are warnings that the play cursor position given by the sound card driver goes backwards in some cases.
The latter problem is not always fatal, but probably indicates a very poorly written sound card driver. It's not possible to tell what the source of the large interruptions is, but if no other likely causes are present (e.g.
Exceeding memory capacity with stimuli, other processing running, etc.) that could also be the sound card driver doing something strange.If possible, I would recommend trying a different sound card. You definitely don't want Presentation pausing for 100-200ms; even if the mixing doesn't crash, that will still impact response and other event timing. Using a multi-core CPU machine might also help.Regarding minimizing overrun errors in cases where for some reason interruptions are unavoidable, you can actually do this manually by scheduling auditory stimuli with delays. This may not be possible for your experiment, but in the case where no other stimuli are presented shortly before a sound, you can avoid telling Presentation to play a sound as soon as possible. To do this, make sure there is some interval between the requested sound start time, and either the previous stimulus or the start of the trial. For example, if there's a 300ms interval before a sound with no other stimuli, instead of placing that 300ms in a previous trial and putting the sound at time 0 in the next trial, end the previous trial early, and have the sound at time 300ms of its trial.Ed Amorim.
2016-09-25 15:49 UTCHi!I’m using presentation for a mismatch negativity project, but the study is crashing half-way through. There is an error pop up coming up, so I was hoping to get some advice on how to solve this problem.#Error message:'Circularlockbuffer stream thread overrun.”The timing of is error pop up seems to be a random.This same error was discussed in the forum in 2013 and commented on above, but I could not find any final recommendations there however.Many thanks for the help!EdMike Blank.