Yeah, I've gone through various different optimizations for windows.
Yeah, like I mentioned, I tried it and it didn't work, only works on windows HD audio.
how to reduce audio latency?
- Thread starter deama
- Start date
marcelnl
500+ Head-Fier
- Joined
- Sep 13, 2009
- Posts
- 516
- Likes
- 95
https://audiophilestyle.com/forums/...-server/page/48/?tab=comments#comment-1095998
load of links to trimming down Windows can be found, I cannot say what is worthwhile as I left W for audio some time ago.
load of links to trimming down Windows can be found, I cannot say what is worthwhile as I left W for audio some time ago.
I posted there, one guy replied saying to decrease buffer size which would decrease the latency according to latencymon, which it does, however it won't result in much of a change as the problem lies in windows itself.
Another guy replied that was I think the first person to actually understand what I was trying to do, however he said there was nothing I could do, at least on windows.
I checked out the thread in general, but they seem to be mostly talking about hardware stuff, or literal latency inside of their computers, and I think their target is DAW or audiophile software/programs rather than everything.
I'm out of options here, the only thing I can think of short of using linux is trying to find some sort of hacker or dissassembly guy and pay him a bunch of money to modify windows' MME/WASAPI/WDS drivers.
Last edited:
marcelnl
500+ Head-Fier
- Joined
- Sep 13, 2009
- Posts
- 516
- Likes
- 95
I have seen your post and the responses there, it IS an audiophile community geared towards folks using a computer as source, I think it is more helpful if you ask further questions or try to clarify what you are trying to accomplish. The thread in which you posted is about building a (high end) music computer server.
To me it still reads as if you are trying to do something that is impossible, reaching 0 latency within a computer is impossible, let alone reaching 0 latency from origin of the impulse to your ears, heck even your brain has latency and that latency is different across people.
That forum has a lot of folks from the industry participating, many playing at the top of the league.So if the information you are seeking is available you'll be able to get to it.
I may be misundersatnding what your objective is, perhaps it is just getting rid of the crackles and pops (which I am sure are very annoying).
From your initial post here it appears to me that the manufacturer fo your DAC may be your best bet for the issue your experiencing, they should know what buffer size works best and what to check if the crackles and pops remain even when all settings are OK. Is there a forum with a section dedicated to this DAC?
To me it still reads as if you are trying to do something that is impossible, reaching 0 latency within a computer is impossible, let alone reaching 0 latency from origin of the impulse to your ears, heck even your brain has latency and that latency is different across people.
That forum has a lot of folks from the industry participating, many playing at the top of the league.So if the information you are seeking is available you'll be able to get to it.
I may be misundersatnding what your objective is, perhaps it is just getting rid of the crackles and pops (which I am sure are very annoying).
From your initial post here it appears to me that the manufacturer fo your DAC may be your best bet for the issue your experiencing, they should know what buffer size works best and what to check if the crackles and pops remain even when all settings are OK. Is there a forum with a section dedicated to this DAC?
Last edited:
sajunky
Headphoneus Supremus
Did you try WASAPI event driven mode? From the description of pops/cracks I suspect that your USB transfer mode is synchronous. In the asynchronous mode effect of breaking synch is different, there is a half second interruption of music, but no pops.
Maybe you should take a look at Foobar 2000, you will see that default buffers for WASAPI event are much smaller than WASAPI push. It means a lower latency. The first step is to enable WASAPI exclusive mode for bit-perfect transfers. It is done in a Windows Sound Control Panel. A data path will be shorter, bypassing Windows mixer&resampler. It will give a similar path to the ASIO, but giving a control over USB transfer mode. Try both push/event WASAPI output devices in Foobar, you can adjust buffers in the advanced settings tab.
Quality of USB cable also matters. Use a short cable as possible, test the cable in WASAPI push mode for the presence of cracks&popups, to see which one accept smaller buffers. Put ferrite clamps on the both ends of the cable to see whether it helps. When finished use WASAPI event mode.
Note that this setting for WASAPI event may be ignored. (?) A clue can give a link below indicating that Scarlett Gen3 device unlike most of UAC class 2 devices do not allow to create explicit feedback endpoint, while Gen2 did. At the moment I don't understand exactly what it means (why a question mark), just passing it through: https://linuxmusicians.com/viewtopic.php?t=20669&start=15
I am not a latency freak, but I remember that Windows XP was giving me constant 20 microseconds latency, with jumps to 30us maximum. Now with Win7/10 it is 70us at minimum and jumps every 10 second to 500us, Win10 is much worse, any network activity cause page fault. The same happens on a number different PC's.
Maybe you should take a look at Foobar 2000, you will see that default buffers for WASAPI event are much smaller than WASAPI push. It means a lower latency. The first step is to enable WASAPI exclusive mode for bit-perfect transfers. It is done in a Windows Sound Control Panel. A data path will be shorter, bypassing Windows mixer&resampler. It will give a similar path to the ASIO, but giving a control over USB transfer mode. Try both push/event WASAPI output devices in Foobar, you can adjust buffers in the advanced settings tab.
Quality of USB cable also matters. Use a short cable as possible, test the cable in WASAPI push mode for the presence of cracks&popups, to see which one accept smaller buffers. Put ferrite clamps on the both ends of the cable to see whether it helps. When finished use WASAPI event mode.
Note that this setting for WASAPI event may be ignored. (?) A clue can give a link below indicating that Scarlett Gen3 device unlike most of UAC class 2 devices do not allow to create explicit feedback endpoint, while Gen2 did. At the moment I don't understand exactly what it means (why a question mark), just passing it through: https://linuxmusicians.com/viewtopic.php?t=20669&start=15
I am not a latency freak, but I remember that Windows XP was giving me constant 20 microseconds latency, with jumps to 30us maximum. Now with Win7/10 it is 70us at minimum and jumps every 10 second to 500us, Win10 is much worse, any network activity cause page fault. The same happens on a number different PC's.
Ok, I have no idea how else I'm supposed to explain this, to me it is very much clear what the issue is and what I want from my first post, and I even clarify it by reposting to other posters.
Guys, for the last time, I do not get pops and cracklings from running my DAC by itself , I get it when I use the voicemeter software to try and reduce latency because I haven't found another way to reduce total audio latency. My goal is to reduce total playback audio latency.
Please, please, please, read the above paragraph very carefully and slowly.
Guys, for the last time, I do not get pops and cracklings from running my DAC by itself , I get it when I use the voicemeter software to try and reduce latency because I haven't found another way to reduce total audio latency. My goal is to reduce total playback audio latency.
Please, please, please, read the above paragraph very carefully and slowly.
bfreedma
The Hornet!
- Joined
- Feb 3, 2012
- Posts
- 3,320
- Likes
- 2,701
It would be helpful to understand why you are trying to accomplish this and what the desired end state is - not just getting to 0 latency, which is not technically possible and actually not desirable. Frankly, it feels like you're chasing something without knowing why. What issue is the current latency level causing? Why do you believe reducing latency will solve it?
Note - I have read your paragraph very carefully and slowly. Sorry, but your lack of sufficient explanation is the issue here, not the attempts to understand by others.
marcelnl
500+ Head-Fier
- Joined
- Sep 13, 2009
- Posts
- 516
- Likes
- 95
what he ^said!
Please read the responses carefully, I think we are miscommunicating.
From what you just wrote it may seem that it could be something silly like having the output level far too high causing clipping.
Please read the responses carefully, I think we are miscommunicating.
From what you just wrote it may seem that it could be something silly like having the output level far too high causing clipping.
castleofargh
Sound Science Forum Moderator
- Joined
- Jul 2, 2011
- Posts
- 10,535
- Likes
- 6,183
Could be, but It's also very likely that his audio artifacts are indeed a result of a buffer issue(not leaving a big enough buffer for some digital operations). I got that sometimes with VSTs in an app or a VST host, and needed to change the buffer settings to remove the artifacts.
sajunky
Headphoneus Supremus
Guys, we all do not understand @deama, so please don't make pressumptions. Voicemixer that gives a trouble is a kind of applictation for playing interactive games: https://vb-audio.com/Voicemeeter/
What I understand from a setup, there is a video game application that is outputing sound to the default Windows sound device. When making Voicemixer default sound device, there are cracks and pops.
Now there are number of things I don't understand. What is A1? Where headphones are connected to? I assume that a Voicemixer output is a Scarlett DAC using ASIO. Then headphones and a mic are connected to the Scarlett and a microphone output from Scarlett goes back to PC using a different USB cable. Right? Or the the same USB cable feeds microphone back? Good to know such details. And a total round-up delay is important. I intentionally don't say "latency" as it can be confused with DPC latency.
In any case a microphone input device shouldn't go back to the Windows Mixer nor a Voicemixer (unless it has special settings), as it creates a loop. It should only feed the game application that sends it to the other player.
What I understand from a setup, there is a video game application that is outputing sound to the default Windows sound device. When making Voicemixer default sound device, there are cracks and pops.
Now there are number of things I don't understand. What is A1? Where headphones are connected to? I assume that a Voicemixer output is a Scarlett DAC using ASIO. Then headphones and a mic are connected to the Scarlett and a microphone output from Scarlett goes back to PC using a different USB cable. Right? Or the the same USB cable feeds microphone back? Good to know such details. And a total round-up delay is important. I intentionally don't say "latency" as it can be confused with DPC latency.
In any case a microphone input device shouldn't go back to the Windows Mixer nor a Voicemixer (unless it has special settings), as it creates a loop. It should only feed the game application that sends it to the other player.
miketlse
Headphoneus Supremus
- Joined
- May 8, 2016
- Posts
- 6,020
- Likes
- 4,015
Yes latency can be a very confusing term - possibly delay or synchronisation are better alternatives.
I keep wondering if he is referring to something similar to 'lip sync', where there can be issues if the game image and audio are not in sync - however many video players enable you to change the delay to suit.
In contrast knowing the delay between audio leaving the hard disk, and being heard in your headphones, is impossible and you wouldn't worry if both the left and right channels are in sync (providing they are also both in sync withe the video image).
I keep wondering if he is referring to something similar to 'lip sync', where there can be issues if the game image and audio are not in sync - however many video players enable you to change the delay to suit.
In contrast knowing the delay between audio leaving the hard disk, and being heard in your headphones, is impossible and you wouldn't worry if both the left and right channels are in sync (providing they are also both in sync withe the video image).
Users who are viewing this thread
