All times are UTC+02:00




Post new topic  Reply to topic  [ 89 posts ]  Go to page Previous 1 2 3 4 59 Next
Author Message
PostPosted: Tue Jun 15, 2021 5:06 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 10270
Quote:
So that's not good, here MME on output works from 484 samples and up (48kHz) KS, wasapi, works just fine. PA shoul'd not do that, we need exact numbers. I hope that does not mean if i set 960 it makes it 760 (?)
I'm not really sure, it might. I have added a different way of controlling it now, but I'm not too sure what's happening internally. If this next build doesn't fix things I'm going to try to get help from the PortAudio developers - but they only appear to have an email list (welcome in 1990...)

(Note that none of this really matters as long as things are stable - and on most sound cards/OS's they are, but not on all. If you really want low latency, you need to use ASIO anyway). Also, on my new pc, the same sound card that caused issues before with input doesn't anymore, so that's just weird. There might have been a bug in an earlier Windows 10 version.


Top
   
PostPosted: Tue Jun 15, 2021 10:21 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 10270
BETA014 posted. The "Override" thing is now a pulldown-menu where you can choose between 2 different methods of determining block size. I hope that this will work better now. On my pc it does - but I'm not too sure about this yet.


Top
   
PostPosted: Tue Jun 15, 2021 11:02 pm 
User avatar

Joined: Tue Mar 17, 2009 2:56 pm
Posts: 3762
BETA014*

And yes, MME is working now "via blocksize"! Still buffer does not start from middle tho.

_________________
control point


Top
   
PostPosted: Tue Jun 15, 2021 11:08 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 10270
Quote:
BETA014*

And yes, MME is working now "via blocksize"! Still buffer does not start from middle tho.
Ah, ok. So does it work on all sizes now or just 480? It does start in the middle here.
Note: I'm not really sure what "block size" really does, in other words: How useful it is. Things are really really confusing (and apparently, purposely so - at least the documentation says that things are "purposely vague", yuck).


Top
   
PostPosted: Wed Jun 16, 2021 12:31 am 
User avatar

Joined: Tue Mar 17, 2009 2:56 pm
Posts: 3762
Quote:
Quote:
BETA014*

And yes, MME is working now "via blocksize"! Still buffer does not start from middle tho.
Ah, ok. So does it work on all sizes now or just 480? It does start in the middle here.
Note: I'm not really sure what "block size" really does, in other words: How useful it is. Things are really really confusing (and apparently, purposely so - at least the documentation says that things are "purposely vague", yuck).
Definetly means something.
In BAOne when i test input or output on block size, let's say i have 480samples with 8% jitter error, that is just fine. When i set some wrong settings, like, 441samples, i have Jitter error 32000% but audio is still fully OK! But when i set 256samples i have 45000% jitter error and audio is then skippy and bad. So acording to BA, up to 30000% jitter error is actually fine. But is it?
I guess that's why majority thinks it means nothing since you have so much room for error. What i think is just a fine tuning for lowest possible latency, and it really means there.
Idk what is difference in this 2 modes in ST but definetly this new one works with MME now. Yes, works with any number now even with 160samples (idk how is that even possible tho).
Also i definetly see better or worse input sync with different block size.
Just tested my USB headphones and 460 or 960samples should be perfect, however 512 and 256 samples are actually perfect. so.. idk anymore :D How do i know, input sync literally do nothing and locked to M1.0000 and x1.0000. Damn also buffer seems to start form center now.
OK, 256 512 or 1024 samples is best for some reason, with 0.05% relative adjust, everything more than then is overfast sync.

Q: Can we have samples start from 128 atleast, maybe even from 64?

edit: damn, Audio Quality makes difference from 100-150% with input sync behaviour. 150% waaaay better and more stable.

_________________
control point


Top
   
PostPosted: Wed Jun 16, 2021 11:48 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 10270
Update from here. I figured that I probably did my earlier tests with my webcam microphone on my old pc, and indeed, on my new pc (same webcam, both Windows 10, but the old pc is running an older Windows build) there are no issues with it. So, most likely, whatever was wrong was a Windows issue. Having said that, that doesn't mean that it's not a problem.

So I went back to my old pc to see what the behavior with the latest build is, and it's at least better than it was.

Wasapi
Old build: Any Overrule setting that wasn't a multiple of 320 failed.
New build: Both with Suggested and BlockSize it works on any setting.

Kernel Streaming
Old build: Any Overrule setting that wasn't a multiple of 320 failed.
New build: Suggested still does that, and with values < 320 it actually just blocks (no input at all). BlockSize works on any setting.

BlockSize depends on one other value (I always have to specify some "suggested" value), if I set that value incorrectly, Wasapi now works fine, but Kernel Streaming does not. I may need to make both configurable... :(


Top
   
PostPosted: Mon Jun 21, 2021 4:57 pm 
User avatar

Joined: Sun Dec 23, 2018 7:44 pm
Posts: 209
Location: Texas, USA
I forgot to post the first time this happened. This is number two. Win 10 64 21H1


Attachments:
st-error-20210621.png
st-error-20210621.png [ 3.02 KiB | Viewed 626 times ]
Top
   
PostPosted: Wed Jun 23, 2021 7:36 pm 

Joined: Wed Nov 08, 2017 3:16 pm
Posts: 118
you are up to beta 17 on 6/23, just wondering what changes have been made, thanks.


Top
   
PostPosted: Thu Jun 24, 2021 1:46 pm 
User avatar

Joined: Tue Mar 17, 2009 2:56 pm
Posts: 3762
I decided to again test "Current AGC" since its 'new' again.
What i see so far it's still a mess. Dont even know where to start but here is 3 things.

- Wide band agc (green) is wrong reading, atleas acording to display. So if it's used it will wrongly drive AGC.
- Window is a must in this agc, so, since there is finaly window speed settings, i tried to put very slow release , and very fast attack speeds inside window. However attack still has hold like' effect with loud burst. So it completely ignores my VERY fast window attack speed, very bad. It seems same happens in release, but it less hurts.. still bad.
- Back to 1. I must use wideband settings, since there is no classic limit between bands. B1 - don't go below B2 by X dB, or B2 don't go below B1 by Z dB. So i MUST use wideband settings to make not-big-diff between bands, wich is ridiculous. Then when you play tracks with mostly bass, it will NOT rise non-bass-bands by zilion dB, but i have wrong WB readings so everything i driven wrong anyways.

_________________
control point


Top
   
PostPosted: Thu Jun 24, 2021 5:24 pm 

Joined: Wed Nov 08, 2017 3:16 pm
Posts: 118
Quote:
I decided to again test "Current AGC" since its 'new' again.
What i see so far it's still a mess. Dont even know where to start but here is 3 things.

- Wide band agc (green) is wrong reading, atleas acording to display. So if it's used it will wrongly drive AGC.
- Window is a must in this agc, so, since there is finaly window speed settings, i tried to put very slow release , and very fast attack speeds inside window. However attack still has hold like' effect with loud burst. So it completely ignores my VERY fast window attack speed, very bad. It seems same happens in release, but it less hurts.. still bad.
- Back to 1. I must use wideband settings, since there is no classic limit between bands. B1 - don't go below B2 by X dB, or B2 don't go below B1 by Z dB. So i MUST use wideband settings to make not-big-diff between bands, wich is ridiculous. Then when you play tracks with mostly bass, it will NOT rise non-bass-bands by zilion dB, but i have wrong WB readings so everything i driven wrong anyways.
honestly, i prefer the legacy AGC, its much more accurate in keeping the gain level of sound content where it should be without perceptively changing the properties and qualities of the sound content, which shouldnt be done with AGC.
AGC is just that, Automatic Gain Control, but to each their own.


Top
   
Display posts from previous:  Sort by  
Post new topic  Reply to topic  [ 89 posts ]  Go to page Previous 1 2 3 4 59 Next

All times are UTC+02:00


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Limited