Page 752 of 804

Re: MQN

Posted: Mon Sep 28, 2015 8:42 pm
by sebna
Is 117 windows 10 supposed to play on W12? I am asking as I do not see in main / archive 117 dedicated to W12.

Cheers

Re: MQN

Posted: Tue Sep 29, 2015 5:28 am
by Aleg
sebna wrote:Is 117 windows 10 supposed to play on W12? I am asking as I do not see in main / archive 117 dedicated to W12.

Cheers
117 win12 is just there in main alongside the 117 win10

Re: MQN

Posted: Wed Sep 30, 2015 11:05 pm
by sbgk
uploaded play v121, control v32 and a new exe called mqncounter.exe, v1, v2 and v3.

think v2 is current favourite, of course the counter exe affects the sq.

play v121 requires mqncounter to work and mqncounter requires control v32

hopefully it works for some.

if you want to stop it early stop mqnplay first, otherwise you'll get a load of noise as whatever was in the buffer is repeated. It should all close down after all tracks played.

by having the mqncounter.exe have moved 4 additions and a memory write out of the render loop and was able to use another 3 registers. Seems to make a difference. This is about as minimal a render loop as you can get.

Naturally, I think this is the best sounding version as well.

Re: MQN

Posted: Wed Sep 30, 2015 11:41 pm
by elaprince
Gordon one thing I don't like it and I think isn't good for SQ is that Mqncounter use so much CPU power
It use 25%
Is there anyway to fix that?
Also I am getting sometime weird buzz from speakers when loading few track
Other then that sound is very good

Re: MQN

Posted: Wed Sep 30, 2015 11:47 pm
by sbgk
elaprince wrote:Gordon one thing I don't like it and I think isn't good for SQ is that Mqncounter use so much CPU power
It use 25%
Is there anyway to fix that?
Also I am getting sometime weird buzz from speakers when loading few track
Other then that sound is very good
can't think of a way to reduce it at the moment, buzz is maybe contention for resources, try counter on a different core ? runs ok on my i5 haswell laptop, maybe depends what else you're doing.

counter v4 takes it down to 3%, with some benefit to sq as well.

v5 is 0.1% cpu, but don't know if it's too close to the buffer refresh rate so might get some crackle, sounds sublime though.

Re: MQN

Posted: Thu Oct 01, 2015 1:06 am
by elaprince
Shall try it tomorrow

Re: MQN

Posted: Thu Oct 01, 2015 6:17 pm
by Aleg
sbgk wrote: v5 is 0.1% cpu, but don't know if it's too close to the buffer refresh rate so might get some crackle, sounds sublime though.
No issues with v5 here, and indeed it sounds sublime
All 3 mqn processes show as 0% on my i7

Re: MQN

Posted: Thu Oct 01, 2015 7:56 pm
by sbgk
wasn't that convinced by play v121, so took some alignment stuff out and think 122 sounds better.

uploaded mqnplay v123, small tweak to v122

uploaded mqncounter v6, code optimised in assembly

v7 better instructions

v8 = v7 with correct alignment, surprised at how much mqncounter affects the sq.

Re: MQN

Posted: Fri Oct 02, 2015 2:51 pm
by janh
I have mixed result with latest versions:
play v121, control v32, counter v2, loader v29 give fine sound.

Then I have tried some newer versions, and I only get distorted sound:
play121, control32, counter v4
play123, control32, counter v8 as well as v11
all these only give distorted sound

Re: MQN

Posted: Fri Oct 02, 2015 4:43 pm
by Aleg
All combinations with play123 don't give any sound.
Using same combinations with play122 all play just fine