Page 272 of 299

Re: lekt player

Posted: Fri Apr 01, 2016 5:03 am
by jesuscheung
kse is superb. Better in so many ways. Bass quantity lesser, treble better.piano notes vibrate more, and more correctly too

Re: lekt player

Posted: Fri Apr 01, 2016 7:31 am
by lekt
jesuscheung wrote:kse is superb. Better in so many ways. Bass quantity lesser, treble better.piano notes vibrate more, and more correctly too
i think you run it for laptop onboard sound card?
it doesn't work for goon-heaven's PCI card, strange, all kse versions never work for PCI card. code need go to other branch that i can't follow up coz haven't this card, use only laptop. will try more...

Re: lekt player

Posted: Fri Apr 01, 2016 8:06 am
by goon-heaven
What can I do to help get kse playing on PCI soundcard?

Re: lekt player

Posted: Fri Apr 01, 2016 10:04 am
by lekt
do nothing. will build versions that test some things i've suspected in your OS with PCI card.
bypassing all kse layers is good but now have much code, not easy make alignment and register usage, sound easy comes to hard and imbalance, need more accuracy.
in win10 code show that audiokse.dll contains the last layer that user application can do something with audio core, after that audio render run in kernel/driver mode in different processes and threads, asynchronous with user side and need high programing skill to make tricks.

Re: lekt player

Posted: Fri Apr 01, 2016 1:05 pm
by sbgk
lekt wrote:do nothing. will build versions that test some things i've suspected in your OS with PCI card.
bypassing all kse layers is good but now have much code, not easy make alignment and register usage, sound easy comes to hard and imbalance, need more accuracy.
in win10 code show that audiokse.dll contains the last layer that user application can do something with audio core, after that audio render run in kernel/driver mode in different processes and threads, asynchronous with user side and need high programing skill to make tricks.
think you would have been quicker to look at the wave-rt code in portaudio, wouldn't have had to muck about with the wasapi stuff.

Re: lekt player

Posted: Fri Apr 01, 2016 2:18 pm
by goon-heaven
sbgk wrote:
lekt wrote:do nothing. will build versions that test some things i've suspected in your OS with PCI card.
bypassing all kse layers is good but now have much code, not easy make alignment and register usage, sound easy comes to hard and imbalance, need more accuracy.
in win10 code show that audiokse.dll contains the last layer that user application can do something with audio core, after that audio render run in kernel/driver mode in different processes and threads, asynchronous with user side and need high programing skill to make tricks.
think you would have been quicker to look at the wave-rt code in portaudio, wouldn't have had to muck about with the wasapi stuff.
@Gordon: Does MQN use wave-rt code in portaudio?

Re: lekt player

Posted: Fri Apr 01, 2016 5:40 pm
by sbgk
goon-heaven wrote:
sbgk wrote:
lekt wrote:do nothing. will build versions that test some things i've suspected in your OS with PCI card.
bypassing all kse layers is good but now have much code, not easy make alignment and register usage, sound easy comes to hard and imbalance, need more accuracy.
in win10 code show that audiokse.dll contains the last layer that user application can do something with audio core, after that audio render run in kernel/driver mode in different processes and threads, asynchronous with user side and need high programing skill to make tricks.
think you would have been quicker to look at the wave-rt code in portaudio, wouldn't have had to muck about with the wasapi stuff.
@Gordon: Does MQN use wave-rt code in portaudio?
no, uses wavecyclic. PA is the only place I've seen a wavert player implementation which is required for PCIE direct access.

Re: lekt player

Posted: Fri Apr 01, 2016 6:23 pm
by goon-heaven
sbgk wrote:
goon-heaven wrote:
sbgk wrote:...
think you would have been quicker to look at the wave-rt code in portaudio, wouldn't have had to muck about with the wasapi stuff.
@Gordon: Does MQN use wave-rt code in portaudio?
no, uses wavecyclic. PA is the only place I've seen a wavert player implementation which is required for PCIE direct access.
Thank you for the insight on WaveRT. But does wavecyclic not support PCIE direct access?

Re: lekt player

Posted: Sun Apr 03, 2016 3:59 am
by jesuscheung
lekt wrote:
jesuscheung wrote:kse is superb. Better in so many ways. Bass quantity lesser, treble better.piano notes vibrate more, and more correctly too
i think you run it for laptop onboard sound card?

yes, onboard.

ses remains the current best of lekt! goon, you aren't missing much here with kse

Re: lekt player

Posted: Fri Apr 08, 2016 10:42 pm
by lekt
i was busy a long time, have less time for coding now. still difficult analys code of kse layer, many objects in protected mode, will do again with kse.

code of v3.55 256 ses is good, sound very nice. rewriten one function in audioses.dll layer and can remove one thing in render loop, i feel sound really better.
uploaded:
v3.55.2 256 , v3.55.3 256 nocrit , v3.55.4 256

try them and compare SQ, what is more correct..
===========================

and v3.55.5 256 ses2 (raw code) for goon-heaven testing with PCI card, there's code join into kse layer but not much only. hope it will work for you. pls. notice me about this.