MQN

[sbgk]

uploaded loader 42 and control 63, hadn't bothered to align them before, but seems to work.

play 183 is same as 174 as far as I can tell, haven't tried to hear if they're different

183 with 42 and 63 is as good as I've heard it, musical to me, hopefully for others as well.

[internethandle]

Can sbgk or others help me confim/parse whether a skylake (i7-6700k) on a H170 mobo would run some of the newer versions of MQn? Sorry if this is a repeated question.

[maroni]

uploaded loader 42 and control 63, hadn't bothered to align them before, but seems to work.

play 183 is same as 174 as far as I can tell, haven't tried to hear if they're different

183 with 42 and 63 is as good as I've heard it, musical to me, hopefully for others as well.

Gordon
play 180 totally agreed, excellent again and not only for guitar and piano :-), like it very much, one of the reference versions to me.
play 183, heard only with control 62 and loader 41, yet. More impressive, more sound colors - could be more popular, depending on equipment and personal taste :-)

Early, but anyway: "Merry Christmas and a happy New Year!"...

[sbgk]

uploaded loader 42 and control 63, hadn't bothered to align them before, but seems to work.

play 183 is same as 174 as far as I can tell, haven't tried to hear if they're different

183 with 42 and 63 is as good as I've heard it, musical to me, hopefully for others as well.

Gordon
play 180 totally agreed, excellent again and not only for guitar and piano :-), like it very much, one of the reference versions to me.
play 183, heard only with control 62 and loader 41, yet. More impressive, more sound colors - could be more popular, depending on equipment and personal taste :-)

Early, but anyway: "Merry Christmas and a happy New Year!"...

180 becomes too much after a while, so maybe 183 is the best solution. They differ in the instruction used to increment the data buffer.

Seasons greetings to you as well.

uploaded control v64, corrected a setting.

[Aleg]

uploaded loader 42 and control 63, hadn't bothered to align them before, but seems to work.

play 183 is same as 174 as far as I can tell, haven't tried to hear if they're different

183 with 42 and 63 is as good as I've heard it, musical to me, hopefully for others as well.

Gordon
play 180 totally agreed, excellent again and not only for guitar and piano :-), like it very much, one of the reference versions to me.
play 183, heard only with control 62 and loader 41, yet. More impressive, more sound colors - could be more popular, depending on equipment and personal taste :-)

Early, but anyway: "Merry Christmas and a happy New Year!"...

180 becomes too much after a while, so maybe 183 is the best solution. They differ in the instruction used to increment the data buffer.

Seasons greetings to you as well.

uploaded control v64, corrected a setting.

+1 in the 183/42/64 combo

[sbgk]

maybe had the alignment wrong, 185, 65, 43 have better presentation.

[John Dot]

uploaded play v149 and control v53, doesn't use counter any more. Does it work for different dacs ?

should be 5 x less noise from interrupts 100 times a second to 5 times a second, works for me, with more depth to the detail.

if it works for different dacs I'll tune it a bit and get the non haswell version out.

Merry Christmas to you all audio freaks.

jD

[sbgk]

uploaded play v149 and control v53, doesn't use counter any more. Does it work for different dacs ?

should be 5 x less noise from interrupts 100 times a second to 5 times a second, works for me, with more depth to the detail.

if it works for different dacs I'll tune it a bit and get the non haswell version out.

Merry Christmas to you all audio freaks.

jD

what do you think of the non haswell version ?

[John Dot]

Wow, I've found the files right now. I'll report later.

[sbgk]

uploaded play v186 and loader v44

play has some further tweaks, loader had an alignment error.

think there's a bit more insight into the music.

piano in particular sounds more realistic, other things improved as well.

If you're listening to a live track you can just about hear what they're ordering at the bar. A surprising lift in detail without bass being affected, which is usually the case.

saw an ms article that said the default device buffer size was 10ms and win 10 allows the driver to select different sizes (the driver would have to be rewritten for this). Interesting as not many players allow the selection of 441 samples which is 10ms for 16/44.1. Somehow we identified 10ms as sounding best for mqn, which makes sense as the internal and device buffer are then aligned and no compensation code needs to run.

http://video.ch9.ms/sessions/winhec/201 ... %2010.pptx