MQN

[jesuscheung]

prefer 3.37 over 3.38.
i cannot remember the last time a software produces zero intensity.

3.37 is unique.

[jorm]

3.37 is beautiful indeed,but the 3.37 mqnplay2496.exe still produces no sound through Audiophilleo without driver in R2.Thesycon 2.20 is working with 3.37 24/96 ,but there is a constant output of soft clicks which are not present through JPLAY,both WASAPI and KS. Both 3.37 versions perform flawlessly with the Young ,Amanero and Young have identical performance vis-à-vis MQn.So still definitely a need for some fine tuning of the 24 bit hirez versions.

Cheers

Jorgen

[nige2000]

didnt get testing 3.38

did 2.97
and 3.37
and 3.27 avx

think you've lost a little bass control and definition since 2.97 (need to get out the bass test tracks for it)
most else seems rosey enough

[Aleg]

JC, there's an interesting debate about mmcss settings and whether they change only at mmcss service start, I think they do, but I'm sure we've all thought there was a sound difference after changing them via regedit without restart.

the solution seems to be to have a proaudio setting for each bitrate sample rate combo eg pa1644 etc

yes. changing registry in mmcss is effective immediately. realtime settings

from time to time i retest each values in mmcss. i cannot do better than pkshan's mmcss registry.
funny he figured this out years ago before win8 existed. total respect.

same mmcss applies from win7 to win8.1. nothing really has changed in windows. apart from graphical rendering.

if you believe me, i think clockrate needs to be discovered depending what DAC one got.
in reality, for pa1644,there is only 2 clockrates that perfectly syn the timerresolution
640=14.5ms. and 352=8ms
448=10ms but the rounding error is big, you can detect little lost of res.

default clockrate doesn't sound that bad.
incorrect clockrate will sound really wrong.

If you use SysInternals ProcesMonitor you can see that Clock Rate is not read again after it has been set by mqncontrol!
Would one not expect to see the Clock Rate being read from registry just like it is when mmcss is started?
There is so little documented about how this actually is done by Windows.

When mmcss is started you can see that svchost is reading clock rate from registry.
When mqncontrol is run you can see clock rate is being changed in registry, but it doesn't seem to be read again.

-

Aleg

[Clive101]

3.37 mqnplay2496.exe is absolutely stunning - beautiful...thank you!

Gordon - as this is so good - are approaching a "final" version and does this mean that you plan to get other sample rates working, like 44.1 16 bit WAVs to work for dacs that only accept 24 bits and those of us with other peculiar dacs or will you perhaps do a Kernel Streaming next...

Clive

[darkpink]

I like the presentation of 3.37 think it's better than 3.38, it got a little more bite to it.

[2channelaudio]

Gents

I'm running a basic win8.1 setup.. Using a e7500 cpu

No pro audio mods or tweaks to windows... As I use this pc for XBMC and music

I'm very much liking 3.37 on my new setup
Plenty of detail, great bass and very transparent midrange

Like it, two thumbs up
Well done impressive work

[jesuscheung]

If you use SysInternals ProcesMonitor you can see that Clock Rate is not read again after it has been set by mqncontrol!
Would one not expect to see the Clock Rate being read from registry just like it is when mmcss is started?
There is so little documented about how this actually is done by Windows.

When mmcss is started you can see that svchost is reading clock rate from registry.
When mqncontrol is run you can see clock rate is being changed in registry, but it doesn't seem to be read again.

-

Aleg

you got me interested. you can detect change in sound during playback.

i guess, svchost has a pointer to the memory to the registry.

[sbgk]

3.39 has a different way of doing things, quite free sounding.

[nige2000]

2.39 > 2.37

Must be due an avx soon
Bad when a sse2 is better