Visual studio 2012 c++ and wasapi minimalist player

[tonerei]

Eventually converted to this fantastic player when I got to hear it recently. It beats Jplay streaming on core server which shocked me. How close are you to playing hires files? Well done with all the work you have come up with something really special

[nige2000]

right

had a bit of a play around with the rpac with the newer xmos firmware and the newer drivers, playing hires and everything else no problem. i gave it a clean power source so it didnt sound terrible at all

difficult for me to review as i changed my dac for a worse one but theres definitely more music there in the

 

are we going to have to source new gear with the xmos chipset and thesycon drivers for hires compatability?

[sbgk]

don't get too excited, the left and right are reversed, I don't know why, yet.

 

Uploaded MQnControl.exe 2 core affinity v2, 4 core affinity v2 and no affinity v2 - fix to hirez swapping channels. remove control line, added file count and file info lines

 

recommend the MMCSS pro audio setting as mentioned a few posts back, am thinking of adding it to mqncontrol so MMCSS always uses the optimum setting depending on the sample rate

[sbgk]

Eventually converted to this fantastic player when I got to hear it recently. It beats Jplay streaming on core server which shocked me. How close are you to playing hires files? Well done with all the work you have come up with something really special

 

thanks, never heard JPlay streaming, but am familiar with beach and river and asked them a number of times why they couldn't produce a neutral engine, maybe they will now.

 

What issue do you have with HiRez ?

[sbgk]

let's cost MQn, couple hours a night for 6 months = 3 x 6 x 20 = 360 hours * £40 = £14440 + testers, sell 50 copies = £288 a copy, any takers ? I thought it would work out more expensive, just as well it's a hobby.

 

Why are there so many free collaborative projects in the linux world, but in the windows world everyone wants to make money.

[tonerei]

thanks, never heard JPlay streaming, but am familiar with beach and river and asked them a number of times why they couldn't produce a neutral engine, maybe they will now.

 

What issue do you have with HiRez ?

 

My understanding is that currently the player will not play hires files in their original format (downgrading them to 16bit) Is that not correct? Maybe I need to go back through all the threads?

 

Yes if I was Josef or Marcin I would be listening to this player very carefully and looking to trying to improve Jplay.

[sbgk]

right

had a bit of a play around with the rpac with the newer xmos firmware and the newer drivers, playing hires and everything else no problem. i gave it a clean power source so it didnt sound terrible at all

difficult for me to review as i changed my dac for a worse one but theres definitely more music there in the

 

are we going to have to source new gear with the xmos chipset and thesycon drivers for hires compatability?

 

don't know, has it a history of problems connecting to players ? Maybe I can get a loan of one to try.

[sbgk]

My understanding is that currently the player will not play hires files in their original format (downgrading them to 16bit) Is that not correct? Maybe I need to go back through all the threads?

 

Yes if I was Josef or Marcin I would be listening to this player very carefully and looking to trying to improve Jplay.

 

people with 24 bit dacs can't play 24 bit files, my dac uses 24 bit in 32 bit container so that works, hope to have that sorted by the weekend. The Thesycon issue is new to me.

 

Any self respecting developer will have had a look at the assembly code, I would have thought. The question is how much is the SQ dependent on it being a bare bones player, not much point in making an exact copy.

[sbgk]

Uploaded MQnControl.exe 2 core affinity v2, 4 core affinity v2 and no affinity v2 - fix to hirez swapping channels. remove control line, added file count and file info lines

 

recommend the MMCSS pro audio setting as mentioned a few posts back, am thinking of adding it to mqncontrol so MMCSS always uses the optimum setting depending on the sample rate

 

Uploaded mqnplay.exe 2.53 sse4 intel eax new loop, thought of an ingenious way of preloading the cache lines. SQ is fuller than the previous 2.53s, my new favourite.

 

 

Just remembered the render loop in MQnPlay for 24/96000 has been left in a dev mode so shall fix it later.

[sbgk]

My understanding is that currently the player will not play hires files in their original format (downgrading them to 16bit) Is that not correct? Maybe I need to go back through all the threads?

 

Yes if I was Josef or Marcin I would be listening to this player very carefully and looking to trying to improve Jplay.

 

there is no format manipulation, so you got that bit wrong, although it will make 16/44.1 sound like it's from a higher resolution, but that's not by manipulation it's more by subtraction of noise.

[jesuscheung]

Uploaded mqnplay.exe 2.53 sse4 intel eax new loop, thought of an ingenious way of preloading the cache lines. SQ is fuller than the previous 2.53s, my new favourite.

 

 

Just remembered the render loop in MQnPlay for 24/96000 has been left in a dev mode so shall fix it later.

 

2.53 sse4 intel eax new loop

-need fuller/longer impacts, and less bright/volume on impacts. this applies in general, e.g. drums, vocal.

-stage got narrower than 2.53 eax and rax no zi. later should be more correct.

-otherwise, it should be great.

-heavily background oriented music sounds better than ever.

[lekt]

...recommend the MMCSS pro audio setting as mentioned a few posts back, am thinking of adding it to mqncontrol so MMCSS always uses the optimum setting depending on the sample rate

 

i think optimally that MQn have recommend about it, let user make it in windows registry by his way. because on depending of windows tweaks, sometimes sound is not open/free when pro audio was changed other than default. and PriorityControl, core affinity too. JPlay have exclusive core affinity, then on my laptop i feel its sound is not open/free.

[cvrle59]

Uploaded MQnControl.exe 2 core affinity v2, 4 core affinity v2 and no affinity v2 - fix to hirez swapping channels. remove control line, added file count and file info lines

 

recommend the MMCSS pro audio setting as mentioned a few posts back, am thinking of adding it to mqncontrol so MMCSS always uses the optimum setting depending on the sample rate

Can anybody brake this MMCSS settings down, please? I'm not sure that I understand how to do it.

Thanks!

[lekt]

2.53 sse4 intel eax new loop

-need fuller/longer impacts, and less bright/volume on impacts. this applies in general, e.g. drums, vocal.

-stage got narrower than 2.53 eax and rax no zi. later should be more correct.

-otherwise, it should be great.

-heavily background oriented music sounds better than ever.

 

agree with you.

vibration still OK. sound more open/free, clearly than rax no zi. but bass volume 99%, mid/treble volume 101%, brightness 102%, soundstage 98%. that sound depth is not enough, a bit more flat and heavy.

[sbgk]

2.53 sse4 intel eax new loop

-need fuller/longer impacts, and less bright/volume on impacts. this applies in general, e.g. drums, vocal.

-stage got narrower than 2.53 eax and rax no zi. later should be more correct.

-otherwise, it should be great.

-heavily background oriented music sounds better than ever.

 

MQnPlay.exe 2.53 sse4 intel eax new loop mov r9 - mov instead of sub for second loop and hirez render loop fix as 2.53 sse4 intel eax new loop used memcpy for hirez

Used a sub instruction which is always problematic with too much stridency, so changed it for a mov and removed it from the loops, this version should be pretty damned good now.

 

Also can use the 2.53 sse4 intel eax new loop version to compare the original windows memcpy in an unoptimised loop for hirez with 2.53 sse4 intel eax new loop mov r9 which has the optimised loop and customised assembly coded memcpy

 

 

 

[jesuscheung]

shall upgrade it and have an ssd which I'll put ws2012 r2 on and then easy to swap from dev to player

 

just been trying the pro audio mmcss timer switch, think it should be 2048/44100 and that works out as 46440 (previously had it at 23220). 46440 totally transforms the sound of 2.53 eac c name, no problem with bass, makes the 2.53 rax cp sound a bit horrible.

 

have you tried 92880? 23220, 46440, 92880 all work. i never knew which one is best. i know that once you use 6 digits clock rate, the SQ starts to get less free

[sbgk]

Can anybody brake this MMCSS settings down, please? I'm not sure that I understand how to do it.

Thanks!

 

change the following key using regedit, take a backup of the registry if you don't feel confident. This is at your own risk.

 

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\Tasks\Pro Audio

 

change clock rate to 2048/44100 = 46440 which is b568 in hex (default is 10000 which is 2710 hex)

 

 

Here are my settings, you could copy them into a .reg file and double click on it to update the registry, again take a back up first. Guess the affinity setting could be set to core 1 ie set it to 00000002

 

Windows Registry Editor Version 5.00

 

 

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\Tasks\Pro Audio]

"Scheduling Category"="High"

"GPU Priority"=dword:00000008

"Affinity"=dword:00000000

"Clock Rate"=dword:0000b568

"SFIO Priority"="Normal"

"Priority"=dword:00000008

"Background Only"="False"

[sbgk]

have you tried 92880? 23220, 46440, 92880 all work. i never knew which one is best. i know that once you use 6 digits clock rate, the SQ starts to get less free

 

Only tried 46440 as that is the result of the calculation 2048/44100, bass gets more lumpy as it goes up and thin if too small.

[sbgk]

i think optimally that MQn have recommend about it, let user make it in windows registry by his way. because on depending of windows tweaks, sometimes sound is not open/free when pro audio was changed other than default. and PriorityControl, core affinity too. JPlay have exclusive core affinity, then on my laptop i feel its sound is not open/free.

 

wouldn't you want to hear MQn the way it was designed to sound ? I think if there was a consensus that it sounded better then I would implement it, maybe have a switch to disable it, but the SQ difference is worth it.

[jesuscheung]

Only tried 46440 as that is the result of the calculation 2048/44100, bass gets more lumpy as it goes up and thin if too small.

 

i use 92880 when my ears are fatigued. coz audio will be cleaner. not sure it is less noise or lost of details. 46440 is safer choice.

[goon-heaven]

2.53 sse4 intel add eax new loop mov r9: unearths yet more detail - very clean sound. Excellent. Thank you sbgk for your gift.

[sbgk]

2.53 sse4 intel add eax new loop mov r9: unearths yet more detail - very clean sound. Excellent. Thank you sbgk for your gift.

 

sound was a bit closed in, uploaded 2.53 sse4 intel full loop which brings it back to life

[LowOrbit]

there was an instruction on how to use wasapi_test a couple of pages back, what results do you get ?

 

Hiya

 

I did try this but the application launched then closed immediately. I will try it again when I have five minutes to spare. It would be nice to find a resolution to this issue, it does seem to be related to the way the Thesycon driver expects data to be packaged. As I said elsewhere it seems to work with other players without issue, but they aren't as lean as MQn (and don't match it for sonic purity either).

 

Mark

[LowOrbit]

Unfortunately mine is too early a firmware version to allow user firmware upgrade.

Waiting (patiently) for Lorien to produce his daughter board and then he will flash the firmware with the latest version.

 

I do have the later version which does allow for firmware updates via the Thesycon control window. I updated to the 384k/32 bit driver without any problem. Lucian's daughter board is highly anticipated, but given the improvements that applications like MQn bring in delivering a cleaner data stream I am not sure how much we have to gain from what the new board will deliver.

 

Mark

[goon-heaven]

Hmmm... Sorry, I still prefer 2.53 sse4 intel add eax new loop mov r9 - no shortage of life here

Listening through headphones (whilst working).

 

2.53 sse4 intel full loop sounded closed in

 

Holophoney returns.