Tír Na HiFi

nige2000 wrote:

sima66 wrote:I'm trying to keep up with you guys here and try to understand if the "Squeezelite Affinityset 4core.exe" should be placed in the same bin like the other excels? Should both be in the bin, in the same time?

yes in the bin file along with everything else

Thanks Nigel and also big thanks for your help with my R2 installation.

BTW, Nigel was trying remotely for a full week to help me install the R2 on the CF card. There are no words that I can find to thank him enough.

jrling wrote:

sbgk wrote:using -a 1 b: 1000:200000 -q with default amanero driver settings getting 0.6 % cpu system interrupts.

Anyone else getting high % Windows System Interrupts? We are running WS2012 R2.

No me neither.
I am running 2012R2 in Core mode, so don't have much running.
In taskmanager the System Interupts stay firmly on 00.

Cheers

Aleg

Just moved to v18 PA and v20 Squeezelite. I'm liking this very much, really clean, balanced and musically involving.

This is on phones in the study, via mixing desk (really clean monitoring phoneamp) and Maudio audiophile 2496 pci card, running lowest wdm buffer (64 samples).

-a 5 -b 400:8000. Squeezelite consumes 0.1% of cpu on i7, W8, no optimisations applied. Interrupts are running well under 1%.

Will drop this into the music system at the weekend.

someone just donated to my MQn unicef fund after using JLP, thanks for your generosity.

Think the future will be that 16/44 and above will play via MQn, probably using KS and streaming will use pa, both squeezelite and pa will be optmised.

After the effect that changing the buffer sizes had on the SQ it's apparent how much squeezelite streaming is affecting it, so best sq is still going to be via a standalone player.

just gave 20/18 a blast
really nice @ -a 0 -b 260:4000

want to play with clockrate is it changed in pro audio too?

also eventually moved to minimal server for mqn jlp comparing and it makes a nice improvement

have yet to move to core, will do a new install tomorrow for that

nige2000 wrote:just gave 20/18 a blast
really nice @ -a 0 -b 260:4000

want to play with clockrate is it changed in pro audio too?

also eventually moved to minimal server for mqn jlp comparing and it makes a nice improvement

have yet to move to core, will do a new install tomorrow for that

don't know what it does with 0.

have mmcss disabled, but yes clock rate would be in pro audio, it uses mmcss if you have it running else sets system clock to 2ms

sbgk wrote: don't know what it does with 0.

dunno either only sounds like 1 only more of it, i could use 1-2 without feeling too shortchanged

have mmcss disabled, but yes clock rate would be in pro audio, it uses mmcss if you have it running else sets system clock to 2ms

ok i had forgot about mmcss ill check that

we going to use squeezelite, portaudio without LMS then?

Think the future will be that 16/44 and above will play via MQn, probably using KS and streaming will use pa, both squeezelite and pa will be optmised.

After the effect that changing the buffer sizes had on the SQ it's apparent how much squeezelite streaming is affecting it, so best sq is still going to be via a standalone player.

Interesting. A few queries, if I may:

16/44.1 and above would seem to cover all WAV files? Or are you suggesting that SQlite streaming would only be used for say MP3?
How will you achieve KS in MQn? That has been a technical issue for you up until now. Are you thinking of rolling your own code for KS or incorporating PortAudio KS code?
I agree that the changes to Squeezelite buffer settings are having quite material effects on SQ. If it is the streaming code in Squeezelite responsible, and I thought that JLP was using 'Local Player' plug-in, therefore not 'streaming' in the LMS sense of that word, could you not cut out the streaming processes altogether and make JLP a dedicated Local Player (using PA KS of course)?

Thoroughly enjoying JLP.

Thanks

Jonathan

Thanks to those who posted their System Interrupt % load when playing JLP.

Although I am alone in using an Atom dual core CPU, which is of course much lower spec than current Haswell CPUs which all respondents seemed to be using, taggart's Haswell NUC still exhibited higher System Interrupt % than he expected.

The conclusion that I make, based upon inadequate data and a very small sample admittedly, is that taggart & I are using the WaveIO USB/SPDIF converter and more importantly, the Thesycon v2.15 Windows driver (their latest version) and others are not. So it is possible that Thesycon KS driver code is causing high System Interrupt % load. Whereas in MQn using Thesycon WASAPI driver code, it was not.

Not complaining though as JLP KS music is sublime. It is just that it could be even more so if System Interrupt % was lower.

jrling wrote:Thanks to those who posted their System Interrupt % load when playing JLP.

Although I am alone in using an Atom dual core CPU, which is of course much lower spec than current Haswell CPUs which all respondents seemed to be using, taggart's Haswell NUC still exhibited higher System Interrupt % than he expected.

The conclusion that I make, based upon inadequate data and a very small sample admittedly, is that taggart & I are using the WaveIO USB/SPDIF converter and more importantly, the Thesycon v2.15 Windows driver (their latest version) and others are not. So it is possible that Thesycon KS driver code is causing high System Interrupt % load. Whereas in MQn using Thesycon WASAPI driver code, it was not.

Not complaining though as JLP KS music is sublime. It is just that it could be even more so if System Interrupt % was lower.

ive had a quick play with thesycon drivers a while back (on others peoples dacs) minium settings cause strain especially buffer, but some other things improve like resolution
cant remember think it was min latency and the second lowest buffer i liked
but second lowest settings for both was good too (this was with mqn)

on my dac system interrupts was 0.02
i didn't look into lowering it or what effects it

i know its sort of useless info
but the small laptop type chips don't preform as well
anyway what does pentium, i5, i3 mean (one of my pet hates)
well not a lot you still have to do the homework
4 real cores would help with affinity but some say more cores more noise i dunno bout that (i get the cheaper ones incase i blow em' up)

http://www.cpubenchmark.net/cpu.php?cpu ... Hz&id=1944
http://www.cpubenchmark.net/cpu.php?cpu ... Hz&id=2020
http://www.cpubenchmark.net/cpu.php?cpu ... Hz&id=2015
http://www.cpubenchmark.net/cpu.php?cpu ... Hz&id=1924

all those figures become irrelevant if your running your cpu at 800mhz like i am then its guess work on which one to get
i dont think all one mini type pc solutions are good for audio