Tír Na HiFi

2channelaudio wrote:Yes I tried 2.71 and 2.59 on my i5 laptop.
Very impressive stuff, love the SQ

I am particularly interested in JLP, due to its convenience and flac playback! as most of my music is in flac...

On a side note, I can't get MQN to work on my audio only PC, which is a core duo 1.86 running windows 8.1.....
JLP works fine...

I think I need to keep trying different versions, until I find one that works.... via my usb to spdif (yellowtec puc2 lite)
Cheers
2CA

try this
https://www.dropbox.com/sh/t4pp3igynpq5xno/Btq1pffSKF

this will play flacs etc as long as its 16/44
just copy track and double click " taggarts " mqnload

should work

2channelaudio wrote:JLP Team (SBGK),


But I thought it would be helpful to have a quick start guide for novice users wishing to early adopt JLP wdm-ks player.
This will prevent any basic questions repeating through the thread.

Personally I struggled a little getting my head around all the posts and instructions to get started....
A how to guide like this would have been a big help.

If appropriate I am happy to further develop/customise these instructions, as I wrap my head around the advanced tweaks/settings. I hope you don't mind me posting it here.

http://2channelaudio.blogspot.com.au/20 ... phile.html

FYI, the post does not cover advanced customisations (local player additional settings/affinity/customised bat files) but it gets enthusiastic & early JLP adopters listening.. Thanks again for a great sounding open source player. Seriously impressive..

2CA

Brilliant piece of work. I am sure many newcomers and may be some old hands will appreciate this.

have you got jlp at mqn quality?
still find bass a little loose muddy

nige - I share your view I think. The JLP bass has always been the weak point on my set-up; I have a Zu Method sub with two very fast 15" woofers and it can produce lovely taut and deep bass. I am now getting some bass depth, but not the tautness that I got with MQn.

Otherwise I prefer JLP which is working beautifully for me.

Jonathan

Have not done a-to-b testing, but tried to improve JLP on itself with the different parameters available.
Have got a lot of detail now, both of instrument and space. Also have got depth of bass, judged on itself, I believe with control and detail.
Need to do head to head testing to see if the control and detail of the bass is up there with MQn on my setup.

Still doing all this with my optimal settings of
-a 2ms
-b 260:8000
-c pcm (because I only play from wav files)
- Timerresolution 2ms
- Clockrate 23220
- + all other MultiMedia/Pro Audio optimisations
- Buffersize DDC-driver 2 ms
- squeezelite-win on own core + realtime priority
- squeezesvr on own core + realtime priority

Cheers

Aleg

Aleg this is very helpful.

I was interested to see that you appear to be using MMCSS? Gordon suggested that with KS MMCSS was not needed? But perhaps the Proaudio tweaks still apply?

I also have a driver set to 2 buffers/2ms latency.
Please could you clarify where in the Registry you are setting Timerresolution to 2ms?

Thanks
Jonathan

jrling wrote:

Have not done a-to-b testing, but tried to improve JLP on itself with the different parameters available.
Have got a lot of detail now, both of instrument and space. Also have got depth of bass, judged on itself, I believe with control and detail.
Need to do head to head testing to see if the control and detail of the bass is up there with MQn on my setup.

Still doing all this with my optimal settings of
-a 2ms
-b 260:8000
-c pcm (because I only play from wav files)
- Timerresolution 2ms
- Clockrate 23220
- + all other MultiMedia/Pro Audio optimisations
- Buffersize DDC-driver 2 ms
- squeezelite-win on own core + realtime priority
- squeezesvr on own core + realtime priority

Cheers

Aleg

Aleg this is very helpful.

I was interested to see that you appear to be using MMCSS? Gordon suggested that with KS MMCSS was not needed? But perhaps the Proaudio tweaks still apply?

I also have a driver set to 2 buffers/2ms latency.
Please could you clarify where in the Registry you are setting Timerresolution to 2ms?

Thanks
Jonathan

Hi Jonathan

Using the MCSS is not the right word as i believe MMCSS isn't present in a Core installation of 2012R2.
It is I just left those values from the MQn era.

TimerResolution cannot be set from registry. It is a realtime request made by a program to the Operating System, and the OS will set it to the smallest value requested by any of the programs that requested a specific Timerresolution.

I have made a small commandline executable, based on a source code I found on a Windows forum, that can send a request for the TimerResolution with any value you like and it will keep it at that for as long as the command window is open. Can't be used in Core mode however because there is no GUI support in Core and this program needs to be able to create a window and taskbar item.

Besides that one I made another version that contains a fixed value for TimerRequest and that can be jnstalled as a service and is started at boot, so I don't have to remember to start the command line version and it will also keep its request alive, where the commandline version needs to keep its window open. Also, being a service, it can be used in Core mode.

I have not put it into the open because it requires some compiling and manual service installation and not everybody is comfortable with that (though I did write a short manual for how to do it).

Cheers

Aleg

JLP squeezelite sets the system timer to 2ms if mmcss is not running

nige2000 wrote: have you got jlp at mqn quality?
still find bass a little loose muddy

Hi Nigel

I just did some a-b testing with another recording which doesn't show its textures as easily as did my other test recordings.

I have to agree JLP is not yet at the level of MQn. But it is very, very good already and I need to do a-b testing to remind me of how extra good MQn really is. I could be happy with JLP if I hadn't known MQn.

I wouldn't say the bass is loose and muddy. I wouldn't go that far. :-)

But it doesn't show yet the amount of textures in deep bass that I get with MQn. I don't think JLP is uncontroled with its bass, it is well defined at the edges and not woozy, but it lacks the textures and insight of the inside of the bass notes.

What I also notice immediately, is that MQn sound much more at ease, more relaxed in its presentation of all of the sound. I think this still has to do with more noise in the JLP playback environment or still some jitter effect.

Cheers
Aleg

sbgk wrote:JLP squeezelite sets the system timer to 2ms if mmcss is not running

Thanks Gordon and Aleg.

I run WS2012R2 in GUI mode (I am a wimp I know). So don't have the issues you describe with Core.

Seems like I have the best settings (by luck!) as MMCSS is not needed and System Timer is being set at 2ms anyhow.

Last night I tried -a 4 and -b 280:8000 and it really changed to the sweetspot for my system. Thesycon buffer latency is set at 2ms 2 buffers, so I wonder if it likes -a 4 being 2x2ms? That's probably rubbish.

As Gordon said, the pros and cons of Squeezelite are that it does seem particularly responsive to settings, much more so than MQn, I found. That may be more to do with KS v. WASAPI of course.

I feel though that I am just guessing at settings with no empirical basis whatsoever. Do you have any clues as to how the settings correlate to driver buffer, timerresolution, 16/44.1 v hi-res etc etc.?

I was little hesitant this time to see what everyone else has to say first, and I've got some courage now. I played versions 18/20 for some time and I switched back to MQn. That was the first time that I wasn't feeling significant difference, if there was any at all. I still need to find some time to optimize my system for it, but I could stick with it from now on.

uploaded squeezelite-win.exe v21 - more optimisations, track update on browser disabled.

I prefer the sound of this version, takes a bit of harshness off and seems to add calmness and depth. Still a lot going on in squeezelite that wouldn't be present if just loading and playing the files from ram like MQn, but one step at a time.

pause wasn't working when running in debug, but seems to work now.

and if anyone saw the recent thread in CA where an expert on jitter was saying that the lack of jitter could be judged by toe tapping, well this version is getting my toes tapping.

http://www.computeraudiophile.com/f8-ge ... dio-19231/

The only difference between v20 and v21 is that v21 has removed the progress bar update code from the portaudio callback routine, PA does warn about putting non essential code in the callback, more from the performance perspective than how it sounds. Whether there is a better place to put it I don't know, but interesting to hear what the effect of removing it is.