JLP wdm-ks player

[2channelaudio]

SBGK,

I thought I would also add, V24 fixed some audio glitches I was having with playback via my yellowtec puc2 lite usb to AES converter..... the issue would present itself as a 1-2 second audio mute/silence while playing tracks, usually once per track or thereabouts (squeezelite v20).....

In local player my additional settings are -a2 -b 700:2000000

I now have perfect playback... interesting...
Would there have been any versions changes that could have improved this?

Cheers
2CA

[LowOrbit]

Hi Gordon

Just installed version 24. Chopping back and forth between tracks, I can't replicate the issue I saw with non-16/44 tracks. All tested rates play happily. Listening to Phronesis "Walking Dark" at 16/88.2k on the headphones as I type this and it sounds bloody fine.

The progress bar works upto a count of 58 seconds then resets. Mercifully the track plays through!

With regard to my earlier comment re "Edgy" sound - I did not mean noisy or digital but defined and dynamic. The rather over-bright washy sounding treble that I associate with digital noise is not audible here. I know others are still stating their preference for MQN. I still don't hear it. I tried every flavour of MQN - JLP is hitting the sweet spot for me, and that assessment is based on listening on two different setups. Colour me deaf if you like!

[Aleg]

Aleg, you've expressed a preference for certain buffer size combinations. Unless you've chosen the minimum buffers that work then you've just tuned the noise to the sound you like from your system, I find the smaller buffer sizes to be digital sounding which is consistent with anything that introduces higher frequencies into the playback chain whether it's system time, cpu frequency etc.

I've seen people describe the latest jplay 5.2 version as more digital sounding than the previous one, I would say that's a bad thing. I don't hear digital noise when I listen to a live band/orchestra, so I'm aiming for transparency.

Gordon

I think it requires the capability to handle high frequency signals to be able to reproduce details.
It is just the sources of high frequency signals that make it sound right or wrong.
High frequency noise is bad and can be heared in a revealing system. It has a certain effect on sound.
High frequency music details is good and need a revealing system to be heared.

Limiting the capability of revealing high frequency signals in order to cover up high frequency noise is not the right way to go, because you will also cover up high frequency music signals that we do want to hear.

The goal imho should be to make a playback chain as revealing as possible, also for high frequency noise if it be, and work at removing the source of that high frequency noise. The latter by improving power supply, grounding, software optimisation to reduce CPU-born noise etc.

MQn is at the moment more transparent and more revealing of high frequency music details than JLP. It comes through in the showing of micro-variations in tones where JLP shows tones in a more opaque / less transparent way.

My 2 cents

Cheers

Aleg

[nige2000]

I think it requires the capability to handle high frequency signals to be able to reproduce details.
It is just the sources of high frequency signals that make it sound right or wrong.
High frequency noise is bad and can be heard in a revealing system. It has a certain effect on sound.
High frequency music details is good and need a revealing system to be heard.

Limiting the capability of revealing high frequency signals in order to cover up high frequency noise is not the right way to go, because you will also cover up high frequency music signals that we do want to hear.

The goal imho should be to make a playback chain as revealing as possible, also for high frequency noise if it be, and work at removing the source of that high frequency noise. The latter by improving power supply, grounding, software optimisation to reduce CPU-born noise etc.

MQn is at the moment more transparent and more revealing of high frequency music details than JLP. It comes through in the showing of micro-variations in tones where JLP shows tones in a more opaque / less transparent way.

My 2 cents

Cheers

Aleg

completely agree

[sbgk]

I've got nothing more to say on the matter, I'm not going to repeat myself if people don't want to read and understand.

[taggart]

Tasker.exe 1.10b has a problem with nested quotation marks if needed in -r option. That's why it couldn't be used to start squeezelite-win.exe with the needed parameters.
I've uploaded a new version 1.20 of Tasker.exe that can handle %% as escaped quotation marks. Here's an example based on Aleg's post some pages earlier:

Code: Select all

taskkill /fi "PID gt 0" /IM squeezelite-win.exe /F >nul
start C:\ProgramData\Squeezebox\Cache\InstalledPlugins\Plugins\LocalPlayer\Bin\squeezelite-win.exe -o "Speakers (Sonicweld Diverter HR2 Output) [Windows WDM-KS]" -a 2 -b 260:8000
c:\tasker.exe -all -x "squeezelite-win.exe|SqueezeSvr.exe" -p normal -a 0101
c:\tasker.exe -set -x squeezelite-win.exe -p realtime -a 0010
c:\tasker.exe -set -x SqueezeSvr.exe -p realtime -a 1000
pause

With version 1.20 this could be replaced now with the following:

Code: Select all

taskkill /fi "PID gt 0" /IM squeezelite-win.exe /F >nul
c:\tasker.exe -exe -x "C:\ProgramData\Squeezebox\Cache\InstalledPlugins\Plugins\LocalPlayer\Bin\squeezelite-win.exe" -r "-o %%Speakers (Sonicweld Diverter HR2 Output) [Windows WDM-KS]%% -a 2 -b 260:8000" -p realtime -a 0010
c:\tasker.exe -all -p normal -a 0101
c:\tasker.exe -set -x SqueezeSvr.exe -p realtime -a 1000
pause

Not a huge benefit of course as start.exe works also very good, but I'd like to make it complete.

Download-Link:
http://bit.ly/1e2tS6v

Christoph

[Aleg]

I've got nothing more to say on the matter, I'm not going to repeat myself if people don't want to read and understand.

Aleg, you've expressed a preference for certain buffer size combinations. Unless you've chosen the minimum buffers that work then you've just tuned the noise to the sound you like from your system, I find the smaller buffer sizes to be digital sounding which is consistent with anything that introduces higher frequencies into the playback chain whether it's system time, cpu frequency etc.

I've seen people describe the latest jplay 5.2 version as more digital sounding than the previous one, I would say that's a bad thing. I don't hear digital noise when I listen to a live band/orchestra, so I'm aiming for transparency.

Gordon

I have read what you said before and I think I understand what you're saying.

I'm just saying that MQn is still more transparant than JLP with any of the buffer settings suggested, mine or yours.
The maximum obtainable with JLP still doesn't show the amount of micro details in the music and that has got nothing to do with noise adding something I like. IMHO it has to do with MQn's capability to show more microdetails from the music and my tweaks minimise the noise from either MQn, WASAPI and whatever other OS-settings allowing the microdetails from the music come through.

What I can hear with MQn I cannot hear with JLP in any configuration.
And again what I hear with MQn is musical detail and not noise.

BTW
I don't like jPlay 5.2 either. It sounds strained and edgy, not as relaxed as MQn or JLP.

Edit:
I just got home and immediately tried the new v24 and I tried Gordon's buffers pf -b1000:20000 and compared it to 2.71 and Gordon I truly wish I could let you hear what I'm hearing, but there is no discussion possible about the differences. The 2.71 still betters it. I'm really sorry and now I will stop posting about it.


My last 2 cents ;-)

Cheers

Aleg

[sima66]

Gordon, MQn is your masterpiece so far and you should be very proud of it. You did something what even most of as could not believe it's possible.

As for the JLP, I have agree with Aleg and Nigel that JLP is still lacking the microdetails-transparency and punch of MQN. It's heading in the right direction, but still miss the fineness of the masterpiece.

The more revealing the system is, the bigger the difference is between the noise and microdetails. So far, in MQn I can find only air and transparency. One reason for that is probably because I never heard anything better! ;-)

[jesuscheung]

Nige, there are threads associated with the buffers and every cycle of the callback it is deciding whether to run a task to fill one or more of the buffers, not something you'd want going on in a minimalist player. Aim is to keep removing things until I understand what is going on and then make a standalone ks player. As someone said programming ks is not a trivial task.

guys, understanding and coding with ks is challenging. give it some time.

[tony]

Would echo Adam's comments. MQn for whatever reason still has the edge but as Jesuscheung's advises give it time. Very heartened Gordon is if I understand correctly trying to get JLP functioning as a standalone player using KS. I think this is might be the key to either closing the gap or surpassing what is achieved with MQn.

Personally for whatever reason I don't like the LMS interface but there is no logic to that view.

I know at the weekend in Fran's we did a quick test using V20 against MQn and soundstage is just so much larger and also retaining huge microdetail. We joked that it was expectation bias but it was clearly audible.

will give the latest version of JLP a bash tonight.

Just on some of the earlier comments by Gordon concerning diyaudio something similar is happening on PFM. Digital poll has a grouping of the usual suspects ascertaining that no either minute or no difference is achievable in digital audio. I am truncating the position but it is broadly similar aggressive 'we no best' mentality that is pointless to discuss with. It is terrible how they hang out on these forums waiting to pounce on any opinions or findings and bullying their opinion but in reality isn't this what happens in most walks of life.