DPS 2.0 Testing (Windows)

[thx1881]

@ jim,

Thank you for that information, I will download version a14 and continue enjoying Music as it should be listened to.

[dps]

As much as I appreciate being part of the alpha programme, I’m not sure how much I can contribute. I use my rMBP as my computer 95% of the time, on which I have installed DPS v1.0.7 on Mac OS X 10.10 Yosemite. Because I wanted to try out the new DPS, I whipped out my ancient (er, ~2007) IBM ThinkPad. Whilst there is of course a quite massive improvement in audio quality before-and-after DPS on the ThinkPad, I have no way of making a direct comparison between DPS v1.0.7 and these new alpha builds on the same hardware unit. If/when alphas are released for Mac, I think I’ll be offer up more in the way of feedback/potential improvements.
Keep up the great work.

[bef522b1-0846-4b91-9c0e-64605f6ae0d5]

Btw does DPS processing (without stereo enhancer) change the original stereo image, e.g. by re-mastering each channel separately?

[xaqmusic]

hans,

DPS applies the same effect processing to both channels.  However, the dynamic range controllers are allowed to work independently.  This means the left and right channels may be affected slightly differently depending on the music or movie you are listening to.

If you are hearing sound stage enhancement when using DPS, this is usually due to the combined effect of the patented dynamic range controller and equalization.  When high frequencies are properly addressed in a DPS profile, the effect is a lifting of the sound stage and bringing the voice “closer” to the listener.

If you want more details just let me know.

[ruslang]

1.3.2015 who dps 2.15 BETA1?

[xaqmusic]

Hey folks,

DPS 2.0 a15 is available in the Dropbox

This version expires April 1.  It includes:

• A checkbox in the Settings tab for enable/disable volume settings during startup.
• A fix for GUI display problemms when the font size is not 100%.
• more skinning options.
• Stability and bug fixes.

This version is very stable.  As soon as our web profile delivery system is ready we will be moving to a more official release.  Stay tuned!

[jimboden]

Just tried a15 and turned off the auto-adjust volume which worked fine. It doesn’t apply to speakers, though, only head phones. DPS still sets the volume to 100% on exit when it’s set to speakers. That is consistent with previous versions, though.

[ruslang]

DPS 2.0 a15 this beta version?
how many beta versions of the final plan (release)?
and whether the RC version?

 

automatic volume adjustment at startup – applicable only for headphones and a laptop?

 

15a – put a load CPU to 8% (icore i3 4130) when minimized to tray GUI, (with audio playback),
at an open window GUI – load CPU is constant from 7.7 to 8.9% – Windows 8.1 X64

[ruslang]

Are there any plans to improve the quality 24bit – 48000khz?
and establish compatibility mode WASAPI Exclusive?

[xaqmusic]

ruslang,

1. Our first official beta will include profile downloading.  Even though all of the other features are finished we consider our software alpha until all designed features are testable.
2. We are going to refine the automatic volume adjustment in the next version this way:
   1. The checkbox enables/disables automatic volume adjustment at startup only.
   2. This will apply to all output modes.
   3. During shutdown of DPS, the current volume of the Digital Power Station driver will always be transferred to the hardware output driver.
3. CPU load is caused by 2 things in DPS:
   1. The frequency visualizer – When data from our C++ processing is brought into the .NET GUI, CPU usage increases.  Un-check “Enable Visualizer” in the Settings tab to prevent this transfer and save CPU usage.  You may see 50% reduction in CPU by disabling the visualizer.
   2. The DPS processing algorithm is very sophisticated.  The code has been heavily optimized but there are still a lot of moving parts.  We also have the ability to bypass some parts of the DPS algorithm if they are not needed.  We will be creating new optimized versions of some of the profiles that will incur less CPU usage.
4. The newest version of the DPS processing code will work at multiple sample rates.  And all of the DPS processing occurs in floating point math which is higher quality than 24bit.
5. Regarding latency and WASAPI compatibility – We don’t necessarily want to have the DPS driver work in exclusive mode.  The DPS driver must work with many different applications at the same time.  Exclusive mode may force the DPS diver into a sample rate or other mode that is not compatible with another app.  Also, we are aware of the small latency that is occurring and we are working to resolve that now.  Our OSX driver has zero latency so our Windows driver will too.

[bef522b1-0846-4b91-9c0e-64605f6ae0d5]

Regarding pt. 4:
Even though DPS processes in floating point mode, for the Windows OS the driver identifies as excepting only 16bit integers, right? Doesn’t that mean, that Windows OS will downconvert 24bit to 16bit *before* passing it to DPS, so losing some quality?

[jimboden]

I’ve experimented with turning enable visualizer on and off and found the opposite of what’s been stated here. When it’s off, memory utilization goes up, not down. When off, memory utilization runs up to over 200mb and about 80mb when it’s on. Whatever this memory utilization issue is, it’s the only task on my system which exhibits that behaviour. I have 62 background processes and 25 windows processes running and not one of them has memory fluctuation of any kind.

None of this may matter, but I find it very strange. There has to be an explanation for DPS acting this way other than claiming it’s got something to do with Net framework. Lots of apps use it and only DPS is behaving this way.

 

[xaqmusic]

hans,

You are correct.  I was thinking of the OSX driver which accepts up to 32bit int audio.  Our Windows driver is older and is currently locked into 16bit.  All we would have to do is provide 16, 24 amd 32bit float->fixed functions in the driver and we would support the higher bit rates.  However, there are some legacy concerns with this driver and, according to Microsoft’s driver team “is not supposed to work at all”.  We are investigating alternatives to the virtual audio driver method.  But virtual audio drivers offer the most flexibility to the average user.  It’s a tricky balance to maintain.

jim,

Have you tested this on a15?  We are using a different method to get the data from C++ into .NET.  It seems to be more efficient and disabling has a good effect on CPU usage (at least in my tests).  The big difference between DPS and most .NET apps is the constant real-time communication between the .NET GUI and the lower level C++ dll that is doing the audio processing.  .NET is like java in that it is interpreted instead of native.  So whenever .NET has to talk native there is some overhead.  We are working on optimizing this.

 

[ruslang]

xaqmusic

whether you can transfer the all text part GUI, from dps.exe> file FontsAndColors, as you did in version 15a, something it was possible to localize the text GUI dps2 – FULL?
p.s. – About dps for anroid: Prompt what better to choose a profile for a tablet Tegra Note7 – with NVIDIA PUREAUDIO 2.0 speakers
a. for BT headphones, b. for 3.5 headphones, and c. listen through the built-in speakers 2.0
[quote=62515]Even though DPS processes in floating point mode, for the Windows OS the driver identifies as excepting only 16bit integers, right? Doesn’t that mean, that Windows OS will downconvert 24bit to 16bit *before* passing it to DPS, so losing some quality?[/quote]

I’m talking about the same and say that windows mode wasapi cuts hkz & bits driver-level DPS, rather than 24bits with “,” and non 48khz

[jimboden]

Yes, my comments are for a15, which is why I posted them. Just tried everything again today and nothing has changed. I’ll try uninstalling and reinstalling it, which might change something. I normally don’t do that because a bunch of important things get deleted during the uninstall process, but I’ll try to work around that.

[Author]

Posts