Hi Arvind,

Thank you for trying. Not hearing a difference is valuable information as well!

regards, Coen

I've been experimenting a lot with xx settings to get a better grasp on what they do to the sound of my system.

Being back on engine 3 with considerably lower sfs, I found a very nice sounding combination, still with the abovementioned 'Not switch during playback' feature on.
What immediately occurred to me was that the sfs peak in the first core was very low and that the processor load indicated something like 58% on the sfs peaks. With the 'not switch...' feature disabled, the peaks are back on 100% again, that is a max peak in all cores.
Also the sound changes from focussed and clear to warm and fuzzier. I tend to prefer the first (with the feature on).

Imho this is an important feature for sq, at least on my system.

Regards, Coen

I experienced this dropout problem. Very annoying.

I reebooted several times to make shure the -latest- driver is properly installed. I guess you also have the nec/renesas usb chip; the usb3 control panel that can be found in the start-programs area provides a checkbox to enable/disable the powersaving features. This should be off.

I did and undid a lot of stuff to solve this problem, but the combination of a recent driver and the powersaving off did the trick best. Second best was increasing the minimum cpu load% in the powersavings area of the control panel. This is not neccesary anymore.
Fyi I downloaded the latest driver from the Intel site (not the gigabyte one). Check your version of the usb driverchip first!

I guess it is the NEC/Renesas chip because the Etron doesn't seem to work at all (complaints all over the place, Google for it). In this case you better buy a pcie card with NEC/Renesas chip.

Regards, Coen

Peter,

When I start with playing Wavs, the peak stays in one core. The PC has to have some extra load when starting a new track to "spread" the peak. A shure bet for switching over to all cores is playing a FLAC file. There are other causes too.
After this switch all cores remain active on SFS peaks (no revert to single core peaks).

I attatched two recent screendumps of my processorload (note: "not switch during playback" is on otherwise the peak would have been in the first core).

Something that caught my eye was that at the loading of the very first track after a reboot, the first peak is in the first core. Next ones are in the second core.

I like the sound best when spread over all cores (shortest duration of the peak, but at 100% cpu).

Does this help?

regards, Coen

Just to share an experience with possible sq benefits.

I was in search of a solution to usb3 related dropouts and fiddled around with processorscemes and playbackpriorities. This afternoon after a fresh boot i found myself having better than usual sq, but no special setting occurred to me. Another break-in leap of the NOS?

It occurred to me that the processorloading also was different than usual with a single sfs loading peak in the second core, where it is usually in the first core with the driver (as long as that lasts). Running again through the settings I discovered that I left 'Not Switch during Playback' in the processorsettings on. Disabling this feature restored the processor loading and warped me back in sq. Sq returned after putting it back on.

Oh and even the sfs peak is now distrubuted over all cores it still sounds -imho much- better with the setting on than off...

Maybe this does something for your sq too.

Happy listening,

Regards, Coen

Well maybe that matters too.

Some guys arereally serious about -ikea- chairs, looks like you bought the right one :


Enlighten yourselves:
http://michaelgreenaudio.com/forum/viewtopic.php?p=4170&highlight=ikea+chair#4170

(couldn't find the topic about colour!)

Regards, Coen

I think these settings adress most of my reservations about the SQ of the PCIe USB3 card. I've been searching in bios and os, but it was there in xx all the time (i did correct some of the other settings though).
The subdued highs that Peter mentiones above have been partly the cause of my search, now they seem to be right again.

Thanks!

Disabeling the left over USB2 controllers (chipset and separate) in the device manager does add a bit to the clarity and experience. Unfortunately if you forget to swich them back on before shut down, the next boot will stall somewhere in the process. You will have to restart to the 'last known good configuration'.

Fwiw I rebooted five times in a row as advised for a cure for reboot related issues on the windows support forum. With the new settings and fresh situation the cpu graph in the task manager looks much more  the usb2 one. Not completely the same, but much more alike than my previous posting.

Regards, Coen

Hi,

I did this and it didn't work. The splitter is not in the formal dual ps2 port standard which is an OR configuration. It was assumed that either the mouse or keyboard use USB.

The custom implementation of a two function AND PS/2 combiport has been in use with mostly older laptops that were supplied with the splitters in the cable set.

The manual of any modern mobo will quickly inform you about the OR nature of this port.

Unfortunately.

Regards, Coen

Thanks Nick,

Very interesting. Is this fix not part of the sp1 update? According to Peter the peaks are the result of xx playing. You have no peaks?

Regards, Coen

They are made with exactly the same settings.
The USB3 driver part of the graph is almost twice as high as the USB2. USB3 peaks at 8%, USB2 at 5% (as you can see). The flat part is 5% on USB3 and 3% on USB2.

I took the best looking USB3 shot to illustrate the difference in load. I have pics where the driverpart is hardly recognisable.

yesterday I updated the mobo Etron USB3 driver to the latest version (21-3-2012), but still I get the "set pauze on pin error". Strangely enabling the USB3 on the mobo did something positive to the sound of the PCIe USB3.

regards, Coen 

Here are my pics.

PE turned on, reinstalled the XX driver on the USB3 port and killed the usb3mon process.

Stefanobilliani made some sq comments that I second. The sound on USB2 is more compact and less controlled. Tonally there is quite a big difference in favour of USB3. Still something irritates me about it. Maybe it is because on USB2 I hear more of the instruments playing together. Maybe my cr*p recorded tracks sound more like they ought to. Still looking for it.

On the other side: I find it hard to believe that the cpu load of USB3 has a favourable effect on SQ, au contraire!

regards, Coen

Hi Scroobius,

What I mean is that in your pics the USB2 core #8 has a nice driver pattern. In the USB3 situation this apparently changed to core #3 (I think) and doesn't look as tidy.

Later I will post some pics of my CPU usage (only 4 cores).

SFS loading peaks on my cpu are 100%, strangely this is no different than on my atom (with lower sfs of course)...

regards, Coen

Fyi peak extend has been off.

Regards, Coen

Yep.

Regards, Coen

Hi all,

I don't share your enthousiasm for this route. In someway I can't reconcile with USB3 and its sound. Dispite cracking issues that seem (hopefully more than temporarily) to be solved after a driver update, I think the sound lacks something that I like and is there on my USB2 ports.

It occurred to me that, even with the latest driver (NEC/Renesas 720200F1 to 2.0.30.0) the CPU isn't loaded as tidy on USB3 than it is on USB2. The latter has a very nice repetitive pattern with the SFS controlled loading peaks, but playing on USB3 it looks messy, with more cpu load on activity. Probably you are able to reproduce this yourself.

I am using a OEM Western Digital USB3 PCIe card with abovementioned NEC chip.

I'm back to USB2 for now.

regards, Coen