Devialet Chat

I use a 'single' network all 100 Mbit/s switched for clients included the music 'boxes'

Mine could be described as a double network: my Internet surfing (when necessary, otherwise it's off) is done via WiFi, while music streaming to the D is done straight from my MacBook Pro to the D with a single Ethernet cable, using Air. No white noise whatsoever.

Hope this helps. Thanks for your efforts!

L


Sent from my iPhone using Tapatalk
MBPro 2008 > Dev 120 > Q Acoustic Concept 20 (soon to be KEF LS50). Location: Buenos Aires, Argentina.

Single network.

A question for all you IT wizards out there. Today I am doing a bit of white noise hunting per Roger's (Devialet) request. He asked me to provide screen shots showing what processes are running following a white noise event. OK, fine. I can run task manager, view the processes, filtered on CPU, hence those processes running are at the top of the list. I can then use the snipping tool to capture a screen shot of the processes running both before and after a white noise event. One point that strikes me is that the processes are very transient in nature, this this only captures what is happening at an exact point in time, that is, something might cause white noise, stop running and drop to 0% CPU usage and effectively become hidden.

Is this the best way to capture decent data? Or can anyone suggest a more effective way of going about this?

I would not sort on CPU usage - I'd sort on the process name, since that way it's easier to compare before and after (as you rightly pointed out, CPU usage is volatalie). So, capture the processes before the white noise occurrence (make sure you click "Show Process from all Users"), then do what it usually takes for you to recreate the white noise, and capture the processes again. Hopefully, there will be a meaningful difference. Try to run as few programs as possible, not to muddle the picture with extraneous data.

(23-Nov-2014, 04:55)AlexS Wrote: [ -> ]I would not sort on CPU usage - I'd sort on the process name, since that way it's easier to compare before and after (as you rightly pointed out, CPU usage is volatalie). So, capture the processes before the white noise occurrence (make sure you click "Show Process from all Users"), then do what it usually takes for you to recreate the white noise, and capture the processes again. Hopefully, there will be a meaningful difference. Try to run as few programs as possible, not to muddle the picture with extraneous data.

The problem is that there are so many processes listed that you need to scroll through the list to view, they will not all fit on the screen. This is why I filtered on usage. I also have some screen shots taken during normal listening, where the CPU usage for Devialet AIR is shown as 0%. (!) So I am beginning to wonder if this method actually has much value? Is there a better way of doing this?

(23-Nov-2014, 10:10)Confused Wrote: [ -> ]

(23-Nov-2014, 04:55)AlexS Wrote: [ -> ]I would not sort on CPU usage - I'd sort on the process name, since that way it's easier to compare before and after (as you rightly pointed out, CPU usage is volatalie). So, capture the processes before the white noise occurrence (make sure you click "Show Process from all Users"), then do what it usually takes for you to recreate the white noise, and capture the processes again. Hopefully, there will be a meaningful difference. Try to run as few programs as possible, not to muddle the picture with extraneous data.

The problem is that there are so many processes listed that you need to scroll through the list to view, they will not all fit on the screen. This is why I filtered on usage. I also have some screen shots taken during normal listening, where the CPU usage for Devialet AIR is shown as 0%. (!) So I am beginning to wonder if this method actually has much value? Is there a better way of doing this?

You could take two separate screen shots, if the processes don't fit one screen (I know they don't).
Observing the processes before and after may have some merit, but I thought you said the problem only happens to you at higher resolutions. If you never get it with JRiver and WASAPI, with the 16/44 format (is this the case?), I don't see how this can be caused by an interfering process.

Day off work today and much extensive listening. I was using JRiver WASAPI with everything set to 16/44 to avoid white noise. But White noise came regularly after a couple of hours, and often repeated even after a full reboot. Gave up and switched to iTunes, which is supposed to be bit perfect if you set everything to 16/44. I don't think it sounds quite as good as JRiver WASAPI, but ran without white noise for 4 hours. Interestingly, I kept the WASAPI bit perfect setting in the AIR GUI when running iTunes, so it looks like the issue is caused when the actual music player is feeding WASAPI. Very odd, as I still got maybe two hours before any issues with JRiver WASAPI.

(29-Nov-2014, 01:01)Confused Wrote: [ -> ]Day off work today and much extensive listening. I was using JRiver WASAPI with everything set to 16/44 to avoid white noise. But White noise came regularly after a couple of hours, and often repeated even after a full reboot. Gave up and switched to iTunes, which is supposed to be bit perfect if you set everything to 16/44. I don't think it sounds quite as good as JRiver WASAPI, but ran without white noise for 4 hours. Interestingly, I kept the WASAPI bit perfect setting in the AIR GUI when running iTunes, so it looks like the issue is caused when the actual music player is feeding WASAPI. Very odd, as I still got maybe two hours before any issues with JRiver WASAPI.

Is this a consistent pattern for you: about one (or so) hour of music before the noise with the HD format, and over 2 hours with 16/44? Smells like a gradual overflow/corruption of some memory buffer, proportional to the amount of data fed to it. Please keep on sharing your observations with Devialet. One day they may have a revelation.
Meanwhile, I've been using JRiver (MC20) in place of Foobar for about a week now. No white noise thus far, but no prolonged listening sessions either. I did notice an interesting thing, though. One day I got 0.5 - 1s dropout in music, which I eventually was able to trace to Windows Defender (anti-virus), which automatically installed an update and ran an auto-scan, 20 minutes into the listening session (when it thought the CPU was idle). The entry in the history this created coincided with the dropout. As a result, I both disabled auto scans in Windows Defender, and also excluded one JRiver folder from the scan (per their recommendations on the Wiki). I don't know if anything like that can be related to the white noise, but I do recall now that I was getting the problem more frequently in those listening sessions when I turned on the computer for the 1st time during the day (and this is when all the updates/auto-scans kicked in, since they were all configured to run nightly, but the computer was always turned off at that time). This does not look like a mere coincidence. I've also disabled automatic windows updates.
All in all, if you have Windows Defender or another anti-virus, could be a good idea to disable it for a while, to rule out (or, maybe, in) this possibility.
BTW, I like JRiver, and will be switching to it from Foobar. Has nothing to do with SQ, but the interface, and JRemote, which beats its Foobar counterparts.

I found some interesting information on the Internet, regarding the white noise with a different sound card called VXpocket. I'm quoting it below. It corroborates Devialet's theory of interferance from an extraneous Windows process. It also does not contradict our own observations around the increased likelihood of the white noise with the HD formats (more RAM used by DevialetAIR, as a result?)

"During playback, I sometimes get white noise all over sudden - what happens?

Several clients have reported this issue on the following laptops:
ASUS L5GXP30-C PRO
Dell Inspiron 5100/8200/4150
HP ZD7280
It seems that another device uses the memory addresses assigned to the VXpocket by Windows, sending non-audio data to the card.

Possible workaround

On the Dell Inspiron models 8200 and 4150, change the I/O resources of the VXpocket in the Device Manager selecting an address above D800.
For the other above mentioned laptops, we haven't found any workaround so far."