Page 1 of 1

Receive Tests / CPU Usage / Firefox vs Chrome

Posted: Sun Apr 30, 2017 9:50 pm
by CalS201
Recently performed a "receive only" test and compared CPU usage using Firefox and Chrome browsers on a Dell Optiplex 580/Windows10/Dualcore AMD@3.0Ghz. Results indicate CHROME requires much lower CPU power than FIREFOX. This could be useful information if your computer is max'ing out on CPU power.

FIREFOX_____________PreBroadcast Video(MoTab)___ Broadcast Video(Joseph Smith)
CPU %----------------------------73 %------------------------------------80 %
Highest Bitrate------------1500 kbps-----------------------------1320 kbps
VideoBuffer Size---------->270 Secs-----------------------------60-70 Secs
AudioBuffer Size-------------60 Secs------------------------------60-70 Secs

CHROME_____________PreBroadcast Video(MoTab)___ Broadcast Video(Joseph Smith)
CPU %----------------------------47 %------------------------------------57 %
Highest Bitrate------------1500 kbps-----------------------------1320 kbps
VideoBuffer Size---------->270 Secs-----------------------------60-70 Secs
AudioBuffer Size-------------60 Secs------------------------------60-70 Secs

I think it would be helpful if the Portal used "receive test" videos of the same bitrate, encoding, & buffer size that is used when receiving a stake conference webcast. Presently there is no option to do a receive test at 2200kbps (the Vidiu HD setting).

Also, the Prebroadcast MoTab audio is compressed, consequently it "sounds" louder than the audio produced (without compression) at a stake conference. As a result, you cannot use it to set the volume levels before a stake conference begins-if you do then you will probably have to raise the level once the conference goes "live".

Using "test" video/audio that has the same characteristics and bitrate ranges as stake conference video/audio is important when troubleshooting/testing/validating and gaining a confidence level for your equipment - especially at the higher bitrates.

I REALLY appreciate having the present "testing" ability and use it often - these are just suggestions to make the portal a little better.

Re: Receive Tests / CPU Usage / Firefox vs Chrome

Posted: Tue May 02, 2017 9:58 am
by rannthal
Great observations, info about the browsers and suggestions. Keep them coming.

The "receive test" at first was there just to see if the locations could connect to the media servers. It now needs to evolve into something more helpful.

Re: Receive Tests / CPU Usage / Firefox vs Chrome

Posted: Tue May 02, 2017 10:30 am
by seanhyte
CalS201 wrote:compared CPU usage using Firefox and Chrome browsers


Interesting info. I would be interested a test with Microsoft Edge on Windows 10, since they claim it uses less battery, I wonder about its CPU usage?

Re: Receive Tests / CPU Usage / Firefox vs Chrome

Posted: Thu May 04, 2017 6:59 pm
by Mikerowaved
There's another aspect of Firefox that should be brought to light and that's memory usage. This past Sunday was a successful stake conference webcast using Wirecast via YouTube Live on a very capable PC. Sunday was great, but I couldn't say the same about Saturday's "dress rehearsal" of trying to webcast the adult session. Wirecast kept unexpectedly crashing, sometimes running for only 10 minutes, sometimes after an hour or more. It didn't do any of this during extensive testing at home. Of course, we're pulling our hair out trying to find what the problem was.

In the wee hours of Sunday morning, I was studying different aspects of Windows, when I noticed the memory usage slowly creeping up. Looking deeper, I saw that Firefox was the one and only app with increasing memory usage. I had a single Firefox window open to download our test stream and display it on a second monitor. When Firefox had consumed all available memory, Wirecast would crash. When Chrome was used in place of Firefox, everything sailed as smooth as could be.

I don't know if the Firefox development team is aware of this memory leak or not, but I just became a much bigger fan of Chrome than I used to be.