Using the screen codec to capture from a device
- I've got a capture device captures a VGA feed from another computer, I'd like to use the new screen capture codec in Expression 3 to capture the feed from this device and then later compress it down to VC1 to maximise quality and minimise CPU use durring the capture. Is this possible either from the UI or the API?
Thanks
All Replies
- We don't currently support using the screen capture codec in our Live encoding mode although it is a great feature request. Are you able to share what brand of capture card you are using?
thanks
James The capture card (well USB device really!) is an Epiphan VGA2USB-lr
http://www.epiphan.com/products/frame-grabbers/vga2usb-lr/
Very handy little thing. I use it for recording the talks at the annual company conference, see http://video.dyalog.com/Dyalog08/ for the results from last year.
At the moment I'm using Windows Media Encoder to capture a VC1 file, then re-encoding that to produce the silverlight & mp4 content using Expression Encoder.
I'd love to use Expression Encoder end to end (under programatic control) AND use a lossless, fast, low processor use codec for the initial screen capture at sizes up to 1280x1024 at 20-30fps. File size isn't an issue, apart from lossless uncompressed capture produces data too quickly to write to disk!
Jonathan- 1) Couldn't you just play the video full screen on the recording pc and use Endoder to screen capture the full screen video playing to acheive the same net effect?
2) I am curious, would video card with hdmi out to hdmi capture card on second pc work? - William,
The capture machine is a headless Vista machine controlled remotely over RDP (remote desktop) from an eeePC running XP (so very small screen). Durring capture it has a couple of applications open, the capture software (a custom wrapper around Windows Media Encoder), a volume meter and the windows mixer applet to control the audio levels. So unfortunately re-displaying the captured image on the capture PC and running a screen capture wouldn't really work :( But it's an intriguing idea - I can certainly think of some cases where it may work well :)
As for the HDMI capture I don't know (haven't got a HDMI capture card to play with). But I suspect it would have the same limitation from the point of view of Expression Encoder, in that you can't use the screen capture codec from the capture API...
But thanks for the idea :) - You can easily do a FREE screen capture with UScreenCapture by Unreal Streaming.
http://www.umediaserver.net/
After the instalation your encoding program EE3/WME9 shows a full virtual capture device.
As a streaming company I can say this filter and "Virtual Audio Cable" program saves my day many times with excelent quality and for a fraction of the cost compared to a professional SDI capture card (in the case you have your player in the same encoder computer).
Christian - Thanks. I tried it. Is a great idea having screen cap as virtual device. It is kind of wild to see the live Cued screen cap of the Encoder screen you are working in. Kind of a recursive mind bend. However, the quality for screen (1680x1050) was not readable at all. Maybe there is knobs I need to turn. Thanks for the info and the v-audio cable looks cool too.
- Manketlow,
I'm planning to use the same Hardware to encode meetings, can you give us some more details regarding the architecture you're using to grab video from vga and encode with Media Encoder?
How much CPU power do you need? What is your current configuration?
Thanks for sharing details.
Angelo Angelfly,
I'm using a 3Ghz dual core AMD chip - (I think it's an Athlon 64x2 6000+) with 2GB of ram running Vista. Memory doesn't appear to be much of an issue - but the cores are maxed out durring capture - so a quad core may help.
I used the WMV power toy from http://citizeninsomniac.com/WMV/ to set:
Filters
In-Look-Filter = on
BFrame
B-Frame number = 1
B-Frame Delta QP = 1 (higher puts too much load on system)
lookahead = 16 (big improvement)
And in Windows Media Encoder:
CBR WMV9 advanced
25fps
video bit rate 5000K (setting the codec to quality VBR gives very similar file size, but puts more load on the capture machine)
video size same as input
key frame interval 1
default buffer size
smoothness 0
Audio: WM10 96kbps, 44khz, 16bit cbr
Which works quite well...
And then I run the files through Expression Encoder to create the silverlight files...
Hope that helps :)
