When looking at the video data, theres a red line to the right of the picture, not sure why its there. It flickers blue once a while, but mostly stays red..
Its like this: (an example)
On the X axis.. ~200 pixels of video data, ~20 pixels of black, ~20 pixels of red line.
The redline could be used to "communicate" data to the computer.. Its possible that its value can be changed slightly (i.e. from 255 to 254 then to 255), but I'm not sure if this is waht happening.
So far, to me, the data seems to be going through the bulk data interface, which is an endpoint in the NT1004 chip.
The NT1004 chip communicates with the SAA7113H through a I2C interface. USB Vision project supports this, and thats why with USB vision, you can see the video input from the camera in real time.
However, USB vision doesnt support audio from the NT1004 yet. The author has made some attempts into doing audio. I've used his beta audio software, which has erros for me, I will work on it later on. The USBVision project supports radio devices, which would communicate through the I2C interface to the NT1004, and so forth.
However I know that the audio isnt being streamed through the audio port (yet) for the USBVision project because in order to do that, the AUDIO_CONT register would have to be enabled, and by default its disabled. The AUDIO_CONT register also controls if the bulk data is streamed through it.
To me, it seems that the data from the sensors is either being streamed into the picture data or is being streamed into the audio endpoint. Remember, I cant eliminate the picture data as a possible sensor data source yet because of the red bar in it.
Using XAWTV and USBVision from CVS - 2/15/2005..
TV Norm: Auto. Also works with NTSC.
Video Source: Composite 1 gives 1st camera output. S-video doesnt give any output from any camera inputs. Both show a red line to the right of the picture frame.
Audio Mode: Mono
Capture: grabdisplay. Overlay doesnt work, but thats because of the way my system is setup.
Using Gnomemeetign and USBVision from CVS - 2/15/2005..
In video devices (in preferences):
Video Plugin: v4l
Input device: USBVision USB Video
Size: Large (small doesnt work well for some reason).. Large is CIF 352x288, small is QCIF 176x144.
Gnomemeeting detects 4 channels (channel 0 through 3). All 4 show the red bar.
When camera is plugged into cam1, and channel is set to zero, it shows video from the camera.
But when channel is changed to 1, it doesnt show video from cam2.
Channel 2 and 3 show a line pattern, but gnomemeeting shows video from channel 0 on channel 3, and the line pattern on channel 2.
All the channels that have the redline, lose the redline when the size is changed to small. However, when the size is small, the video looks messed up. It looks normal with a large size.
I downloaded and ran Palantir, which allows streaming of audio and video over the LAN/Internet. It has a server and client part.
The output of the server is:
[main] -- palantir 2.5.5 starting --
[main] Definitions for 0 devices found
[main] No serial port specified
[main] No named pipe specified
[video] Card: USBVision USB Video (/dev/video0)
[video] Capabilities: 41 (capture overlay clipping)
[video] Size: (64x48)-(320x240)
[video] Channel no. 0 ('Composite 1') tuners: 0, flags: 2, type: 2, norm: 3
[video] Channel no. 1 ('S-Video') tuners: 0, flags: 2, type: 2, norm: 3
[video] Channel no. 2 ('S-Video') tuners: 0, flags: 2, type: 2, norm: 3
[video] Channel no. 3 ('S-Video') tuners: 0, flags: 2, type: 2, norm: 3
[video] Unmuted audio input 0
[video] depth: 24 palette: 4
Connecting from the client to the server, gave me the same video that I got in gnomemeeting and xawtv.. Except that the FPS was much better. My laptop that I'm testing USBVision on is a P3 733MHz, and doesnt have a very good video card. So naturally video applications on the laptop will be a bit slow.
Also the color via palantir seems much better. If USBVision can be modified to support audio from the NT1004, maybe Palantir can be modified (with permission) so that it can be used for Xanboo remote monitoring. The fps using palantir is ~16.50 fps on average.
The red line is still there.
Trying 300x200, the video is messed up, same with other modes. 320x240 works. I took a screen shot of it. I set the window so it will not stretch the video to the window, so I can say for sure that the video being outputted by the NT1004 is smaller than 320x240 but its enlarged to 320x240.
For example: When I reduced y to 100 (320x100), the y axis of the video window reduced. However, the X stayed the same, and I could still see the picture. However if I reduced the x axis (for example, 160x170) the video always got messed up. Therefore, the x-axis is fixed to 320, while the y axis can be resized. So the red line could be a way for the xanboo client to know where to "clip" the screen. The picture is in this folder.
Since I'm doing more research into this, I may start putting documents in that location instead of here because this topic is in "News, Breaking news" and I'm not sure if the administrators/moderators approve of me constantly updating this topic as I find more information. I'd rather not have my access removed. I'll also save a copy of this thread into that location because alot of my notes are on here.
I looked into usbvision.c, and it seems that usbvision is the reason why xawtv and other programs print out "composite" and "s-video".
Remember that USBVision is mainly for tuners, capture devices and some other devices like that (as far as i know) and wasnt made for Xanboo and does not include support for xanboo natively. I'm working on that. I may get in touch with the developers of USBVision about myself writing code for it that will support Xanboo..
However this will be my first attempt at programming a driver, and thats one reason why I'd rather work with USBVision than to start my own driver since I'd have to program it from scratch. However, there are some specifics in Xanboo that are not found elsewhere.. Such as the possible usage of the audio endpoint for bulk data, which would require coding the driver in a way that may end up making a specialize device (/dev/xanboo?) just to output the audio. And thats not including the camera select chip.
In addition to this, theres a chip that supports switching between multiple video sources.. This is connected to the SAA7113, and would explain why there are only 1 workable video source in the programs that I have tested so far. I'm not sure why.
Last but not least, the red line could be used for sync. In other words, when the picture frame is synced right, the red line will show up right. If its synced wrong (i.e. X isnt the right value), the red line will not show up.
Last but not least, jsmmd, U7 is NT1004 not PIC16C57-HS/P.