I am using the resizer in continuous mode. My camera input is D1, NTSC 720x480i 60Hz using UYVY format. I am trying to using resizer B to convert the format and upscale the input video. A few questions:
1. I see in the DMAI demo applications that the capture used 736 instead of 720. But I could not find exact reason. Is it because each horizontal line has to be 32 bytes aligned? If yes, is this alignment because of capture, previewer, resizer or DDR memory?
2. In my case the input format is UYVY. It is actually 32 bytes aligned by itself for each horizontal line. I think I dont need to use 736 right?
3. Assume if my input format is 420SP, I would have to use 736 instead of 720. But what is the addtional 16 bytes come from? at the end of each line? are they only padding which contain no video data?
4. If I use resizer B second output to do format conversion from UYVY to 420SP, the input is 32 bytes aligned. But the second output is not, because after converting, it will be 420SP format. Do I have to set resizer output2 width as 736? If yes, does it actually do resize? In this case I dont want resize but only format converting.
5. If I use resizer B to do resizing as well, from 720x480 UYVY to 1080x720 (that's 1.5x) 420SP format, the output2 1080 horizontal line is not 32 bytes aligned. Do I have to make it 1088? If yes, output2 video is not exactly proportional.
Thanks.
