I'm trying to come up with the best (as in: highest quality for a given compression ratio) way to encode a video stream using the ADV212. It seems the compression is better with progressive-scan video, but I'll certainly have to support interlaced video as well, in fact I may *only* have access to interlaced video.
So, ideally I'd want to convert the interlaced video to non-interlaced before handing it off to the '212 to perform its compression on the full frame rather than individual fields. I've read that the '212 can do some basic de-interlacing but it's tough to get the parameters correct (and they have to be perfect) between the video decoder chip and the '212, especially in PAL (which would be a problem for me). The de-interlacing done by the '212 is also not motion-adaptive so you see visible tearing on high-motion areas.
So, I started thinking I ought to de-interlace before presenting the video to the '212, but I can't find much that AD does that would support that.
- There's the 780x parts, but they seem to be non-adaptive as well (so I'm assuming this haspoor performance on high-motion areas as well).
- The ASV7186 seems to offer 'edge-adaptive interlaced-to-progressive conversion' but I can't find a decent datasheet for it. I have some other requirements (which is why I'd like to have a datasheet), for example I want the VITC decoded if it's there, and it's not clear if the ADV7186 supports this.
Is there anything else that I've missed by looking in the wrong parts of the site ? Doing interlace->progressive seems like it ought to be right up AD's street, as it were. I'm surprised at the small number of options I'm finding, which suggests I'm looking for the wrong thing Any help appreciated!