Proper Motion Compensated deinterlacer (ala Snell)

Archive of historical feature requests.
Please use the GitHub link above to report issues.
Forum rules
*******************************
Please be aware we are now using GitHub for issue tracking and feature requests.
- This section of the forum is now closed to new topics.

*******************************
Post Reply
mcpish
Posts: 14
Joined: Wed Apr 20, 2011 9:19 pm

Proper Motion Compensated deinterlacer (ala Snell)

Post by mcpish »

This would be my dream feature. I say Dream because I've never seen any piece of general consumer grade video encoding software (open source or not) ever include it. I would love to see a proper form of "Motion Compensated deinterlacing and standards conversion using proper tweening"

To give you an idea, I'm thinking along the lines of something like what a SNELL broadcast standards converter like this does:
http://www.snellgroup.com/products/conv ... -ph.c-hd1/

I'd love Handbrake to have some sort of deinterlacing filter that did proper "field extension deinterlacing" (ie. bob up to 50/60Hz rather than blend down to 25/30Hz as it currently does). It should also be able to do proper hi-motion 50Hz PAL/DVB to 60Hz NTSC/ATSC conversions with proper motion compensation tweening (ie. interpolate and generate new frames based on motion and timing, not simply duping or droping frames to match the new refresh rate). It should be able to convert from any frame rate to any other frame rate using motion compensation to generate new frames that match the temporal dimensions of the target framerate. This is the kind of stuff a proper hardware based standards converter such as the SNELL converters do for broadcast operations.

I know I'm dreaming here, but this is ideally what I'd love. I'd love to be able to take a 1080i25 (DVB) MPEG-2 transport stream, and PROPERLY encode it to an h264 encoded mp4 in 1080p60 and maintain perfect fluidity at the new 60Hz rate by having the deinterlacer in Handbrake be able to take the original 50Hz interlaced signal and properly tween-up and interpolate the missing frames and with proper timing. I know this is technically possible, since broadcast standards converters can do it, but that it's also extremely CPU intensive since you're basically generating brand new frames to match the new timing (ie. creating data that isn't there based on interpolating the two closest frames).
Post Reply