this post is regarding a question i have about barlow's nattress slow-mo method as described by him in this thread:

http://www.dvinfo.net/conf/canon-xl-h-series-hdv-camcorders/87670-slow-motion-question.html

barlow says:

"My trick is to put a 1080i clip into a custom 1080 60p (50p for you)
sequence. Usually I make it an uncompressed timeline to minimize conversion artifacts. (most slow mo clips are short lengths anyway)

You add the Nattress Standards Conversion/Map Frames filter to an HDV 1080i clip on the timeline and choose "Fields to Frames HQ" and choose upper field first. Also, choose "normal" for deinterlace, rather than "smart". You should see the "combing" effect of interlace fields on movement within the image go away if you've done it correctly. It may have a slight amount of aliasing/jaggies on the image, but it should basically look progressive at that point. You then output a 1080 60p (50p again, in your case) clip and bring it into Cinema Tools and conform the frame rate to 24/25p."
------------------------------------------------

if I'm understanding correctly then what is ultimately displayed in the final output would be frames with every other field missing. So when you run nattress's map frames filter wouldn't this just add black pixels in place of the missing fields? Wouldn't this significantly decrease the apparent brightness of the image?

I suppose you could do some kind of a field interpolation, but wouldn't this defeat the point of doing the nattress method in the first place?

i'm trying to decide if i should by nattress's software so any illumination on the subject would be much appreciated.

Hi Cal. I'm not familiar with the natress method, but when you create a frame from a field you don't replace the "missing" field with black. What happens is you get a frame that is half the height of a full frame, then you use a scaling algorithm to resize it back to full height. It makes the image a bit softer but there are no interlace artifacts.

Richard

aahh, that makes so much more sense! thanks for explaining that. so i guess the trade of is just a loss of vertical resolution? Sounds like a pretty good trade off for some smooth slow-mo...

Hi Cal. Yes, straightforward deinterlacing causes loss of vertical resolution. Then there are the smarter or adaptive deinterlace methods that try to maintain vertical resolution as much as possible, by using the details from both fields in the areas where there is no movement. The success of this is heavily dependent on the content of the footage, e.g. locked down shots with a static background will stay pretty sharp.

Richard

so i guess thats why barlow tells us to check "normal" deinterlace as opposed to "smart"? Am I correct in assuming using the smart option would create motion artifacting in subject moving across the frame?

oh duh, i guess he addresses that in his very next sentence,

"You should see the "combing" effect of interlace fields on movement within the image go away if you've done it correctly."