www.ssontech.com

[PaulF]

Hello.

I'm currently trying to add some overscan/padding to my NTSC footage, after removing the lens distortion. I'm using the 'Crop' tab in the image prep dialogue to do this. However, when I save out the sequence, then load it back in to start matchmoving I find that it has been saved out with square pixels. When I change the pixel aspect to 0.89 the picture looks like this:



As you can see, with all the jaggies it's totally unsuitable for tracking.

Is there any way to fix this?

Many thanks.

[milesl]

Is this progressive or interlaced footage? Are you setting it to non-square/DV size pixels when you import initially? On export of the un-distort did you set aspect ratio to 0.89? Though i believe 720x480 footage is 0.9 and 720x486 is 0.91. Maybe using those numbers will help?

[PaulF]

Thanks for the reply milesl.

It's 24P footage which I shot with my DVX100. It's 720x480 NTSC, and when I import it, Syntheyes automatically detects all the above and sets the pixel aspect ratio to 0.89. So, inside Syntheyes, everything looks OK. It's just that the image prep renders out the padded footage with square pixels. There isn't anywhere (That I can see) in the output section to force it to write a particular pixel aspect. I just assumed it would write it out the same way that it came in.

Cheers.

[milesl]

In the output box there is an option to resample, and you can set the proper aspect ratio there. You'll have to add in your padding to the final output size in order to get a 1:1 pixel ratio. For example, if you've added 100 pixels in your crop to width and height your output size would be 820x580 with the correct aspect ratio for the footage.

[PaulF]

Thanks again Milesl . I'll give that a try.