I’ve been looking for an answer whether to use a frame width/height value which is a multiple of 8 or 16 pixels. The VP6 codec doesn’t seem to have any restrictions regarding framesize. I did find out that frame width/height values are not only codec related, but also depend on the used color space.
For FLV I use the YV12 color space. Both image dimensions have to be a multiple of two, if the video is interlaced the height has to be a multiple of four.