Be very careful on your tests. Your techs won't be able to tell the difference between 'no delay' and delay within 100 milliseconds or so. The brain just fixes A/V sync below that - it's a feature!
So your testing may actually be 3 frames delay versus 4 frames delay, and it won't look like 'a bit more delay', it will appear to suddenly be quite a bit off.
The main cause of latency for you is probably de-interlacing. Any time you go from an interlaced to progressive display, such as in a display device, or a DVE in a switcher, 1.5 frames is a usual minimum.
Have you tried 720p, if you're using 1080i right now?
For testing - make a video with a clear frame indicator. I have one knocking around somewhere with a white stripe that rolls past some frame markers - 5 or so, I recall. Feed that to a monitor. Shoot the monitor with a camera and feed the camera through your signal chain right to your display. If you can frame it to see the 'original' monitor and the post-processing monitor or display, you end up with a video you can pause and examine.
Bruce
Regards,
Bruce Wheaton
http://www.synchromeshDDR.com