But NEVER go to H264 in an extra step, it will never help, it will always hurt, it will take extra time, doesn't make any sense.
I never recommend using aif audio for DVDs especially if you are using DVD-Rs, unless you are producing a professionally mastered musical event. No hollywood title is released with PCM audio except for concert type titles. This of course means encoding before bring it into DVDSP.
Sorry for lack of clarity. My workflow is that I have a library full of photo jpegs and am making DVDs out of different selections. I have been told by other editors that going to h.264 will introduce sufficient frame blending that will (partially) solve the interlacing issues that we have by these DVDs being shown on progressive monitors (and most of the library is 29.97fps with pulldown added). So i'm wondering about advice as to if h.264 quality can approach photo jpegs.
I reread this thread, and a I got a little more confused on somethings and a little clearer on other things, I think :-/
What are you starting with? Still photographs? Uncompressed video? Uncompressed video of still photographs?
What you are saying the benefit of H264 is helping would be because it is softening the image with its compression scheme, so it maybe helping because the loss of quality.
But please explain exactly what you are doing and what tools you are doing it with. There maybe no reason to add interlace to the footage anyway. A DVD player will interlace progressive video better than it will deinterlace interlaced footage. So you maybe looking for a way to solve a problem with step 4 that you have created in step 2. Your solution maybe to figure out a better step 2 and be done with it.
I think the main confusion may be the photo jpegs. That is the quicktime codec that all of our uncompressed media was conformed to for our library, which is a few thousand clips. I'm not sure why that was/is the standard used here where I work but that decision was made before I came. The original sources of this material is all over the place; pal, ntsc etc.. and is saved on 10bit uncompressed digibetas then brought in to the library through blackmagic ntsc capture cards at 29.97fps. The photo jpegs are for DVDs and for archiving, as our web stuff is all h.264s at small resolutions that look pretty good. Making DVDs, we are realizing that people are usually watching them on computer monitors and we have a ton of interlacing problems whereby all our library media is 29.97 no matter what the source material. So I am testing methods to get rid of the interlacing without going through every cut of 2000 pieces of media in after effects...etc...And was hoping converting them back to h264, however lossy, might maintain enough quality while solving the interlace problems.
You are trying to mask a serious problem. You may get it to work but you will be walking a tight-rope. If someone just dumped every and all footage to NTSC, a lot of your PAL footage will have interlaced problems. Wouldn't want to be in your shoes, but I will say that if quality is a concern you should do it right, which means looking at each clip and fixing the problems. If not you have been dealing with unhappy with what they got.
This is such a "special" situtation that you will need to do testing to figure out what works. There is no established workflow for what you are dealing with and the more you explain it the more troubling it gets.