If you ever want to deliver in hi-def (Blu-ray, web), then you need to continue to shoot in HD. You can go from HD to SD with minimal loss of quality compared to shooting in SD, but you can't go to HD from SD without a huge and unacceptable loss of quality (IMHO).
An advantage in shooting HD is that you have that available, when you need to do a BD Project.
For HD to SD conversion, see this TUTORIAL.
If you are shooting HD tape, and your camera supports it, you might Capture in SD through the cameras settings, letting it down-rez for you.
Another advantage to shooting in HD for SD delivery is that you have extra pixels for post-production "zooms" (scaling up) and "pans" (just reframing and cropping).
David Van Brink shows an excellent example of this on his omino pixel blog.
Jeff, I have seen many of your threads and your tutorial about converting hd to sd with avi synth and virtual dub and you have provided a lot of useful information which has been helpful. Could you please answer one question for me?
In your experience, do you still recommend the use of virtual dub and avi synth in converting an original HD timeline (let's say actually HDV or DVCPRO HD original interlaced footage at 30 or 29.97 fps) to an SD 720x480 interlaced 29.97 fps NTSC output rather than just doing an export using AME with Premiere Pro CS5 or 5.5? (I have searched the forums for this and couldn't find the answer, so forgive me if I missed this.)
When using original Panasonic P2 media in the 1080i/1080p 60hz dvcprohd preset, I often had difficulty when converting to SD for local television broadcast, but in CS5 I had gotten decent results by using the "Field Options>flicker removal" on those troublesome clips. And in more difficult cases it was often sofved by exporting to an uncompressed AVI and then applying the "Field Options>flicker removal".
Sometimes I have also taken the footage from a DVCPRO HD sequence and cut and paste them into an NTSC sequence preset before output and I usually get a pretty good preview of problems to come when the footage scales to size and at this stage the Field Options>flicker removal works pretty well. (I usually scale the footage to have a letterboxed final output.)
I have to deal with a lot of camera pans and in-the-field jittery handheld work which makes it more difficult, and when I can and budgets permit, I will stabilize the footage as much as possible beforehand using AE, but I was curious to know in your experience, if you still recommend going the avi synth/virutal dub route with CS5 or CS5.5 versions for most HD to SD conversion.
I haven't tested the method using Dan Isacc's script. But I have tested a direct export using DVCPRO HD to a DVD, and I found nothing lacking in the quality that made me want to use Dan's script.
I start with AVCHD (Canon Vixia HFS100) and down-rez with CS5 to then create a DVD (no BD burner as yet)
Since the quality of the DVD is MUCH better than when I was digitizing 8mm tapes via Dv500, I will stick with just using CS5
The quality of Premiere's scaling and deinterlacing is improving with each version. I don't think Pr will ever get better than Dan's AviSynth filter, but it's possible that it's now just as good.
I continue to use hd2sd because I've got an effective, established workflow that I don't even have to think twice about any more. If I get some time where I have absolutely nothing else to do (not likely!), I'll revisit an all-Adobe workflow and compare it again to hd2sd.