Canon XH A1 Down Converting HDV to SD

Eric VS9 · Post by **Eric VS9** » Thu Jun 19, 2008 12:18 am

I have a new Canon XH A1 I have been shooting a few projects in HD
and I am a little confused. The HD gets captured at MPG 2 UF First 1440 X 1080 MPG audio
( in VS11 Plus or 10 there are no options to change any capture settings)

I then make my edits and down convert to SD. Now here is the confusion
I have allways be told if you capture in upper field or lower field then you render in the same field order

When I render the HD video to SD video in the UFF and play the DVD it looks like Crap ( I hope I can say that) So then I tried LFF it also looks like Crap. So then I tried no fields (Frame Based)and it looks good

So my question is what is the proper way to down convert from HD to SD, Have I firgured it out? Also it takes over 2 ½ hours to render HDV to SD is this normal ( I have Intel 2Quad 2.40 and 4 gig of Ram

Last thing I do think the video looks better shooting in HD and down converting to SD
rather then shooting in SD

Post by **Ken Berry** » Thu Jun 19, 2008 1:10 am

Your camera uses the HDV format and as such, its format is 1440 x 1080, not 1920 x 1080. So your captures are correct and you don't have to change capture settings.

I have a Canon HV20 which also uses HDV. I edit in its native UFF high def mpeg-2 format and, apart from producing high def output and converting to high def hybrid AVCHD discs, I also, for some people, downcovert to SD mpeg-2 for burning to standard definition DVDs. In VS11.5+, I do this by first converting my edited HDV into a single HDV file (Share > Create Video File > HDV). I do this anyway for use elsewhere. Then I clear the timeline, and insert that new HDV file in the timeline and downcovert that (Share > Create Video File) but here I choose Customise and set the bitrate to 8000 kbps (if it isn't that already) and make sure the Field Order is UFF. And they come out fine.

Rendering to Frame Based will look fine on a computer monitor. But probably would look not so good on standard TVs which must have interlaced video. They would probably appear OK on HDTVs which use progressive scanning, which is akin to the concept of Frame Based.

I too have a Quad like yours, and the conversion times are long, even with such a computer. You don't say how much you are converting, but if it take 2.5 hours to convert one hour, that seems excessive. It seems to take mine about 90 minutes for conversion of one hour, or thereabouts. And I have done no special tweaks of the computer.

Eric VS9 · Post by **Eric VS9** » Thu Jun 19, 2008 1:55 pm

Thanks for the reply Ken

When I play back a sample of HD to SD rendered UFF on a Standard TV is also Looks Good
The HD TVs are the problem HD to SD in UFF really looks Bad.

I have a 26 inch HD TV in my office for clients to view their projects on.
The only rendering that looks good on my HD TV is HD UFF to SD frame based

I don't know what I can be doing wrong with the HD to SD I can't give a
client anything other then Frame based because more and more people
are purchasing HD TV and I don't know what they have

Have you viewed your down conversions (UFF)on HD TVs

Eric

Post by **Ken Berry** » Thu Jun 19, 2008 11:17 pm

Yes, a 115 cm Samsung LCD HDTV which displays in 'true' HD resolution of 1920 x 1080p. Apart from the SD being obviously not HD when seen on a large HD screen, it is still pretty good, no artifacts, no distortions and nothing suggesting any problem with field order.

Eric VS9 · Post by **Eric VS9** » Sat Jun 21, 2008 7:27 pm

Thanks Ken

I've been away for a few days
so I haven't had a chance to reply back

I must say I am really confused I think for now I will
keep rendering from HD UFF to SD Frame based

I am shooting NTSC is there any difference in
down converting between PAL and NTSC HD to SD

Thanks Eric

Post by **Ken Berry** » Sat Jun 21, 2008 9:50 pm

None -- apart from the obvious speed difference (25 fps for PAL, 29.97 fps for NTSC) and frame size (720 x 576 for PAL and 720 x 480 for NTSC).

Eric VS9 · Post by **Eric VS9** » Sun Jun 22, 2008 12:37 am

Just thought of one other thing My Cannon is Shooting HDV 1080I (Interlaced)
is there a clue there I noticed your TV is P

I also subscribe to Videomaker Magazine and there is an article this month
on Progressive or Interlaced Shooting ( also available on line http://www.videomaker.com/article/13755/)
If I understand correctly 1080 I is not upper or lower field first
so maybe VS11 should capture it frame based I know that with AVI DV type 2 capture it is not UFF or LFF

This thing is driving me nuts
any thoughts

Post by **Ken Berry** » Sun Jun 22, 2008 2:22 am

My HDTV is indeed what is called a true high definition one as it displays in a resolution of 1920 x 1080p -- the p of course standing for progressive which is the top (?) form of HD. (As the article says, it also has its drawbacks.) But my TV is also capable of playing back interlaced (i) high def video as well as standard definition. So that is not the problem. My camera also shoots in 1440 x 1080i yet its video plays back magnificently on the TV...

To repeat, 1080i means interlaced, and interlaced in turn means it uses a field order -- in this case Upper Field First -- rather than progressive (p) scanning which is akin to frame-based. So I suspect you might have misread that article.

Your other misconception concerns DV. DV format is also interlaced, whether Type 1 or 2. And it is always Lower Field First.

Eric VS9 · Post by **Eric VS9** » Sun Jun 22, 2008 3:07 am

Ok so I am stumped
What would you suggest I can only go by what i am seeing from my down conversion
the only thing that looks good is the frame based (the UFF looks very bad)

do you think I should be giving clients UFF or Frame Based
Does CBR or VBR play into this in the conversion
The HD captured appears to be CBR and in the down conversion i have be doing VBR ??

I really do appreciate your help and experience.

Eric

Post by **Ken Berry** » Sun Jun 22, 2008 5:02 am

Well, all I can suggest is that you make two DVDs, one using interlaced video and the other using frame based. Then test those on stand-alone players (at friends or family members places) which are connected to standard TVs and high def ones. See what plays best and stick to that. Or else make stocks of both type... Sorry but otherwise I am out of ideas.

One thing, though. I am assuming you are down-converting to DV Type 2 (since you mentioned it). If so are you first converting that to DVD-compliant mpeg-2 while still in the Editor module of VS (Share > Create Video File > DVD)? Or do you use the DV directly in the burning module (Share > Create Disc > DVD) and allow the conversion to mpeg-2 to take place during the burning process. The latter is something we would not recommend... And at which step exactly do you change the field order to frame based?

As for CBR vs VBR, that shouldn't make much obvious difference in quality, though in some senses it depends on the subject matter of the video. It is a point a division between certain groups of users who prefer one or the other. But essentially, the argument in favour of VBR is that is represents a more sensible use of video space. With VBR, in high action sequences of your video, the bitrate speeds up to cope with the action. But in more tranquil shots, it can slow right down. Those who support CBR says it gives a constant and predictably good product. But of course it does so by using a higher bitrate than might be necessary in the quieter parts of the video.

So CBR vs VBR is really a question of the efficient use of video space vs perceptions of efficient. But in most cases you would normally be hard pressed to detect any visible difference in quality with the naked eye. And certainly not to the extent that you seem to be experiencing.

Eric VS9 · Post by **Eric VS9** » Sun Jun 22, 2008 2:12 pm

Hi Ken
I do the down converting in the time line
I do not go to DV I make my edits then create video file HDV 25000 UFF to dvd mpg2 7000-8000 100% VBR DD 192 ( I am on the VBR team)

I have tested UFF vs Frame on many friends and family (AKA as my R&D department)
if funny if I call first the don't answer the phone if the know I just got a new piece of equipment
or software so I have to sneak up on them.

so the bottom line is I will render in frame based because that's what looks best.
but I don't know why

Thanks Again