HDV Workflow -> Upsampling/Deartifacting/Deinterlacing

[Rafael Metz]

Hi All!

I´m just trying to figure out the best workflow for HDV (on Final Cut Pro).
The question is which steps to go to have the best possible image quality out of HDV. I´ve read some stuff yet but I´m very interested to have a good way and the right tools thrown together here.

The final result should be a 25p/30p/24p film.

Should we shoot (e.g. with FX1/Z1) in 1020i and then deinterlace it (which is the best deinterlacer?)? When should deinterlacing be done in the workflow?
Or should we shoot in progressive mode?

How can we raise color quality with upsampling?

How can we lower the compression artifacts?

What else can be done to improve picture quality? (from a technical point of view)

Best way of digitizing? Which compressor would work best? How about thru SDI?

How about HDVxDV? Is it still useful with FCP5? And if why?

??? More ideas???

[Lewis Lehman]

Originally Posted by Rafael Metz

Hi All!

I´m just trying to figure out the best workflow for HDV (on Final Cut Pro).
The question is which steps to go to have the best possible image quality out of HDV. I´ve read some stuff yet but I´m very interested to have a good way and the right tools thrown together here.

The final result should be a 25p/30p/24p film.

Should we shoot (e.g. with FX1/Z1) in 1020i and then deinterlace it (which is the best deinterlacer?)? When should deinterlacing be done in the workflow?
Or should we shoot in progressive mode?

How can we raise color quality with upsampling?

How can we lower the compression artifacts?

What else can be done to improve picture quality? (from a technical point of view)

Best way of digitizing? Which compressor would work best? How about thru SDI?

How about HDVxDV? Is it still useful with FCP5? And if why?

??? More ideas???

lots of questions... lots of answers...
these are mine...

1.de-interlace, yes. If going to DVD
2.de-interlacer? I use natress. from g., Look at joe too. boris etc. $$$ you
get what you pay for.
3. max pic info in 60i 1080
4. CF30 interesting, be careful.
5. start w. 4:2:0 @1440x 1080 then what?
6. artifacts. articfacts. artifacts. Compressed video... be careful...
7. compression...Compressor, bitvise etc...Knowledge and practice will save you money.
8. SDI= analog, HDV = Digital, you decide.

To all the Gurus...
Will HDDVD or Blue Ray solve my HDV problems? Yes and No I think...

Lewis

[Rafael Metz]

Originally Posted by Lewis Lehman

lots of questions... lots of answers...
these are mine...

1.de-interlace, yes. If going to DVD
2.de-interlacer? I use natress. from g., Look at joe too. boris etc. $$$ you
get what you pay for.

What´s the point for de-interlacing when going to DVD?
I use deinterlacing to have a more film like look. The plug in I use is Re:Vision Fields Kit.

3. max pic info in 60i 1080

For sure. So is it the better way to shoot interlaced and de-interlace it to have the resolution?!

5. start w. 4:2:0 @1440x 1080 then what?

Going to 4:2:2.

6. artifacts. articfacts. artifacts. Compressed video... be careful...

???

8. SDI= analog, HDV = Digital, you decide.

SDI is not analog and offers 4:2:2 upsampling from 4:2:0.

[Dylan Pank]

Originally Posted by Rafael Metz

Hi All!

I´m just trying to figure out the best workflow for HDV (on Final Cut Pro).
The question is which steps to go to have the best possible image quality out of HDV. I´ve read some stuff yet but I´m very interested to have a good way and the right tools thrown together here.

The final result should be a 25p/30p/24p film.

Erm, like, pick one. 24p is convertable to 25p (just speed it up 4%) but 30p is very difficult to satisfactorily transfer to 25p

When you say "25/30/24p film" do you mean a film out to 35mm for a theatrical release or do you mean just that you want a progressive scan video master for DVD and Broadcast?

Should we shoot (e.g. with FX1/Z1) in 1080i and then deinterlace it (which is the best deinterlacer?)?

Magic Bullet has the best reputation, but the Nattress filter also comes very highly recommended (and at a MUCH lower price.) You should test the different approaches and see which you prefer.

When should deinterlacing be done in the workflow?

Tricky. Do it at the start and you'll have to re-encode all your footage (which, unless you're going to an uncompressed format, means a varying degree of quality loss, though it may be minimal). Do it at the end and you save yourself disc space/rendering time, but you won't really know how the deinterlacing looks on your footage till the final render out.

It depends how much drive space/time/patience you have and how picky you'll be on the effect de-interlacing has on your footage.

Or should we shoot in progressive mode?

Depends on your camera. At the moment you can either get a 720p progressive scan camera (the JVC cams) or a 1080i Interlaced camera (the Sony cams). Until the (non HDV) Panasonic HVX200 arrives there are no consumer/prosumer cameras offering interlaced and progressive.

The Sonys have the infamous CineFrame 24 mode which some like and some hate, but I believe is tricky to deal with in post and has a negative impact on picture quality. There is CF25 on the 50i/PAL cams and the HVR Z1/A1 line which can be converted to 24fps by slowing DOWN 4%. The 60i/NTSC cams have CF30 which, as I said can't easily be converted to 24/25fps.

With the Sony's, in general the consensus seems to be that the best quality comes from shooting interlaced (at 50i or 60i) and then de-interlacing with the best software available to you, rather than shooting with the CF modes.

How can we raise color quality with upsampling?

You can't. Only with colour correction tools/effects filters. You can upsample to a 4:2:2 codec (DVCproHD) or even a 4:4:4 codec, but the colour quality will stay the same till you start to affect it. It's probably advisable to colour correct at a better colour sampling rate (I'm not a CC guru) but upsampling itself cannot improve picture quality.

How can we lower the compression artifacts?

Careful shooting; not too many random, fast changing images; keep the shutter speed at 60th/sec. Some have suggested that light diffusion in the lense may help too, though this would of course reduce the image sharpness/definition, but many shots in 35mm and HD are diffused anyway, especially close ups (see all those "why Britney hates HD" articles around the web.)

It seems likely that shooting in CF25/CF30 would also reduce artifacting a little, as uninterlaced footage at a lower resolution should compress easier. I've not personally tested this yet. For SD distribution this might be a way, but when distributed at HD res, the lower resolution of the CF mode might be apparent, depending on the nature of the images.

It seems that the JVC HD100 footage (24/25/30p, 1280*720)has fewer artefacts under similar conditions than the Sonys (60i, 1440*1080) but as of yet this is based on comparible sort of similar footage. No-one has yet done a side-by-side test

What else can be done to improve picture quality? (from a technical point of view)

Good shooting/lighting really, try and get colour balance as you like it in camera rather than relying too much on post. Too much agressive colour correction will start to show up the artifacts in HDV footage.

Shoot tests, see what works and what doesn't.

Best way of digitizing? Which compressor would work best? How about thru SDI?

Again, depends on your needs. Some people recommend native as it leaves quality at it's highest, and you should need to do only one output pass at the end.

Others prefer to edit with an i-frame intermediate codec (in FCP, DVCproHD) as it makes editing easier.

You can digitise over firewire and do a software convert, taking time and HDD space, or over SDI which takes money: you need an component-SDI converter and a suitable capture card. Some SDI cards (Kona, BlackMagic) will digitised directly to DVHproHD, though you'll have all the compression artefacts of the HDV footage, plus any that the DVCproHD codec introduces (they'll be pretty imperceptable though)

How about HDVxDV? Is it still useful with FCP5? And if why?

As far as I can tell, HDVxDV is redundant with FCP5 as everything that HDVxDV does can be done internally with FCP's media management tools, unless HDVxDV offers better quality conversions between codecs.

??? More ideas???