Any verdict in producing SD from HD w/ FCP?

[L.J. Morelli]

A few months back this was a hot and heavy topic for me. I've done a bunch of shoots, all of which are coming to completion in editing. I tried it this way:

Outputting Standard Definition in FCP from the Sony XDCam EX1

I couldn't complain about results, seemed fine. Close friend of mine who is extremely knowledgeable and runs a post house insists I use Compressor, then bring that file into DVD Studio Pro. Recently did one that way too. Also seemed fine. I didn't do a side by side comparison.

I'd like to hear from people what they've been doing. Only those using FCP and DVDsp.
Thanks, John.

[Matt Davis]

Originally Posted by L.J. Morelli

I'd like to hear from people what they've been doing.

Interesting that you link to that article...

I spent some time playing with ways to do the SD convert in FCP, then found (much to my eternal relief and joy) that I should work entirely in XDCAM (I like the 720p mode which is almost as fast as DV), export it a a self contained QT movie (very important), then just dump it on a Compressor droplet to make a HQ SD DVD using the 120 minute preset. I habitually use 120 for even short stuff because the 90 min setting can just tip the balance of lesser DVD players and older DVD enabled laptops.

It's worked excellently for all my EX1 work.

[Paul Frederick]

This recent topic discusses it in detail.

http://www.dvinfo.net/conf/sony-xdca...d-footage.html

[Craig Seeman]

Great if your target is DVD but what about other forms of delivery?

I believe the Vortex training DVD mentions editing HD in SD timeline. I do local cable spots and delver standard def MPEG-2 Program Streams.

I create an SD timeline (still toying with codec choice between AppleProRes, 8 bit uncompressed, DVCPro50). This gives me the ability to reposition shots of course as well as to chose between center cut or 16:9. I then export Quicktime Self Contained "master" that I can use for various forms of SD delivery.

There's PLENTY of downsides to that workflow which includes, no HD edit master, having to chose between 16:9 or center cut to name two.

I should experiment with HD edit and then doing downcovert in Compressor (or Episode?) targeted delivery type needed.

There's also some big title/graphic decisions to make also between the 16:9 HD vs letterboxed 16:9 on 4:3 screen vs center cut.

Since you're from the UK I think these days most of you SD tv sets are 16:9 which is not the case here in the USA.

Another issue I'm facing is that while even local channels in my area are moving to HD, they all still require SD spot delivery (Grrrr!) so I have to think about how my HD to SD downconvert looks on an HDTV. How backwards is that!

Originally Posted by Matt Daviss

Interesting that you link to that article...

I spent some time playing with ways to do the SD convert in FCP, then found (much to my eternal relief and joy) that I should work entirely in XDCAM (I like the 720p mode which is almost as fast as DV), export it a a self contained QT movie (very important), then just dump it on a Compressor droplet to make a HQ SD DVD using the 120 minute preset. I habitually use 120 for even short stuff because the 90 min setting can just tip the balance of lesser DVD players and older DVD enabled laptops.

It's worked excellently for all my EX1 work.

[Matt Davis]

Originally Posted by Craig Seeman

Great if your target is DVD but what about other forms of delivery? I should experiment with HD edit and then doing downcovert in Compressor (or Episode?) targeted delivery type needed.

Compressor isn't just for MPEG2! I've got presets for standards conversion, the ever important 'make this 720p into PAL DV at 16:9' and ditto 'to DV 4x3'. The Ripple Training Compressor 3 training stuff is very good at demystifying what's under the hood of Compressor.

I'm firmly in the camp of letting FCP edit, and Compressor transcode.

Originally Posted by Craig Seeman

There's also some big title/graphic decisions to make also between the 16:9 HD vs letterboxed 16:9 on 4:3 screen vs center cut. <snip> I have to think about how my HD to SD downconvert looks on an HDTV. How backwards is that!

UK terrestrial is still officially 4:3, so the BBC has adopted a sort of faux wide screen - 14:9, which means you lose a little either side, plus some other odd title safe wrinkles for 4:3 crop circumstances. 16:9 and all its modes is still a bit of a mess at this side of the pond.

So FCP users may like to add this to their toolbox:

Digital Heaven - The Hottest Tools for the Coolest Apps

Keeps me sane.

[Alister Chapman]

I edit HD in FCP and then export using compressor. I'm perfectly happy with the results I'm getting.

[Sverker Hahn]

Originally Posted by Matt Daviss

... export it a a self contained QT movie (very important) ...

Why is self contained important?

[Matt Davis]

Originally Posted by Sverker Hahn

Why is self contained important?

1) Stops any probems with sync
2) Keeps FCP and the system active (whereas sending from FCP to Comp does not) so you can continue doing other things
3) Creates 1 master file of final programme, therefore no need for plug-ins and treats in te future

What you don't get:
1) All effects re-rendered frame by frame within a 4:4:4 colour space (if you really can spot it)

[Dominik Seibold]

My experience is, that Compressor has some kind of a bug which prevents it from doing this correctly:
interlaced-input -> rescale -> interlaced-output
(which completely is: interlaced-input -> deinterlace -> rescale -> interlace -> interlaced-output)
All other combinations of interlaced/progressive-input/output are working fine.
if you do need interlaced->rescale->interlaced, then you can use this 2-step workaround:
1. interlaced-input -> deinterlace -> progressive-output
2. progressive-input -> rescale -> interlaced-output

[Jon Braeley]

As far as I know, over many years of working from FCP to Compressor, there is nothing to be gained by exporting a self-contained QT movie if you are dumping this into Compressor. In fact you are creating a massive size file and spending way too much time making FCP do all the work when it should be Compressor.
Export a reference QT movie - much faster, still plays even in QT player but with a small file size. Then import into Compressor for SD delivery.

I have tried both ways and see no difference in quality but a big difference in time and file size.

[Sverker Hahn]

Originally Posted by Jon Braeley

Export a reference QT movie - much faster, still plays even in QT player but with a small file size. Then import into Compressor for SD delivery.

I have tried both ways and see no difference in quality but a big difference in time and file size.

This is what I have done lately with good results. Of course I have to render all sequences but that always takes it´s time. But exporting to a reference file is a matter of seconds. The Compressor can do the transcoding and during that time I can work in FCP.

If there are no quality issues I will continue with reference files most of the time - it saves time and space

[Matt Davis]

Originally Posted by Jon Braeley

I have tried both ways and see no difference in quality but a big difference in time and file size.

My reason is twofold:

1: Backup - I value having a full quality, self contained file I can back up separately.

2: Outputting to Compressor from FCP ties up the whole machine, and locks out FCP for the duration. Outputting a file is quicker, leaving the machine free to do FCP rough chops, logging, and the rest of the machine to invoicing (love it), research, writing, whatever whilst I compress in the background.

FWIW, there is some reference into the ins and outs of either method in Philip Hodgett's guides, which may have a bearing on specialised technical requirements and use of 8 bit over 10 bit, but essentially it's a workflow issue for me. And I ain't giving up my backups!

[Jon Braeley]

Matt, only outputting directly to Compressor from the FCP timeline will tie up Final Cut. We are not advocating this. Also, a self contained movie ties up FCP also.
From FCP you export a reference file, not a self-contained movie which will tie up Final cut for some time. A reference file is much faster.
When imported into Compressor this file works in the background. Almost all Apple techs and production people I know use this method.
As long as your source files and Final Cut project is backed up, I see no reason to have a full size self-contained QT movie. I would never have a use for such a large beast.

[Craig Seeman]

There are a couple of gotchas with using reference files.

Unless you name it, there's nothing inherently in the file name or icon to let you know it's a reference file. At some later date, when sources are offline, you click on what you think is a self contained file and . . . surprise, failed searches for source elements. You're then faced with reloading the sources and exporting again. BTW when I use reference files I usually put REF in the title but even I forget sometimes.

Often when dealing with codecs like EX one may set render to AppleProRes while timeline codec is EX. The reference file may well be similar to self contained in size so you haven't really gained anything in smaller size and you can still end up with the situation above.

Sometimes there are issues finding sources or renders. Using a reference file has that risk. It's small but it's real annoying when that happens. Not good when you're trying to save time (although not a common occurrence).

Having a self contained file gives you a master you can back up for safe keeping.

______________
so when do I use reference files?

Source and render codec match (which is not the case when editing EX codec with ProRes render)

I'm only making an interim encode with no intention of saving the FCP export.

I'm really short of hard drive space and the source and render codec match (otherwise the reference can be as big as self contained).

So I really don't find reference files all that useful when working with EX.

[Christoph Gelfand]

I am attempting to down-convert HD to SD and wondered if I should leave titles in my HD sequence when I drop it in my SD sequence, or add titles in after I've already converted to SD. Any thoughts?