I wouldn't send the original, as the potential for damage in projection is too 

> Or would I be better off to send the rolls out to be scanned in 2K, cut it 
> together digitally, and send digital media files (how big is big enough to 
> look like film?).

That's the safest route. The projection of digital transfers never looks the 
same as 16mm projection, but on a good video projector (e.g. 3 chip DLP) it 
will look good. And digital scans of film always look "like film" — as opposed 
to looking like they were shot with any kind of video camera — regardless of 
the resolution of codec.

I would guess a good scan of the film is something you'll want to get anyway, 
as some venues just don't do 16mm projection anymore, and you'll want the film 
to have the widest range of screening options possible. Although I have not 
personally had any film scanned in this century, I would look for a lab that 
uses a Kinetta. Nobody knows the technical issues involved in scanning better 
than Jeff, or has a higher commitment to quality. I also have the distinct 
impression that the transfer houses that have acquired Kinettas are more 
indie-filmmaker-friendly than those that deal primarily with clients from 'The 
Industry'. I'm sure Jeff can recommend a service that would be the most 
accessible to whatever part of the world you might be located.

As far as sending media files goes, no-compromise quality would be a ProRes 422 
file, assuming the venue can play that back. ProRes files have a very high data 
rate, but for 300' the file wouldn't be so large as to make it impractical to 
transmit with a file sharing service. 

Smaller, more highly compressed files can still look excellent, depending on 
the compression software and setting used. H264 is pretty much the standard, 
and can be enclosed in either a Quicktime (.mov) or .mp4 container. All H264 
encoders are not created equal, and the free-to-use x264 is definitely the way 
to go. 

(To be clear, H264 is a codec format; x264 is one of several software encoders 
that can create H264 files. If you have x264 installed on a Mac, and do a 
conversion with Quicktime, the format selection drop-down menu will offer 
options for "H264" and "x264encoder" which makes them seem like different types 
in the same category, i.e. that x264 is a codec format. It's not. Choosing 
"H264" processes the source file using Apple's default H264 encoder, which is 
notoriously mediocre. Choosing "x264encoder" produces the same kind of thing — 
an H264 output file — but just uses a different means to do so, and one that 
produces superior results.)

You can download the X264 encoder separately, install it in the Quicktime 
Library, and use it within MPEG Streamclip, or other 3rd party file conversion 
front-ends. Or you can just install the cross-platform freeware program 
Handbrake, which has it built in. (Adjusting all the x264 settings in Handbrake 
is not exactly intuitive or straightforward though, but it's not rocket science 

There are tons of options within x264, most of which you can ignore as they're 
just for major techies. The key things are:

1. The setting for maximum data rate. This is generally used to force smaller 
files with higher compression. If you leave it blank, you should get the 
encoder's version of 'best' rate. I have never tried setting a higher max data 
rate to see if it yields a better-looking file than the "leave-it-blank" 
default, but I doubt it does.

2. The field that tells x264 the fps of the original. This defaults to 29.97, 
so you'll want to change it to 23.98.

3. The x264 quality presets. These are labeled by speed, e.g. "Medium" "Slow" 
"Slower" and "Very Slow". These mainly function to make the output file smaller 
while preserving quality, but they do so in part by having the software parse 
the original file more carefully, so they have some affect on the output file 
quality in general as well. "Slower" should be good enough.

4. Finally, there's an option for a Third Pass, which I would use in your 
situation. The faster quality presets are one pass, the slower ones two-pass. I 
don't know how much improvement comes from a third pass, but it can't hurt. 
Even with a short film, the encoding it processor-intensive enough that it's a 
start-it-and-thn-go-do-something-else-for-awhile kind of thing anyway, so 
adding the time for a 3rd pass shouldn't be a big deal.

Of course, once you've encoded the film in H264, you can play it back on your 
computer and compare it to the original output of the digital edit (which I'm 
assuming would be a ProRes file), to see what, if any, differences in quality 
you can spot. If it's done right, it's pretty amazing how good the H264 looks 
at a mere fraction of the size of the original file. You just say to yourself, 
"If the file has been shrunk that much, it just CAN'T be that good! But it is.

That said, if the ProRes file from the digital edit is a sendable size, and the 
venue can play it, I'd still send that, just to avoid having to take the time 
to set up and do the compression on a tight deadline. 

good luck,

FrameWorks mailing list

Reply via email to