Low budget feature workflow

[Justin Marx]

I'm shooting a feature in OCT with my EPIC and would love some input on workflow..

What is the best and most efficient way to:
Transcode for edit
Then CC
then out for mastering.

I'm confused with from the transcode to the CC, lets say I transcode with the look I set in RED CINE X, but they want to change it after the edit is complete.. What happens then? Re-cc the transcoded files or re-transcode from raw and re edit?

Imsure there is some eSy answer for this?

Thanks!!

[Kalani Prince]

wouldn't you want to go to something like Davinci's Resolve after the Red Cine stage?

Good question, I'm curious as well..

[Mark Toia]

Resolve is the best (cheapest) over all way to grade nicely from your R3D's from a EDl/XML from FCP.. then export directly to what ever...

simple. :))

[Jamie Metzger]

use adobe premiere.

[Tom Gough]

I'm shooting a feature in OCT with my EPIC and would love some input on workflow..

What is the best and most efficient way to:
Transcode for edit
Then CC
then out for mastering.

I'm confused with from the transcode to the CC, lets say I transcode with the look I set in RED CINE X, but they want to change it after the edit is complete.. What happens then? Re-cc the transcoded files or re-transcode from raw and re edit?

Imsure there is some eSy answer for this?

Thanks!!

Hi Justin,

There are many ways to skin this cat...

But a very useful way of approaching it is to use the "Caboose Effect". This is a term coined by none other than the maestro of DI and all things digital Michael Cioni of LightIron.

The concept is to work out your workflow by starting at the end of the chain - your deliverable!

I have added my comments in Red where appropriate.

"The CABOOSE EFFECT

1. How are you going to finish the film?

Base your process on how you are ultimately going to exhibit the film. This will reveal to you ways in which you should employ and/or avoid specific steps that may or may not be useful in distribution. For example, if you plan to distribute primarily on BluRay, there is no benefit to color correcting in P3 space for film exhibition.

2. How are you going to edit the film?

Most films are edited in Final Cut Pro and AVID Media Composer. I recommend working in either, but using DNxHD115 or ProRes HQ for your offline codec. If you can do this, you can easily work at full raster and a very high fidelity. But I do not recommend working with RED RAW files in offline, as that workflow can be cumbersome as the footage count goes up. If you are going to skip the offline process and work at online quality and grade off your transcodes use ProRess422, ProRes4444 or DNxHD175x.

3. How do you plan to color correct?

The nice thing about modern day DI is that everyone has access to it in some capacity. I prefer to work with data-centric systems for color mastering, but more importantly is a colorist. Do not color correct yourself. Even if you are working with the oldest coloring system around, it is probably better to take RED files to HDCAM tape and grade with an experienced colorist tape-to-tape than to use Apple Color and do it with an amateur. This really makes a HUGE difference-especially in an independent film. I cannot stress this enough. Talent in color trumps toolset any day of the week.

4. Test your gamma spaces

I recommend working with REDLogFILM for your gamma space to grade in. This may mean to process your dailies in REDLogFilm so you don’t have to do a reconform after you finish cutting. Otherwise, when you do reconform, use this gamma space since it is making RED footage look absolutely spectacular! A great combination is REDcolor2 and REDlogFILM this gives you maximum flexibility in the grade.

5. If you are doing on-set downloading, try to use 16GB CF cards

The pace of working with CF cards enables a downloader to keep up with the pace of production. This means there isn’t an hour of footage sitting on the camera at any one time waiting for downloading. Do it frequently in small chunks. This is the best rhythm for on-set processing. Obviously you will be using RED MAG SSDs with your Epic but Michael's point is still valid. If there is a DIT or DAM on-set, the producers don't want to see them sitting around most of the day and then start offloading at the end of the shoot. Obviously YMMV!

6. Shoot with 2 cameras

I find that productions that shoot with 2 cameras (especially indies) end up using the additional coverage to help solve performance issues. Giving an editor more choices is worth it-especially if you are running a tight budget. Sure the camera costs more, but the price of footage is so small with RED that a 2nd camera doesn’t increase the cost at the same rate as it does with tape and film. By doing your own dailies using a RED Rocket, you are easily improving the film by getting twice the dailies shot and more coverage."

http://michaelcioni.tumblr.com/post/1245348269/hi-michael-im-a-huge-fan-after-the-social-network-i

As I said at the start, there are many ways to skin a cat. This is not conclusive and the "post world" and DI processes are constantly evolving but the solid concepts that Michael has shared still stand.

If you want to learn more, I highly recommend REDucation. They have a 2 day Epic course at the end of this month for $750 which is an absolute steal!

http://epicreducation.eventbrite.com/

Hope this helps.

All the best,

[Tom Gough]

Here is one Low Budget feature workflow using your question as a prompt...


What is the best and most efficient way to:

Transcode for edit
Using the RED Rocket and REDcine-X, transcode to ProRes422/ProRes4444/DNxHD175x with REDcolor2 and REDlogFILM.

Then Edit
Using Avid or FCP...

Then CC
In your editor or export an XML, AAF or EDL to your weapon of choice (grading system) and grade off your transcoded material.

Then out for mastering
Rountrip back to your editor and playout to tape (HDCAMSR) or export to Prores4444. To go to a film print, take the HDCAMSR tape or files to the the filmout company that you choose and they will do the transfer from tape to film.

I'm confused with from the transcode to the CC, lets say I transcode with the look I set in RED CINE X, but they want to change it after the edit is complete.. What happens then? Re-cc the transcoded files or re-transcode from raw and re edit?
If you transcode with a look that you set in REDcine-X then that is more suited to an offline online post workflow where you edit offline transcodes and then online and grade off the R3Ds or re-transcoded neutral look media (REDcolor2, REDlogFILM with default camera metadata) to give the colorist the most options.

In your question you ask for the "best" way which nobody can truly answer as every production has different demands, restrictions and timelines. So the "best" way for one production is not necessarily the best for another.

Anyway some food for thought. :)


All the best,

[Justin Marx]

Here is one Low Budget feature workflow using your question as a prompt...


What is the best and most efficient way to:

Transcode for edit
Using the RED Rocket and REDcine-X, transcode to ProRes422/ProRes4444/DNxHD175x with REDcolor2 and REDlogFILM.

Then Edit
Using Avid or FCP...

Then CC
In your editor or export an XML, AAF or EDL to your weapon of choice (grading system) and grade off your transcoded material.

Then out for mastering
Rountrip back to your editor and playout to tape (HDCAMSR) or export to Prores4444. To go to a film print, take the HDCAMSR tape or files to the the filmout company that you choose and they will do the transfer from tape to film.

I'm confused with from the transcode to the CC, lets say I transcode with the look I set in RED CINE X, but they want to change it after the edit is complete.. What happens then? Re-cc the transcoded files or re-transcode from raw and re edit?
If you transcode with a look that you set in REDcine-X then that is more suited to an offline online post workflow where you edit offline transcodes and then online and grade off the R3Ds or re-transcoded neutral look media (REDcolor2, REDlogFILM with default camera metadata) to give the colorist the most options.

In your question you ask for the "best" way which nobody can truly answer as every production has different demands, restrictions and timelines. So the "best" way for one production is not necessarily the best for another.

Anyway some food for thought. :)


All the best,

Tom,

THANK YOU SO MUCH!! Your posts were exactly what I was looking for..
I think the REDucation sounds like a good idea as well!
One last question: how do you online and grade off of the R3D files?

Thank you so much again!!

[Gino Papineau]

Stick with R3D as long as you can, at least for the color correction process...

[Tom Gough]

Tom,

THANK YOU SO MUCH!! Your posts were exactly what I was looking for..
I think the REDucation sounds like a good idea as well!
One last question: how do you online and grade off of the R3D files?

Thank you so much again!!

Hi Justin,

You're very welcome :)

With regard to the online and grading with R3Ds, that entirely depends on what system you use for the finishing process and whether or not it supports working with R3Ds natively.

Another factor is budget and timeline. You can go the fast and expensive route or you can go the slow but cheap route to finish your project - there is always a trade off!

Mark Toia's recommendation on using Resolve is well founded and a great option for working with R3Ds and roundtripping with FCP and Avid - but there are many other options but all depend on budget.

There are other more qualified people (Michael Cioni, Mike Most, Jake Blackstone, Bruce Allen, Marc Wielage) in this forum that can go into futher detail on the pros and cons of various finishing systems but I have to say that I would heed Michael Cioni's wisdom in "The Caboose Effect":

"...Do not color correct yourself... it is probably better to take RED files to HDCAM tape and grade with an experienced colorist tape-to-tape than to use Apple Color and do it with an amateur. This really makes a HUGE difference-especially in an independent film. I cannot stress this enough. Talent in color trumps toolset any day of the week."

Finally TEST, TEST and TEST.

Test every part of the workflow from on-set through to post. Create test footage, edits, roundtrips, outputs... Make sure everyone in the chain from DIT, Post Supervisor, Assistant Editor, Editor, Online Editor and/or Colorist are on the same page. Through testing various workflows and systems you will find the right one for your production.

I can't recommend REDucation enough, getting information from the RED coal face really does fill in a lot of gaps regarding RED in production and in post - it did for me :)


All the best,

[Michael Cioni]

Tom,

One last question: how do you online and grade off of the R3D files?



Hi Justin-
My advice to you is not to concentrate on doing an online. Take things in small strides, go slow, learn from each experience and move up from project to project.
Embarking on what you're doing is going to be enough of a learning curve on its own. For this job, Tom is right in suggesting just to transcode once to ProRes HQ or DNxHD175. Make that your negative. For some of my most successful über-indies, I encourage them not to do any color correction at all in dailies. Transfer REDColor2 / REDLogFilm and let that be the cutting material. With an independent, you can explain how the log-like-look is paying it forward to a cheaper, faster and better finish.
If you do this, you're literally working at a higher quality level than most people are able to achieve with the SONY F35...so it literally pays to work in 1080 or 2K REDLogFilm files and eliminate the reconform altogether.

Once you are done, simply export a self-containted textless ProRes444 or DNxHD175x file from your cut separated into 5 or 6 reels that are 15-20 minutes in length each. Then, depending on who you choose to work with, you can load that media in as a pre-conformed sequence and allow color to be done with a real professional. As I said before, I cannot stress enough how important it is to have a very experienced colorist in your corner and working with a true infrastructure. On our films, we use the same talent on big jobs as we do small jobs - they all present their own unique challenges and everyone can learn something new on every job. But what we don't do on every job is run everyone through the same cut-and-paste pipeline. So should be the case with your job.

Filmmaking is an unending series of compromises, but the trick is to be educated enough to know which ones have little to know consequence.

R3D + RED ROCKET + REDLogFilm + ProRes 444 = ideal backbone for an independent film

This formula, if followed correctly, represents unmatched simplicity when compared to all the other noise out there - RED or non RED. This formula makes a few tiny sacrifices while protecting the only thing that matters: what you ultimately put on the screen. Ensure that every call you make and every dollar you spend is headed to the final film. Many components of indies are one-way streets; or processes that end up getting you to a checkpoint, only to be re-done for the final piece later (like offline and online or temp vfx or temp conform for test screenings, etc.)

Workflow is one of my least favorite words. Mainly because the more money you have the less people depend on how the work flows. When money is more available, money can control the flow of work. But when money is less available, how well the work flows can change the amount of money you spend.
m

[M Most]

I agree with what Michael said. Simplicity is your friend. Quality isn't really sacrificed if the intent is to deliver HD video anyway, and color flexibility isn't really sacrificed if you use the correct transcoding parameters.

I would add that if the light, flat look of the RedlogFilm files is really annoying, and you happen to be editing on a system that includes a client monitor, you could easily use something like an HDLink Pro to insert a Log to Video LUT in the viewing path and see a more "normalized" image on the monitor. That would allow for more director and producer friendly screenings in the cutting room. Not necessary, but potentially useful and definitely affordable.

[Subhadip Sen]

The easiest possible workflow is get your look "in camera", drag and drop to Premiere Pro. That's about it, really. Minor tweaks with Source Settings menu of course, which has everything you find in RedCine-X, including curves. The keyword here being "easiest". There have been plenty of absolutely gorgeous films in the last century which had no color timing whatsoever. The idea is to keep it simple. With Sandy Bridge and CUDA today's mainstream hardware (with the right software) has got to a point that transcoding is simply a waste of time and space. 2K ProRes or 4K R3Ds played back at 2K - both give you real-time performance without a hitch on a PC that costs about the same as the software running it (i.e. ~$1000). It's a no-brainer, personally. By using a native workflow your transcode occur at the end of editing - thereby eliminating an entire step in the process. Of course, only Vegas Pro or Premiere Pro offer this kind of performance. Any other NLE and the transcoding step becomes mandatory, whether online or offline.

[M Most]

There have been plenty of absolutely gorgeous films in the last century which had no color timing whatsoever.

Every film ever made and printed in a photochemical process had color timing by definition. Somebody had to come up with the printer lights, and in almost all cases, those lights were changed on a scene by scene basis. Color timing didn't suddenly start with digital intermediates, it was done in a different way prior to that. How extensive the timing was, and how many light changes were incorporated differed, but every picture had lights assigned, and that is what color timing is.

[Subhadip Sen]

Every film ever made and printed in a photochemical process had color timing by definition. Somebody had to come up with the printer lights, and in almost all cases, those lights were changed on a scene by scene basis. Color timing didn't suddenly start with digital intermediates, it was done in a different way prior to that. How extensive the timing was, and how many light changes were incorporated differed, but every picture had lights assigned, and that is what color timing is.

Just to be clear, I was referring to changes in timing on a shot-by-shot (or scene-by-scene) basis. Obviously, without light there is no print. In an interview Hsiao-Hsien Hou talked about how the technology (or talent) for precise timing simply did not exist even in late-90s Taiwan. Yet the Taiwanese New Wave produced the most beautiful pictures, in my opinion. For many early TNW films it was a simple matter of sending the negatives to the labs and accepting whatever came back.

Edit - Now that I think of it, I think by "beautiful" I am referring to the composition or mise-en-scene more than anything else. In terms of the general "grading", they all looked kind of similar. But pleasant. Colour was used extensively, but most of it came from production design, filters and gels. So, a lot of possible with only basic color correction.

Another extremely colourful film would be Double Life of Veronique. A high profile film by all means, and apparently Slawomir Idziak lost all control of the photochemical process. It was a total surprise to him when in the first screening the film appeared with the now famous green tint. It's an interesting story - it's there in the Criterion extras on the Blu-ray for the film.

[Michael Cioni]

With Sandy Bridge and CUDA today's mainstream hardware (with the right software) has got to a point that transcoding is simply a waste of time and space.


Every film is different, every film has it's own unique set of limitations and every film has its own personalities. While I recognize the above statement may be applicable in circles, I want to make perfectly clear to any user of file-based technology that transcoding is not a waste of time or waste of space. Unsure if the writer realized this or not, but (perhaps ironically), transcoding is actually a saver of time and a saver of space:-)

In fact, for some circles, transcoding is not just the obvious option, it's the only option we have. For example, 3 projects I have in motion right now are each shooting an average of 1.2 terabytes per day across A, B, C and D cameras, which will total about 75-125 terabytes of source dailies. Coupling that significant of a data footprint with the fact that 50% of these films are visual effects means that if you wanted to forego transcoding you would have to:

* have 200 terabytes of networked storage hooked up to a AVID, Final Cut Pro, Premiere or Vegas Pro
* have enough CPU horsepower to multi-cam 2-4 cameras together for playback of 2-4 streams of 4K or 5K files
**have enough GPU horsepower to debayer hundreds of exports for previews, previs or visfx
•*have enough thruput to handle 2-4 streams of 4K = 350MB/s should do the trick
* have enough horsepower to get 4K files back form vfx and cut into the timeline for continued editing at full res

Most people know I'm a big fan of pushing technology, but even the most forward-thinking people on this board know that the above work will not flow.

Subhadip, I have no intention of picking on you. I trust it doesn't come off that way. I want to point out that Subhandip's point of avoiding transcoding does, in fact, have merit to it. But my intention is to make perfectly clear to everyone that transcoding does not make you any less of a filmmaker or file-based expert. It is not a waste of time or space and there are massive advantages to the transcode process that may improve the workflow for some and simply be an annoyance to others.

m

[Subhadip Sen]

Subhadip, I have no intention of picking on you. I trust it doesn't come off that way. I want to point out that Subhandip's point of avoiding transcoding does, in fact, have merit to it. But my intention is to make perfectly clear to everyone that transcoding does not make you any less of a filmmaker or file-based expert. It is not a waste of time or space and there are massive advantages to the transcode process that may improve the workflow for some and simply be an annoyance to others.

m

Absolutely, I completely agree that the transcode process has immense benefits to many workflows, no doubt about that. In the above situation you mentioned, with extensive VFX and multiple cameras, transcoding is a necessity. However, the said situation doesn't really fit with what we are discussing here - "low budget feature workflow". I was just speaking for the easiest possible system, i.e. native editing in Premiere Pro. In this case, what we are really looking for is the simplest way to work. Of course, "low budget" itself is open to interpretation, but I would think a small team, a single camera shoot, mostly 1 stream of R3D, no VFX, no complicated color grading etc. And I do believe in such a situation natively cutting R3Ds is simply the easiest and most fool-proof solution. We can discuss more complicated situations such as the one you mention in another thread. Hope that clarifies.

[Marc Wielage]

I'm shooting a feature in OCT with my EPIC and would love some input on workflow.. What is the best and most efficient way to: Transcode for edit... Then CC... then out for mastering.
I agree with The Two Mikes (Most and Cioni) above. And to add my own two cents: there's no harm in setting a temporary look in Redcine-X, then doing a rough cut with compressed files just to finalize the edit. Once you know who's going to distribute the project -- theatrical to film-out? festivals via D-Cinema? home video? pay TV? foreign? -- then make a decision on conforming and final color based on your budget and the eventual destination. The number and type of deliverables is crucial.

Like Michael says: figure out the caboose, then determine how the train's gotta run. You haven't laid down the tracks yet -- don't get ahead of yourself. And don't forget, there's a lot of steps ahead of conforming and color correction, like the mix. I know of some indie films that spent years in post. (Even some major ones, actually.) And I know of too many projects that got close to the finish line and then realized they had grossly underestimated all the remaining steps, like pan/scan, standard-def, TV versions, QC, special sound mixes... all the minutiae that drives post people crazy. And some of those are absolutely critical, because they're line-items tied to a contract (as in money).

Going through the Reducation post class is a wise move, too. At least it'll make you aware of all the various processes available. There's a dozen different ways to get to the finish line, and quite a few of them work pretty well. Some differ mostly in time and money, others in terms of technology. But they all require attention to detail and moving one step at a time -- carefully.

[John Bannister]

Tom, your posts in this thread have been invaluable. Thank you for sharing you knowledge it's bookmarked for my next project. Cheers!

[Tom Gough]

You're very welcome John.

All the best,