Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

WHY!!!! are people still using color baked DPX's F&%*@#%*#K!

Mark, and other this may be a goo place to ask this. For someone who has yet to create their Red work flow, what would you guys suggest for a documentary series. Keep in mind I'm talking about a LOT OF DATA! My next series is 6 episode, so I'm looking ay roughly 70hrs of footage. In the work flow I work with, I've found it's best to transcode everything. Like MMost said, playback performance is key to me being able to crank through many episodes quickly. From what I hear AMA may not be the best for me, but would love to hear what people have to say.

All stuff will be shot over seas, and graded here. I shoot an edit. I'll use Avid and Resolve lite. I'd like to convert to DNX to edit, conform, but then send my stuff to resolve where I can then source my r3d for the final grade. Is this possible??

Te more people that start off with the right workflow, the better. So help us out.
 
>>> I'd like to convert to DNX to edit, conform, but then send my stuff to resolve where I can then source my r3d for the final grade. <<<

and that is exactly how you do it. just resist the urge to rename the video files on the harddrive :-)
 
Sorry Mark...

When I was doing software design and the user did the wrong thing... you don't blame the user. You blame the software you're designing for leading the user astray.

So looking at it from that perspective:

1. Where is the nice PDF on RED's website that explains the right way to do things?

2. Designing a codec that requires post houses to buy a separate hardware board is very... 1980s. Imagine if every manufacturer did it. I'd have my computer filled up with cards - SonyROCKET, ArriROCKET, PanaROCKET. WTF!!!! R3Ds should decode via GPU. That's RED's fault.

3. With Epic, you should be able to monitor with one color setting... but record metadata that tells the post software to develop using different settings (RedLogFilm / RedColor2). Arri Alexa makes it very easy to do. If RED made it similarly easy, we wouldn't have this problem.

4. The problem with RedLogFilm is that it wasn't there from day one - and it has not been properly publicized by RED. Lots of people don't know it exists. Again, where is the explanation on RED's website? Alexa had LogC from day one, with a clear explanation, so people know to use it.

5. The total lack of documentation / downloadable LUTs makes things difficult to understand.

6. DPXes are not inherently bad. DPXes with RedLogFilm are great in fact. The problem is that RED is pushing a native R3D decode solution... which just doesn't work once you start adding VFX. Instead they should push a clear, simple, DPX-based solution.

Bruce Allen
www.boacinema.com
 
Ummmm.... not often i would disagree with Mike, but as my bread is buttered by the ablity to work with r3d's in the timeline while i grade... i have to say - there is options worth considering.

i think i should mention that i do NOT do tv episodic.. i'm wayyyyyy to slow for that pipeline, and i agree with Mike that working from r3d's is slower than DPX, but not every show needs to be squeezed inside the DPX pipeline

I do work on indie features, and i am getting jobs that the DP's have insisted the show be pulled from faclilities much larger than i am due to working with r3d's directly... three shows in the last three months.

I get a lot of mileage from setting debayering choices in context, get to work with great folks, and great images

d
 
Sorry Mark...

When I was doing software design and the user did the wrong thing... you don't blame the user. You blame the software you're designing for leading the user astray.

So looking at it from that perspective:

1. Where is the nice PDF on RED's website that explains the right way to do things?

2. Designing a codec that requires post houses to buy a separate hardware board is very... 1980s. Imagine if every manufacturer did it. I'd have my computer filled up with cards - SonyROCKET, ArriROCKET, PanaROCKET. WTF!!!! R3Ds should decode via GPU. That's RED's fault.

3. With Epic, you should be able to monitor with one color setting... but record metadata that tells the post software to develop using different settings (RedLogFilm / RedColor2). Arri Alexa makes it very easy to do. If RED made it similarly easy, we wouldn't have this problem.

4. The problem with RedLogFilm is that it wasn't there from day one - and it has not been properly publicized by RED. Lots of people don't know it exists. Again, where is the explanation on RED's website? Alexa had LogC from day one, with a clear explanation, so people know to use it.

5. The total lack of documentation / downloadable LUTs makes things difficult to understand.

6. DPXes are not inherently bad. DPXes with RedLogFilm are great in fact. The problem is that RED is pushing a native R3D decode solution... which just doesn't work once you start adding VFX. Instead they should push a clear, simple, DPX-based solution.

Bruce Allen
www.boacinema.com

We run on Flame and most other composting / grading tools and I have to say that you are quite wrong. The Red rocket is only needed if you are on slow/small computers todays hardware can manage R3D with no problem. The card is actually the cheapest solution available to play 4k RAW with realtime grading. (Possible to do on your mac air laptop) But if you do serious work your you should have a big enough workstation to be able to chew the 5K files without to much difficulty, and at least playback at half resolution or such. When using DPX files there is no such option. So Red in this case is making it easier for the little guys, since you do not need the big machines even. The problem is all the old hardware used out in the major post houses are not up to spec while 4k and above will make demand upgrades no doubt.

And for DI work the post side needs to learn a lot more complex things than handling r3d files which is more or less a plug and play thing to do. As for flame, nuke etc you are very much able to just open the r3d file select. camera setting, red user setting or, user setting. If the last option is chosen you then go trough and sellect film gamma film gamma and set your iso, kelvin and so on. Not so tricky. But if you do not know you will probably do it wrong. Read your manuals, do your test etc. But off course things will go wrong when people do not know what they are doing.

As a fact knowing what you are doing will be the only way to make a living in the post industry as things will evolve. Since the time of great investments in this field is long gone. Buy a computer, buy a red rocket, buy a raid, download resolve and call yourself "colorist"... No special funds needed really.
 
Shouldn't EXR files be preferred to DPX?
 
Totally agree, Bruce. Most of these post facilities are busy working and don't have time to search reduser all day for information that is buried and or changes day to day. That said, I can understand your frustration Mark. Also, Redcolor / Redlogfilm is not always appropriate for all scenarios and because there are so many different options for treating RED you really have to take it on a case by case basis. When you work on the bleeding edge of technology you have to be dilligent about educating and questioning these facilities that have systems and workflows in place that were set up long before RED was thought of before you employ them.

Sorry Mark...

When I was doing software design and the user did the wrong thing... you don't blame the user. You blame the software you're designing for leading the user astray.

So looking at it from that perspective:

1. Where is the nice PDF on RED's website that explains the right way to do things?

2. Designing a codec that requires post houses to buy a separate hardware board is very... 1980s. Imagine if every manufacturer did it. I'd have my computer filled up with cards - SonyROCKET, ArriROCKET, PanaROCKET. WTF!!!! R3Ds should decode via GPU. That's RED's fault.

3. With Epic, you should be able to monitor with one color setting... but record metadata that tells the post software to develop using different settings (RedLogFilm / RedColor2). Arri Alexa makes it very easy to do. If RED made it similarly easy, we wouldn't have this problem.

4. The problem with RedLogFilm is that it wasn't there from day one - and it has not been properly publicized by RED. Lots of people don't know it exists. Again, where is the explanation on RED's website? Alexa had LogC from day one, with a clear explanation, so people know to use it.

5. The total lack of documentation / downloadable LUTs makes things difficult to understand.

6. DPXes are not inherently bad. DPXes with RedLogFilm are great in fact. The problem is that RED is pushing a native R3D decode solution... which just doesn't work once you start adding VFX. Instead they should push a clear, simple, DPX-based solution.

Bruce Allen
www.boacinema.com
 
Shouldn't EXR files be preferred to DPX?
yes and no.. DPX is real time in most finishing systems & across most platforms, EXR is not, minimising the value of transcodeing the r3d's as you are not gaining the one thing that DPX does & you are loseing the ablity to change debayer settngs in context
 
Nowadays, those that don't 'get' the RAW workflow are simply fighting it... It is the film workflow -exactly- only 100% digital.

Unless proprietary walls hold it back, the future is fast & bright. A fast CPU, a pair of top-shelf gaming videocards, a card to provide proper output and a RAID controller... That is all an artist will need to start. Of course there is also software, a proper 4k monitor/projector, a control surface and proper RAID arrays. But these enthusiast boxes will let you decode RAW 5k on the fly on one video card, accelerate VFX on the other video card and/or CPU/s, push a proper signal out with the output card and read/write with the RAID card -- all RT RAW -- all in one box that will cost about $3k.

I am not a post house guy, not trying to nail the timing here, or the exact way that these everyone-boxes will handle the data, just ranting... But the simple truth is that there will be MORE than enough power to do EVERYTHING without ANY baking-in on an enthusiast system within in a few years, that's a given... Software is all that's missing.

Adobe's mentality seems to be aiming square at that future. Imagine AE with the power of Premier... Now imagine all others reacting and catching up... EVERYTHING will run on gaming GPUs and/or CPUs. There is no need for anything else 'other than' proper peripherals. 'All' we should be forced to buy is proper signal output, monitors/projectors, control surfaces and safe-deposit-box RAID systems.

I don't think there is any need to freak out whatsoever, but if these insanely powerful and equally affordable workflows worry some, don't... TALENT is what producers really pay for. That is why Toia doesn't have time to do it all himself.

[/rant]

[cheers!]
 
Sorry Mark...

When I was doing software design and the user did the wrong thing... you don't blame the user. You blame the software you're designing for leading the user astray.

So looking at it from that perspective:

1. Where is the nice PDF on RED's website that explains the right way to do things?

2. Designing a codec that requires post houses to buy a separate hardware board is very... 1980s. Imagine if every manufacturer did it. I'd have my computer filled up with cards - SonyROCKET, ArriROCKET, PanaROCKET. WTF!!!! R3Ds should decode via GPU. That's RED's fault.

3. With Epic, you should be able to monitor with one color setting... but record metadata that tells the post software to develop using different settings (RedLogFilm / RedColor2). Arri Alexa makes it very easy to do. If RED made it similarly easy, we wouldn't have this problem.

4. The problem with RedLogFilm is that it wasn't there from day one - and it has not been properly publicized by RED. Lots of people don't know it exists. Again, where is the explanation on RED's website? Alexa had LogC from day one, with a clear explanation, so people know to use it.

5. The total lack of documentation / downloadable LUTs makes things difficult to understand.

6. DPXes are not inherently bad. DPXes with RedLogFilm are great in fact. The problem is that RED is pushing a native R3D decode solution... which just doesn't work once you start adding VFX. Instead they should push a clear, simple, DPX-based solution.

Bruce Allen
www.boacinema.com


Bruce,

you make excellent points here.

I also beleive that RED has to be sold as a complete CAMERA and POST system and not just a camera system. It is recent that RCX P has been the single great replacement for all things post with RED, but in times before these a RED POST WORKFLOW could have been greek to many post houses:

-apple compressor
-monkey extract
-clipfinder
-red rushes
-red alert
-RED CINE
-FCP transcode

Not to mention the color spaces

REDcolor
REDcolor2
REDgamma
Redgamma2
REDSPACE
REC 709
Liniar
RedLog
RedlogFilm

Enough to make any post house (lazy ones) say F it! Just make it all DPX.

RED needs to be as proactive (besides their work with Michael Cioni) at promoting workflow from RED positive houses as they are about promiting Spider man.

It's really that important.

A post house who gets RED footage should be able to go to RED.com and have a plethora of tutorials to help them deal with the multitude of post workflows.

i would put my name forward to help in this task.

I'll start by identifying "some" key workflows.

SHOOT RED EDIT FCP7 FCPX FINISH FINISH APPLE COLOR FINISH FCP 7 FINISH FCPX RESOLVE
SHOOT RED EDIT PREMIERE ----FINISH PREMIER FINISH RESOLVE
SHOOT RED EDIT AVID ----------FINISH AVID FINISH RESOLVE
SHOOT RED EDIT SMOKE -------FINISH DISCREET

SHOOT RED EDIT ANYTHING FINISH PABLO


R3D Native, DPX, Film. It should all be at RED's website.

With the Scarlet onboard, I am seeing the same post questions NOW as I did in 2007, that means there is a basic lack of information.


Thanks,
Sorry for the long winded post. I am passionate about this one though.

David
 
Sorry Mark...

When I was doing software design and the user did the wrong thing... you don't blame the user. You blame the software you're designing for leading the user astray.

So looking at it from that perspective:

1. Where is the nice PDF on RED's website that explains the right way to do things?

2. Designing a codec that requires post houses to buy a separate hardware board is very... 1980s. Imagine if every manufacturer did it. I'd have my computer filled up with cards - SonyROCKET, ArriROCKET, PanaROCKET. WTF!!!! R3Ds should decode via GPU. That's RED's fault.

3. With Epic, you should be able to monitor with one color setting... but record metadata that tells the post software to develop using different settings (RedLogFilm / RedColor2). Arri Alexa makes it very easy to do. If RED made it similarly easy, we wouldn't have this problem.

4. The problem with RedLogFilm is that it wasn't there from day one - and it has not been properly publicized by RED. Lots of people don't know it exists. Again, where is the explanation on RED's website? Alexa had LogC from day one, with a clear explanation, so people know to use it.

5. The total lack of documentation / downloadable LUTs makes things difficult to understand.

6. DPXes are not inherently bad. DPXes with RedLogFilm are great in fact. The problem is that RED is pushing a native R3D decode solution... which just doesn't work once you start adding VFX. Instead they should push a clear, simple, DPX-based solution.

Bruce Allen
www.boacinema.com

Bruce,

great post as usual.

Also to add that in general any RAW doesn't help you so much in the case that your camera has a limited range of color gamut, sensitivity, dynamic range and exposure latitude.
 
Sorry Mark...

When I was doing software design and the user did the wrong thing... you don't blame the user. You blame the software you're designing for leading the user astray.

So looking at it from that perspective:

1. Where is the nice PDF on RED's website that explains the right way to do things?

2. Designing a codec that requires post houses to buy a separate hardware board is very... 1980s. Imagine if every manufacturer did it. I'd have my computer filled up with cards - SonyROCKET, ArriROCKET, PanaROCKET. WTF!!!! R3Ds should decode via GPU. That's RED's fault.

3. With Epic, you should be able to monitor with one color setting... but record metadata that tells the post software to develop using different settings (RedLogFilm / RedColor2). Arri Alexa makes it very easy to do. If RED made it similarly easy, we wouldn't have this problem.

4. The problem with RedLogFilm is that it wasn't there from day one - and it has not been properly publicized by RED. Lots of people don't know it exists. Again, where is the explanation on RED's website? Alexa had LogC from day one, with a clear explanation, so people know to use it.

5. The total lack of documentation / downloadable LUTs makes things difficult to understand.

6. DPXes are not inherently bad. DPXes with RedLogFilm are great in fact. The problem is that RED is pushing a native R3D decode solution... which just doesn't work once you start adding VFX. Instead they should push a clear, simple, DPX-based solution.

Bruce Allen
www.boacinema.com

Hey Bruce,
Great post.
-Sam
 
I think it all boils down to: WE NEED PROXIES IN THE RED MAG DIRECTORY!!!!!!

It's the only way to convince people to embrace the R3D concept, I find that MANY don't have the appropiate hardware (i.e. Red Rocket, or CUDA, etc, etc.) Proxies, like on the Red One, right next to the RAW clip would make the whole experience more accessible either to small companies or lazy big production houses.


PLEASE WE NEED PROXIES ON THE RMD FOLDER PLEASE!!!!!!!!!!!!!!!!
 
But the enthusiast box will already let you decode RAW 5k on the fly on one video card, accelerate VFX on the other video card and/or CPU/s, push a proper signal out with the output card and read/write with the RAID card -- all RT RAW -- all in one box that will cost about $3k.

Could you build this system for me, please?!

Right now I'm on a few year old Mac Pro sans Rocket, and am considering upgrading or just buying a new machine and switching software. Right now I'm starting from scratch on what to get.
 
Also to add that in general any RAW doesn't help you so much in the case that your camera has a limited range of color gamut, sensitivity, dynamic range and exposure latitude.

from a post i made last week;
===========================================================

did it this afternoon

A cam is Epic with HDR on, B cam (also and Epic) is off shooting a splinter unit, C cam is an MX, day int, windows are blown out, all good with the Epic +HDR, not good with the reverses on the MX..

Going to print, cinemascope....

many answers, the one i used is one button click to drop the clip into a composite, link to, and debayer the r3d a second time set at 125 ISO, luma key the windows in from the second pass... one button click gets me back into the color interface - took me longer to type this than to do it...

===============================================================
And that is excatly why one size does not fit all, and why it makes sense to have the debayer settings alive in your gradeig timeline.. if it was DPX debayered @ 400ISO for good exposure on the skintones, the highlights were toast, if it's debayered @ 125 the highlights have a ton of detail & the skintones are down in the mud... i have the best of both worlds and a finishing / color system that can make take advantage of the options quickly and in context.

Not as fast as working with DPX, vastly more flexiable... not so good for a pipeline, great for a boutique with a realistic sched.

d
 
Could you build this system for me, please?!

Right now I'm on a few year old Mac Pro sans Rocket, and am considering upgrading or just buying a new machine and switching software. Right now I'm starting from scratch on what to get.

Read my post carefully. I'll edit out the 'already' in that sentence to make it clearer... We shouldn't derail this thread. There are tons of other threads with specific, up to date information and thorough discussions on the subject.
 
And that is excatly why one size does not fit all, and why it makes sense to have the debayer settings alive in your gradeig timeline.. if it was DPX debayered @ 400ISO for good exposure on the skintones, the highlights were toast, if it's debayered @ 125 the highlights have a ton of detail & the skintones are down in the mud... i have the best of both worlds and a finishing / color system that can make take advantage of the options quickly and in context.

Um: something is wrong with RED's 400ISO decode if it's hard clipping the highlights.

DPX files have worked great for describing the large latitude of film for decades. If RED can't fit the (smaller) latitude of R3D files into the DPX file... it's RED's problem.

Ideally if you just tell the R3D to decode at 400ISO, it should mean good exposure on the skintones... AND the highlight detail still present (nicely rolled off).

Personally, my workaround is to just convert at 250ISO (which maintains all highlight detail) with a little bit of FLUT to get overall exposure right (which moves exposure but doesn't clip highlights as much).

There is no benefit to going to 125ISO - if the highlights are clipped at 250ISO, there's no extra detail to be found at 125.

I know that you're using this as an argument for R3D... but all it sounds like to me is that you just don't know how to make a good DPX file.

Which comes back to my point: it's too easy to make a DPX file that is clipped, using RED's software.

Bruce Allen
www.boacinema.com
 
There are many valid points in this thread about the situation with RED work-flows but I wanted to return to one that Bruce Allen mentioned. The fact that the RED SDK doesn't support GPU decoding of R3Ds thus forcing you to get a RED ROCKET for faster work.

This is one of the things that RED should reconsider. It might have been a nice solution 3 years ago but with the way GPU processing power has increased and prices have reduced lately it's a shame we can't use what's already available to us. GPU utilisation for R3D decoding would make easier the life of many people and help improve the perception of difficulty to working with RED footage.
 
Back
Top