Testshoot with Alexa and Epic Dragon in germany

[Björn Benckert]

It's ironic, that you're talking about lack of knowledge and general confusion and in the same breath you are doing exactly the same. 4:2:2 and 8 bits are not the same thing.



Yes, using pre-graded material is a standard type of workflow. That is done for many reasons, one being, that there is no such thing as RAW pipeline. You can't edit, grade or composite RAW material. It must be debayered first and more important, once debayered, you can't go back to bayered image. Use of pre-graded EXR for compositing is quite common. The bigger problem is maintaining the proper color science, so everyone involved in the project are looking at the same image, regardless of platform. FilmLight is making great strides toward this goal by providing a Baselight plugin for Avid/FCP and Nuke. So, everyone in editorial, grading and VFX are on the same page. That is how Elisium was done. Everything was pre graded and exported as EXR and VFX placed and matched everything perfectly with existing pre graded plates. Once color grading received VFX shots back, they were placed in the timeline and usually there wasn't even a need for any additional grading.

4:2:2 log, is "about" 8 bit after grade.

I work raw as we speak as I do most of the time. The pipe is as follows: r3d 5k straight into flame batch, composite nodes fed with a live feed from redlogfilm input node, sharpening, grain, grade nodes, log2lin conversion, batch output straight into a 10 bit timeline at chosen resolution.....

The above flows without any "double dipping" in other words it's only one flow graph for each clip, composting, reframing, downsampling, grading etc, so at any time I can choose to change iso or any other r3d setting and I can also choose where I want my down sampling to happen for any given layer and every layer can have it's iso or any other raw seting changed at any given time, if wanted I can run it all trough at 16 bit even though I find that a bit overdue as I can dial colors to land within 12 bit node with no clipping and render faster downstream from there.

To me the above is a raw workflow. I mean, at some point of time the file needs to land as a result clip and then it's of course not raw any more but if there is no sub renders between raw source clip and final output... well, then thats what I call a raw workflow.

 

[Adrian Jebef]

BTW high speed fps Alexa is BOTH 422 AND 8bit as far as I know.

It is also very true that the level of FUD and ignorance (you are extremely knowledgeable, just a quibbler ) in how to deal with Red RAW (or any RAW) is insane.

Re: ALEXA...

ProRes 4444 is 12-bit 4:4:4 color sampling
ProRes 422HQ is 10-bit 4:2:2 color sampling


Re: RAW...

All raw files are fundamentally useless until they are transformed into a common color/gamma space. The idea that a raw file allows one to "re-develop" the image in a different way at any given point in the workflow pipeline is powerful but can lead to more headaches and inconsistencies with larger-scale VFX-heavy projects. That's why the big projects tend to prefer using DPX Log or EXR as a mastering/intermediate file format instead of continuing to track raw files through the pipe. Consistency is key.

 

[jake blackstone]

One would imagine, that by now people would understand the big picture without trying to reduce it to "that's how we're doing our stuff". I'm talking about a director having creative vision, translated to the big screen by a trusted collaborator- a colorist. Colorist's job is to tell a cohesive story, regardless of camera, format, workflow. So, how does a VFX person to know what is going on visually, if he's getting a RAW file for compositing? He just grades the image from RAW the way he sees it, regardless of director's vision? Or you propose, that the director would supervise every VFX session? And, I can't imagine how would that be helpful anyway. May be a colorist would sit in?
Does that sounds like a quibble, Robert or it may be more of a fundamental issue of cohesive storytelling? The same goes for a commercial work, Bjorn. You can't have many cooks on the kitchen. Someone must be grown up and be responsible for color decisions. But if you're dead set on using RAW for VFX, then right now there is only one way to use RAW files for compositing by many different companies (very common these day for feature work), while maintaining common look set by the colorist and that is what I already had mentioned- FilmLight's approach. You could use LUTs, but that is nowhere as seamless and it doesn't eliminate the faulty/incorrect metadata settings.

 

[Björn Benckert]

Re: ALEXA...

ProRes 4444 is 12-bit 4:4:4 color sampling
ProRes 422HQ is 10-bit 4:2:2 color sampling


Re: RAW...

All raw files are fundamentally useless until they are transformed into a common color/gamma space. The idea that a raw file allows one to "re-develop" the image in a different way at any given point in the workflow pipeline is powerful but can lead to more headaches and inconsistencies with larger-scale VFX-heavy projects. That's why the big projects tend to prefer using DPX Log or EXR as a mastering/intermediate file format instead of continuing to track raw files through the pipe. Consistency is key.

I think the dpx dip is more due to the fact that most composting programs will crash when r3d's are introduced or they can not digest them, for example there is no 3D program that can work with r3d's sadly...

 

[jake blackstone]

I think the dpx dip more due to the fact that most composting programs will crash when r3d's are introduced or they can not digest them, for example there is no 3D program that can work with r3d's sadly...

And the reason for them not working with R3D because there is no reason to do that. EXR quickly replacing DPX at becoming the new Lingua franca for the post world.

 

[Brian Merlen]

I thought that was the purpose of LUT creation? Say the Director and DP have decided they like the LUT I created on set. If I were to send that LUT along with the r3d files it would provide a gist of what the DP and Director wanted. Sure I can still encode it all to something else with that LUT on if that is what you want to work in, but hypothetically we could stay in r3ds and use the LUT all the way through post. Its just not an attractive option as far as you guys needing to process r3ds every step of the way then instead of offline codecs... or at least less intensive codecs I should say...

 

[jake blackstone]

I thought that was the purpose of LUT creation? Say the Director and DP have decided they like the LUT I created on set. If I were to send that LUT along with the r3d files it would provide a gist of what the DP and Director wanted. Sure I can still encode it all to something else with that LUT on if that is what you want to work in, but hypothetically we could stay in r3ds and use the LUT all the way through post. Its just not an attractive option as far as you guys needing to process r3ds every step of the way then instead of offline codecs... or at least less intensive codecs I should say...

The RAW file is placed in the timeline, debayered, pre-graded and sent on it's way to VFX in the form of Log DPX or EXR. The viewing LUT is sent along, so everyone is on the same page. There is almost no advantages to maintain RAW for VFX. But I can only think of number of disadvantages. This myth of RAW superiority vs EXR or DPX is just that- a myth...
The LUT's created on set may get used for offline or it may become the basis of grade, but eventually a real LUT is created for grading a feature, unless that is the whole thing is done using ACES. In that case, there are no LUTs used, but color transforms.

 

[Björn Benckert]

The RAW file is placed in the timeline, debayered, pre-graded and sent on it's way to VFX in the form of Log DPX or EXR. The viewing LUT is sent along, so everyone is on the same page. There is almost no advantages to maintain RAW for VFX. But I can only think of number of disadvantages. This myth of RAW superiority vs EXR or DPX is just that- a myth...

Well, yes colorist hate me for grading in flame, but for a lot of short format productions (read commercials) I find it quite convenient DPX files and EXR does not come with built in proxies, they do not come compressed and even if they are capable to hold everything that was captured they also add weight, as in the case with RED the original source is compressed raw 16 bit. So personally I think that some, like my self will continue to work straight from trimmed r3d's. Flame/lustre allow for this workflow and I find it to be a very good setup not for everyone for sure but, I think that a solution where the comps are floating in terms of grading until final render is a real plus and for certain applications and VFX work especially I find it to be the way of the future... Especially if you think of 6k source then even more so. I love to fiddle around with r3d's and and be able to suck them straight over san network and do my sketch composting at a quarter res or so, then it's like melted butter and when I like what I got in my comp I can simply just set the inputs to full debayer, adjust my settings and render the final, especially when doing 4k work then the r3d's shine in comparison to EXR or any other uncompressed formats. As most things we do are shot at 5/1 or even higher compression then bumping the files to uncompressed 16 bit makes things...well, a lot slower. So even if you do not like it I think others will... But fine with me, I like the idea of my workflow being a myth.

Here is a screen shot of my myth of the day... 29+ layers of r3d's that gets multiplied and comped together into a crowd dup shot.

 

[jake blackstone]

Well, yes colorist hate me for grading in flame, but for a lot of short format productions (read commercials) I find it quite convenient DPX files and EXR does not come with built in proxies, they do not come compressed and even if they are capable to hold everything that was captured they also add weight, as in the case with RED the original source is compressed raw 16 bit. So personally I think that some, like my self will continue to work straight from trimmed r3d's. Flame/lustre allow for this workflow and I find it to be a very good setup not for everyone for sure but, I think that a solution where the comps are floating in terms of grading until final render is a real plus and for certain applications and VFX work especially I find it to be the way of the future... Especially if you think of 6k source then even more so. I love to fiddle around with r3d's and and be able to suck them straight over san network and do my sketch composting at a quarter res or so, then it's like melted butter and when I like what I got in my comp I can simply just set the inputs to full debayer, adjust my settings and render the final, especially when doing 4k work then the r3d's shine in comparison to EXR or any other uncompressed formats. As most things we do are shot at 5/1 or even higher compression then bumping the files to uncompressed 16 bit makes things...well, a lot slower. So even if you do not like it I think others will... But fine with me, I like the idea of my workflow being a myth.

Here is a screen shot of my myth of the day... 29+ layers of r3d's that gets multiplied and comped together into a crowd dup shot.

Given enough time, you may create some interesting looking image, but if that "grade" doesn't work with the rest of the storyline of the film set by the director and interpreted by colorist, then what? Do you think it's better for a colorist to have a full set of metadata controls to grade with or for VFX guy to fiddle with the grade of the single shot? I think you know the answer…
On the side note, if you need 29 layers to grade something, there usually something wrong with either the material or the person grading Flame/Smoke are awful devices for grading- overly complicated, very slow (can you say render?) and not really optimized for such operation. And that is why AD still offers Lustre.
And one final point, pushing some colors around is not color grading. Using colors to enhance the story, is.

 

[Adrian Jebef]

Bjorn is a true original. There aren't many out there that have the technical background and experience to pull off his preferred "raw workflow". It CAN BE DONE. Bjorn is proof. And I'm sure his results are pretty fantastic. That being said, he gets it done this way because he has full control of the pipe. That's not the case with most large-scale productions. Though there are production, post-production, and VFX supervisors that oversee a large-scale film the day-to-day demands of a tight schedule don't leave any room for one individual to "do it all".

Again, consistency is key. A raw file is the opposite. It's "undeveloped" and can be "interpreted" a thousand different ways. Tracking metadata correctly through a post pipeline is one of the biggest hurdles to overcome at this stage in the game. Transforming R3D's or other raw file formats into "properly developed" DPX or EXR files for DI and VFX gives much more consistent results. If done properly you lose exactly ZERO information recorded in the corresponding raw file. And doing these transform/transcodes correctly really isn't that big a deal. Is is an extra step? Of course it is. But it needs to be done. The only real way to loose this step is to get your camera to record DPX or EXR files natively. But then it wouldn't be "raw" and everyone might start feeling like they're getting the short end of the stick...

 

[Blair S. Paulsen]

My only quibble would be that you inevitably lose some information when making an RGB file from the RAW source. But, assuming that process includes landing that RGB image in a good place for the rest of the pipeline, and describes it in a robust container/format, I'd argue you haven't given away anything you were ever going to be able to use. No delivery format will hold the complete data set of the RAW file anyway. My goal is to make good choices about best practices for each "tributary" based not just on its inherent properties but the real world parameters of the finishing process.

If the post infrastructure can evolve to the point where RAW essence/metadata/monitoring(LUTs) are handled seamlessly I'd love to save the storage space - Björn has that part dead right - but as others have noted that is impractical for many (I think we're mostly jealous ;-).

There's also the practical question of whether it serves the project to put off color decisions just because you can - but that's another thread.

Cheers - #19

 

[Robert Ruffo New]

Re: ALEXA...

ProRes 4444 is 12-bit 4:4:4 color sampling
ProRes 422HQ is 10-bit 4:2:2 color sampling


Re: RAW...

All raw files are fundamentally useless until they are transformed into a common color/gamma space. The idea that a raw file allows one to "re-develop" the image in a different way at any given point in the workflow pipeline is powerful but can lead to more headaches and inconsistencies with larger-scale VFX-heavy projects. That's why the big projects tend to prefer using DPX Log or EXR as a mastering/intermediate file format instead of continuing to track raw files through the pipe. Consistency is key.

That's why I pointed out that it can't work on VFX-heavy projects. But some projects have no VFX at all.

Thanks for the clarification on Alexa - I have never used it for high speed and didn't know much about that function.

 

[Robert Ruffo New]

One would imagine, that by now people would understand the big picture without trying to reduce it to "that's how we're doing our stuff". I'm talking about a director having creative vision, translated to the big screen by a trusted collaborator- a colorist. Colorist's job is to tell a cohesive story, regardless of camera, format, workflow. So, how does a VFX person to know what is going on visually, if he's getting a RAW file for compositing? He just grades the image from RAW the way he sees it, regardless of director's vision? Or you propose, that the director would supervise every VFX session? And, I can't imagine how would that be helpful anyway. May be a colorist would sit in?
Does that sounds like a quibble, Robert or it may be more of a fundamental issue of cohesive storytelling? The same goes for a commercial work, Bjorn. You can't have many cooks on the kitchen. Someone must be grown up and be responsible for color decisions. But if you're dead set on using RAW for VFX, then right now there is only one way to use RAW files for compositing by many different companies (very common these day for feature work), while maintaining common look set by the colorist and that is what I already had mentioned- FilmLight's approach. You could use LUTs, but that is nowhere as seamless and it doesn't eliminate the faulty/incorrect metadata settings.

In our workflow the colorist is handed an XML file and then does a session with the director and DP on Resolve and/or whatever other softs he/she wants to use.

Before edit, same three people (if those are three different people) do a rough one-light in RCX and re-save the metadata for Premiere, or make offline Prores or DNX files if editor is using Avid or Finalcut 7. I don't really see the problem, or how this is not what one could call "RAW workflow". As I 've told you before, my projects are not the stupendous and huge and prestigious ones that you work on, so there is not always millions and millions worth of VFX, sometimes none at all. When there is VFX effort is made to use high bitrate DPX when needed. No one fudges the color without asking "the key three".

Not sure how this breaks anyone's ability to tell stories, or do anything else. It works for our projects, and having seen Bjorn's impressive body of commercial work for top-drawer clients, I don't see what right you have to tell him what to do, it's obviously working for him too. Not everybody wants to be just like you or do things the way you do them - it doesn't make them "wrong" or you "wrong" - it just makes their circumstances completely different from yours.

I like the way we do things here in our little shop and so do our clients. We like our clients. Everybody's happy all around. Deal with it.

 

[Robert Ruffo New]

Given enough time, you may create some interesting looking image, but if that "grade" doesn't work with the rest of the storyline of the film set by the director and interpreted by colorist, then what? Do you think it's better for a colorist to have a full set of metadata controls to grade with or for VFX guy to fiddle with the grade of the single shot? I think you know the answer…
On the side note, if you need 29 layers to grade something, there usually something wrong with either the material or the person grading Flame/Smoke are awful devices for grading- overly complicated, very slow (can you say render?) and not really optimized for such operation. And that is why AD still offers Lustre.
And one final point, pushing some colors around is not color grading. Using colors to enhance the story, is.

You have real trust issues, which on a small team of people you know well (like our team, likely Bjorn's) is never a problem. I would say you seem to have trust and mutual respect issues all around. You seem to think everyone is an idiot except you. I would venture that is not actually true, and that if you stopped thinking that you would be much happier and more successful.

 

[Jan Wenstad]

I don't know how many of you guys have worked on VFX shots for a films... But this is how it goes... Color is the last piece of the puzzle... The film scan (DPX file sequences) get sent to many post houses to do the tracking, 3D and comps. Everything, including grain, is matched to the scan and re-exported back out as a file sequence. Then when all the pieces are put back together, the colorist does his job.

All this back and forth that I'm reading can be so mind numbing.

 

[jake blackstone]

You have real trust issues, which on a small team of people you know well (like our team, likely Bjorn's) is never a problem. I would say you seem to have trust and mutual respect issues all around. You seem to think everyone is an idiot except you. I would venture that is not actually true, and that if you stopped thinking that you would be much happier and more successful.

You know what Robert?
I finally had it with you. Your psychobabble, like saying, that I'm having trust issues, endlessly trying to analyze what makes me tick, which usually has nothing to do with the subject at hand is getting pretty tiresome. From now on, feel free to ignore my posts, as I do yours most of the time. Actually do me a favor, put me on your ignore list.
What makes you think, that I need your advice on what I need to do in order to be successful? Should I follow your road to success?
And finally, no, I don't think everyone is an idiot, just you…
BTW, while visiting your site http://www.robertruffo.com browser gives a suspected malware warning. You need to look into that

 

[Antoine Baumann]

I understand your point but as Paul said, I think "there are weird workflow stuff going on...". Like if they prefer to stick to the good old DPX workflow. But I'm not (anymore) a post-production specialist. I won't argue here if a 10bit DPX is as good as a R3D, but why switch format, do you need to embed the look at that stage of the workflow?

Pat

Because we want your cpu and gpu to do some compositing tasks, not debayering. Of course if you have 32 cores and multiple gpu, we may want to do compositing on r3d files.

 

[Nick Morrison]

You know what Robert?
I finally had it with you. Your psychobabble, like saying, that I'm having trust issues, endlessly trying to analyze what makes me tick, which usually has nothing to do with the subject at hand is getting pretty tiresome. From now on, feel free to ignore my posts, as I do yours most of the time. Actually do me a favor, put me on your ignore list.
What makes you think, that I need your advice on what I need to do in order to be successful? Should I follow your road to success?
And finally, no, I don't think everyone is an idiot, just you…
BTW, while visiting your site http://www.robertruffo.com browser gives a suspected malware warning. You need to look into that

Having worked with Jake recently on several jobs, I can saw w/out hesitation he's not only EXTREMELY TALENTED, but also a joy to work with and someone I found myself "trusting" right away...so certainly no "trust issues" of any kind (quite the opposite actually!). These jobs were usually very last minute with fast turn-arounds, and we only survived because of our quick short-hand, and the trust we built so effortlessly. In fact I ended up trusting Jake with my favorite piece of the year, which of course...he hit it out of the park! You can see his sublime, mesmerizing grade here:

Robert, I think you're being provocative. Not everyone likes that BTW. You're a great super smart dude, but if you can try and make comments just a bit less personal...would probably make things a lot easier!

best

 

[Hans von Sonntag]

Robert, I think you're being provocative. Not everyone likes that BTW.

best

I've been pretty active on these forums in the past when digital cinema was new and exciting. But with the introduction of the Red Scarlet a new crop of folks joined here and the place git pretty crowded. Back then I felt that I cannot start over again and took a sabbatical of 3 years. Now with the introduction of the Dragon sensor I logged in here again and was pleased to see valuable sources such a Michael Most, Jeff Kilgroe, Jake Blackstone, Bjoern Beckert, Gunleik and others still around letting redusers participate in their deep knowledge. They are all seasoned experts in this industry with a wide horizon only a few have.

Robert, please do Red the favour and don't scare them away. They are the heart of these forums, not the masses that have questions.

There is no need to be personal. If their opinion does not comply with your findings say so without personal attacks.

Thanks,

Hans

 

[Hrvoje Simic]

Good thing human's fine tuning capabilities are applicable to many realms.