Ask Mike Most Anything

[M Most]

Are there any guidelines regarding this or very heavy reasons to pick one or the other or are they freely exchangeable?

thanks

They are freely exchangeable. If one is posting in the US, 23.98 will avoid a lot of headaches because it's usually what's presumed to be used. Anyone making a DCP will in the course of making that DCP pull the sound rate up and play the images at 24fps because that's required by the DCI spec. That is a common operation and not problematic in the least. Same goes for film prints, the sound will be pulled up on request in the course of making the optical tracks. Europe tends to use "hard 24," but in the US, that is not the case.

Not to mention that video deliverables - which are still always required - will need to be at 23.98 anyway.

 

[Elsie N]

Well, one thing I would definitely recommend is some sort of proper viewing monitor, because without that, you can't really have any confidence that anything you do will look the way you want it to in a festival screening. There is nothing wrong with a DIY approach, in fact, at least in theory, there is really no quality difference between desktop tools and "high end" tools. However, there is often a serious difference in the monitoring of the intended result. As long as you have some reasonable assurance that what you're seeing and hearing is what the audience at the festival screening is going to see and hear, you should be fine. Everything beyond that is just workflow specifics. So make sure you set yourself up with proper monitoring and let me know how it goes.

Mike, I haven't acted on this advice just yet because I felt that everything that would get me what you suggest is out of my budget range. But now, the 39" 4k Seiki monitor seems to fit within budgetary limits. Don't know if you are yet acquainted with it, but if you are, I would appreciate your comments on its use.

EDIT FOR CLARITY: I would hope to use one of the inexpensive monitor/TV calibration tools to get something close to what a 4k TV audience might see. (Can't find the link to it at the moment.)

 

[M Most]

Mike, I haven't acted on this advice just yet because I felt that everything that would get me what you suggest is out of my budget range. But now, the 39" 4k Seiki monitor seems to fit within budgetary limits. Don't know if you are yet acquainted with it, but if you are, I would appreciate your comments on its use.

EDIT FOR CLARITY: I would hope to use one of the inexpensive monitor/TV calibration tools to get something close to what a 4k TV audience might see. (Can't find the link to it at the moment.)

I have no direct experience with it. However, from what I've been told by those who do, the out of the box setup of this monitor is not even close to a "proper" calibration. Of course, since there is also no standard that anyone is adhering to for UHD color at this point, I have no idea what a "proper" calibration would be. But for most material, some sort of adherence to Rec709 colorimetry is probably in order, if only for the fact that the vast majority of viewers are going to be looking at that material on an HD monitor for the foreseeable future, regardless of what you're using to look at it.

I would also caution that even in this world of cheap and free stuff, you still get what you pay for, at least to some extent, and especially with things like displays. That monitor may "look good," but that doesn't mean it reflects any kind of standard, and it most definitely doesn't mean it qualifies as a "grading" monitor, which by definition must adhere to some kind of standard. Do what you want with it, but don't expect it to be any kind of reference device, especially when 98% of your viewers are going to be looking at the material in HD. The fact that this monitor can display a higher resolution doesn't mean crap because you're not making color decisions based on resolution. Higher resolution doesn't have anything to do with color accuracy. If you like what it looks like in your living room, that's very nice. But it has no bearing on its suitability for professional use.

 

[Elsie N]

I have no direct experience with it. However, from what I've been told by those who do, the out of the box setup of this monitor is not even close to a "proper" calibration. Of course, since there is also no standard that anyone is adhering to for UHD color at this point, I have no idea what a "proper" calibration would be. But for most material, some sort of adherence to Rec709 colorimetry is probably in order, if only for the fact that the vast majority of viewers are going to be looking at that material on an HD monitor for the foreseeable future, regardless of what you're using to look at it.

I would also caution that even in this world of cheap and free stuff, you still get what you pay for, at least to some extent, and especially with things like displays. That monitor may "look good," but that doesn't mean it reflects any kind of standard, and it most definitely doesn't mean it qualifies as a "grading" monitor, which by definition must adhere to some kind of standard. Do what you want with it, but don't expect it to be any kind of reference device, especially when 98% of your viewers are going to be looking at the material in HD. The fact that this monitor can display a higher resolution doesn't mean crap because you're not making color decisions based on resolution. Higher resolution doesn't have anything to do with color accuracy. If you like what it looks like in your living room, that's very nice. But it has no bearing on its suitability for professional use.

Thanks Mike... words to live by... (but hoping someday there will be something that gets us closer, both soft and hardware-wise, to an Indie grading solution.)

 

[mark haflich]

Hey Mike,
I've been debating upgrading our facility projector for a few months and have been hesitant due to the added cost of installing 220w power required and a little bit of contractor work on our room. I recently heard about the new Sony VPL-VW1000 4K home theatre model. It says it's DCI complient and all the reports I'm hearing are great. But I was wondering if you had any experience or advice in this area? Is the Barco DP-12C the better way to go with the added cost factor? Our room is approximately 22ft x 18ft and we have a 147" Stewart Screen so it should be perfect to hang in our current spot approximately 16ft from the screen. Any info would be appreciated.
Thank you sir!

I own a Sony VPL-vw1000ES and it is definitely not DCI compliant and the Sony literature does not say it is. It will however replicate the DCI color space. I had the opportunity before the projector came out to review and offer edits for some of the literature on that model and offered revisions which Sony accepted to remove the word DCI compliant from the specifications and literature.

 

[mark haflich]

Well, one thing I would definitely recommend is some sort of proper viewing monitor, because without that, you can't really have any confidence that anything you do will look the way you want it to in a festival screening. There is nothing wrong with a DIY approach, in fact, at least in theory, there is really no quality difference between desktop tools and "high end" tools. However, there is often a serious difference in the monitoring of the intended result. As long as you have some reasonable assurance that what you're seeing and hearing is what the audience at the festival screening is going to see and hear, you should be fine. Everything beyond that is just workflow specifics. So make sure you set yourself up with proper monitoring and let me know how it goes.

And two years later the OP because of financial limitations which all too many here face proffers a wish seeking reassurance that somehow a 39 inch 4k 16/9 Chinese panel for $699 is sufficient to do 4K grading and mastering and give a jack of all trades digital filmmaker the tool (the so to speak final link in the chain) to see the results of his work with reasonable assurance it will appear the same way the audience at a film festival will see it. Mike as usual with grace and knowledge states the obvious answer. Now let me ask two questions. If you are shooting, mastering, and displaying in 2K whether it be DCI or rec 709, what is a suitable monitor size? And I think this will really be a two part answer, leading up to a conclusion that some sort of small panel, 50 inch D in size, will be inadequate to see everything that a festival audience will see and NOTICE. Don't you in fact need to use a projector with at least a 6 ft wide screen for that to see the errors you will likely have made working on a 50 inch panel? Likewise, I ask the same question for 4K where something like a 10 ft wide screen would be required. I suspect this post will not be popular. Many here a producing not for a large screen venue. Many are used to 2K and because of Red are now thrust into the venue of 4k before there are high quality affordable4K monitors and before there are affordable 4K projectors which are desperately needed by the production industry. But yes kiddies there are film producers out there that are using the Sony 4K VPL-vw1000ES for exactly that purpose. Expensive but B stock can be had for under $15K. But where does the industry go from here? Obviously, there will be cheaper HT 4K projectors coming later this year and the cash starved industry will employ those especially if they are brighter than the 1000ES. And what will Red be doing/. Red has booth space at the Cedia show (consumer HT) in Denver late Sept. Why are they going? To sell expensive 4K cameras to consumers? I think not. To hawk Redrays to consumers factory direct at $1750 per pop? I think not. To give Odemax some booth space, yep, a small part of why they are going. So why are they going? Maybe they will be hawking for future coming soon delivery a playback only Redray. But MAYBE they will have a 4K consumer projector that the production industry can use. Will it be true laser? Not a Fing chance. But it could be a laser lit phosphor replacement for a bulb light source. Such sources are available for use now, kick out a huge amount of white light, as class 2 lasers, avoid the normal regulatory laser hurdles and presently effective bars to market entry. The light engine beyond light source is pretty much straight forward and does not require great innovation. Market entry by Red for under $10K sold factory direct at low margin should be possible. But Red would have to break its historical track record and deliver product within 60 to 90 days of showing. Stay tuned. Elsie, your monitor and the monitor needed by all makers of 4k films for large venue displays might be coming.

 

[Peter Moretti]

Hi Mike,

Thank you so much for contributing to this thread! I have a few questions about how Red footage is graded in Resolve with visual effects composites.

Since only Red cameras can write .R3D files, it seems that there will be a file format mismatch between the sources for the non-VXF shots and those for the VXF composites. For example, let's say 30% of my final edit is going to need VXF work in AE, so 70% will not. Now I realize that I can use an EDL, AAF or XML to transfer my edit into Resolve and relink to the original .R3D files. But that will only work for the 70% of the show that is non-VXF. For the VXF shots, the VXF composites effectively (no pun intended) replace the .R3D's as "original" source files, don't they?

Which brings up a few questions. The first being, if I want to work in 4K in Resolve with the intent of providing the necessary output (uncompressed TIFF's I believe) for making a 2K DCP, what file format and raster size should I have the AE composites rendered to? I've considered DPX, ProRes 4444, DNxHD 444 or 185x (but they are only HD) but I really don't know the answer. And won't the AE renders grade "differently" than the .R3D's?

My other question is about conforming the AE renders into the Resolve timeline. Because this material will be intersperesed through out the film, there needs to be a way to get Resolve to understand which clips have AE renders for their source media and which have .R3D's. But how is this done?

I hope that is clear. And thank you very much for answering!

 

[M Most]

...if I want to work in 4K in Resolve with the intent of providing the necessary output (uncompressed TIFF's I believe) for making a 2K DCP, what file format and raster size should I have the AE composites rendered to? I've considered DPX, ProRes 4444, DNxHD 444 or 185x (but they are only HD) but I really don't know the answer. And won't the AE renders grade "differently" than the .R3D's?

Taking the last part first, they won't render "differently" provided you do the pulls using the same settings you use when you grade. That means the same camera metadata, the same color space, and the same gamma. In most DI situations today, log format images are usually used, so that would likely mean Redcolor3 (or 1 if you want more saturation), RedlogFilm, and whatever settings were either on the camera or in the RMD file. You want to leave out any curve corrections, though, as those are not universally supported, nor do they contribute to a usable log image in most cases. This is especially true in television work, where there's not much time and everyone wants things done as efficiently as possible. If this is done correctly, you can usually grade the non-effected shot - either directly from the R3D file or from an RGB image made with those settings - and apply the same grade to the final VFX shot when it is delivered, and get essentially an identical result. As for the first part, what you use depends on the circumstances you have, particularly in terms of equipment, storage, and software choices. So if you happen to have some big, fast storage and you're using, say, Resolve, you probably want to use DPX files, because they require no decompression overhead. So if you have fast storage, they will give you better performance than any of the compressed formats. If you're really going to work in 4K - and mind you, I really don't know why you would do that unless you have a required deliverable in that format, but that's just me - you might want to use Prores, in part because it's widely supported, and in part because the performance of those files is pretty good on commodity hardware. It's really the only commonly available format that I would say that about right now. Using Prores4444 will usually avoid any annoying gamma issues. But at the "pro" level, where fast storage is available and Linux based grading systems are usually used (be it Resolve, Lustre, or Baselight) DPX is by far the most common choice.

My other question is about conforming the AE renders into the Resolve timeline. Because this material will be intersperesed through out the film, there needs to be a way to get Resolve to understand which clips have AE renders for their source media and which have .R3D's. But how is this done?

Well, one of the reasons for using DPX files is that they support time code. So in most cases, time coded DPX sequences are produced, with specific naming conventions, and each shot in its own folder, also with a specific naming convention. Editorial media is then made from these sequences, and the editor cuts them into the show. EDL's are then passed on for conforming and if everything is handled correctly, the metadata in those EDL's matches that of the DPX folders/files, so they auto conform pretty simply. With very late arriving shots, they might be dropped in directly, but this is usually done only if they are updated versions of already edited shots.

 

[Blair S. Paulsen]

Insert Blair's usual rant about how great it would be to have a decent mezzanine codec for UHD/4K post workflows. As Mike suggests a 10bit log DPX based pipeline is the obvious choice but at 4K that's north of a GB (byte, not bit) per second. If 2K would work for your project, its hard to beat a well supported and robust format like DPX. I further agree with Mike that ProRes4444 is the most practical alternative if you have to be at UHD/4K rez and don't have oodles of extremely fast storage. If you go the ProRes route I suggest you test out gamma behaviors in advance, including reviewing test images in the deliverable format.

Cheers - #19

 

[Peter Moretti]

Thank you very much Mike and Blair. I really appreciate the time you've spent on this! Here are a few follow-up questions if I may.

It seems that you are both saying that grading a combination of 2K VXF files (be they DPX or ProRes 4444) and 4K .R3D's does not cause a significant resolution mismatch if your delivery requirements are 2K DCP, BD and DVD. (Which makes me wonder about a rather "ghetto" way of getting the edit into Resolve: render the entire sequence out as 2K ProRes 4444 and use Resolve's scene detection to create the clips. That would of course necessitate having a truly final cut because there would be no handles with which to work.)

It also seems that the conform of the VXF files is done in the NLE not in Resolve. This way the XML, AAF or EDL created by the NLE points to the correct files for the VXF shots. Am I understanding that correctly?

Thank you very much again!

 

[Frank Cueto]

Hey Mike, we recently acquired a Flame Premium 2013 20th AE. As such it comes with Lustre as part of the suite and we are considering using it for client sessions as well as our Baselight (we are awaiting our new gen BL2 as I type this :emote_happyhappy.

What kind of familiarity do you have with Lustre? What strengths would you say it brings above Baselight? What are the cons? Also, I am undecided if we want to get the autodesk control surface or go with the tangents. What do you think?


I have tons of other questions but I don't want to abuse your generosity.


thanks.

 

[jake blackstone]

Obviously, I'm not Mike. But I do have some familiarity using Lustre for couple of years.
Lustre had been a bustard child for a few years now and as such, there is not much development or support, that had been going on.
The bigger problem of Lustre is it's close to none existent conforming tools. Luckily, with Flame Premium, that's not a huge problem, as the conforming usually is performed in Smoke. Still little annoying, especially, if your client asks you to move shots around a bit.
Native support in Lustre is pretty much limited to non encapsulated formats, like DPX or TIFF. Using anything else, like Prores or DnxHD has to be done with the assistance of Wiretap Gateway. For real time performance with those kinds of files you need a very fast network, which if memory serves me right, you may not have.
Another big issue for Lustre is very limited single GPU support. Also, unlike Smoke and Flame, Lustre GPU support is limited to an outdated and overpriced Quadra GPU with an SDI daughter card. Everything needs to be rendered, unless you want to use RAM cache playback For real time performance Lustre needs a serious Backburner render farm, which is not trivial endeavor. Otherwise, every time you want to see real time playback, prepare to wait for Lustre to render to disc cache.
Majority of Lustre users prefer CP-100. That's what i chose. I'm pretty sure you can pick one now for next to nothing.
As much as it pains me to say it, Resolve 9 and even more so, Resolve 10 is head and shoulders better, than present day Lustre. And don't forget free and quick Resolve updates, unlike may be, once a year updates for Lustre, as well as wealth of being able to read and write pretty much anything using Resolve, free support, expendability and universal acceptance.
I know Resolve does't fit your business model, but that's today's reality. Lustre is a beautiful software, but it has one foot in the grave and, unfortunately, it's way past it's prime.

 

[Frank Cueto]

Hey Jake, thanks for the reply.

I kinda feel the same way about Lustre, but I have been to a couple post houses where they have both (Resolve and Lustre) and they seem to run their big theaters with Lustre. Not really interested in leaving Baselight, but since Lustre came with the package I thought that I might as well try it. My only experience with Lustre are some trim passes that I have supervised in the past and it seemed mighty responsive back then.

As far as the Prores issue, they solved it, you no longer need a wiretap gateway to do the prores READ, that is.

I guess i'll play around with it and maybe try it for the odd job, who knows. Its not a bad thing to have laying around

 

[Björn Benckert]

Hey Jake, thanks for the reply.

I kinda feel the same way about Lustre, but I have been to a couple post houses where they have both (Resolve and Lustre) and they seem to run their big theaters with Lustre. Not really interested in leaving Baselight, but since Lustre came with the package I thought that I might as well try it. My only experience with Lustre are some trim passes that I have supervised in the past and it seemed mighty responsive back then.

As far as the Prores issue, they solved it, you no longer need a wiretap gateway to do the prores READ, that is.

I guess i'll play around with it and maybe try it for the odd job, who knows. Its not a bad thing to have laying around

Even if Luster is supposed to be a compliment to flame. it's to me at least a compliment to flame that makes it somewhat possible to bring the grading of complex comps out of the flame suite and then back. Which maybe was a good idea, but it's not worth any extra charge, resolve beats them as stated above.

Curious, how do you like your flame 2013? Personally I'm sticking with 2012 as I find the 2013 to be to much of an unfinished remake.

 

[Frank Cueto]

Bjorn, wish me luck, I am new to Flame. I have lots of experience with Smoke though, about 10 years, so I hope I wont be too lost. I have a mostly CGI/COMP background so I hope to be up and running in a couple weeks. I honestly got tired of all the roundtripping amongst 8 or 10 apps to get my work done.

Should I request them to install the 20th ae and the last previous release? I have seen some online bickering regarding the changes.

 

[M Most]

What has been said already has validity. My own feeling is that since you already invested in Flame Premium, you already have Lustre by definition, and you already have Smoke, also by definition. In the hands of the right artist(s), those programs represent a pretty damned powerful combination with "no excuses" capabilities that quite frankly exceed those of any standalone solution such as Resolve, Baselight, Film Master, Smoke (standalone, on Mac), DS, or anything else I can think of. Lustre itself has what are probably the best and most accurate keyers anywhere, as well as a number of features related to grouping, sorting, and then unsorting that can be very helpful in many circumstances. There is, as always, a rather steep learning curve, but you had that with Baselight and your artists dealt with it. To me, having both Flame Premium and Baselight sounds a bit superfluous. In Flame Premium, you have a package that can do everything you can ever need to do, and all with world class tools. As you've already noted, they've taken care of most of the weaknesses in native format support. I would suggest that you let your artists go at it on all three programs (Flame, Smoke, and Lustre) and see what efficiencies you can gain by using all of them together. As Jake noted, you probably need to look at some additional hardware such as panels and/or Backburner licenses, to make it really sing. But I can tell you that there are a number of very high end companies that have given Autodesk enough business lately to ensure continued support of the Flame Premium package, including Lustre. So dive in and judge for yourself.

 

[Scott Crawley]

This has likely been asked before, so I beg your pardon in advance.

I am beginning to consider training options in a serious way. Where did you receive your most valuable training as a Colorist? Who in the U.S. has the best formal programs in your opinion?

 

[Rudi Herbert]

Mike, if I may,

Saw Elysium in 4K. Yes it looked sharp enough even if uprezzed from 3.3K, and no I particularly did not love the grade, BUT what impressed me was the actual quality of the grade. It is the best looking RED footage I've seen in terms of color accuracy and stability, even if other films have had more expansive, saturated looks (which came at the expense of mildly to wildly varying tones and casts). I have to assume this precision comes from the ACES workflow? I wanted to learn more about and I headed to your blog, where your explanation made sense of course. All three stages of the process are of course self explanatory, but the first stage, where in-depth mapping, or whatever the term may be, of each camera is performed feels a bit "unachievable" to me. How can a camera/sensor be mapped out in such way that its response is know to the extent it can be anticipated under a variety of conditions? Wouldn't this require access to privileged information from the manufacturer? How would it be, or how is it, done in reality? And to what degree of accuracy and completeness? Obviously, this would be the most important part in achieving a truly universal color pipeline, the rest is just a matter of implementation and adherence to standards, but mapping out cameras feels a little unrealistic to me...

 

[M Most]

This has likely been asked before, so I beg your pardon in advance.

I am beginning to consider training options in a serious way. Where did you receive your most valuable training as a Colorist? Who in the U.S. has the best formal programs in your opinion?

I've been doing this for a VERY long time. When I started, the only way to learn was to get a job in a post house, find someone who was already doing this kind of work, be willing to spend an awful lot of your own time to get access to the equipment, and hope they'll help you. Back then, we were called Telecine Operators, and it was a relatively new discipline with very few people actually doing it (probably less than 15 or 20 in all of Los Angeles some of whom are still doing it, such as Lou Levinson, Jan Yarbrough, Bob Festa, and a few others). The term "colorist" only came along in the 1990's, and mostly to justify the salaries that had gotten a lot higher than when I started. What I'm trying to say is the best way to learn anything in our business that involves a combination of creative and technical skills is still to learn it via mentoring with someone who has already achieved a level of proficiency and acceptance. That includes coloring, but it also includes editing, cinematography, sound design, music, and a host of other things. I think it's very unfortunate that the rise of the Internet has diluted that to the point where people watch a few online videos and immediately think of themselves as experienced artists. There is no real substitute - then or now - for direct interaction with an experienced operator. So if you really want to be all you feel you can be, my advice is still to try and find a way to put yourself in a position to do just that. Since I have been in the industry for a long time, I'm still pretty old school in that belief, and I really don't know of, nor do I keep track of, any "formal" programs, if indeed there are any. I think the books written by both Steve Hullfish and Alexis Von Hurkman can be valuable, as are various video training courses such as FxPHD and International Colorist Academy. So feel free to make use of any or all of those things. But understand that the best training is one on one. Always has been. Always will be.

 

[M Most]

Mike, if I may,

Saw Elysium in 4K. Yes it looked sharp enough even if uprezzed from 3.3K, and no I particularly did not love the grade, BUT what impressed me was the actual quality of the grade. It is the best looking RED footage I've seen in terms of color accuracy and stability, even if other films have had more expansive, saturated looks (which came at the expense of mildly to wildly varying tones and casts). I have to assume this precision comes from the ACES workflow? I wanted to learn more about and I headed to your blog, where your explanation made sense of course. All three stages of the process are of course self explanatory, but the first stage, where in-depth mapping, or whatever the term may be, of each camera is performed feels a bit "unachievable" to me. How can a camera/sensor be mapped out in such way that its response is know to the extent it can be anticipated under a variety of conditions? Wouldn't this require access to privileged information from the manufacturer? How would it be, or how is it, done in reality? And to what degree of accuracy and completeness? Obviously, this would be the most important part in achieving a truly universal color pipeline, the rest is just a matter of implementation and adherence to standards, but mapping out cameras feels a little unrealistic to me...

Characterizing a camera is really performing a series of tests, just like many other things. Charts are used to provide known references and measurements are taken. And yes, because nearly all cameras do internal signal manipulations, manufacturer input is usually needed. All of the major camera manufacturers keep in close touch with the Academy and all are supporters of the ACES project, likely to a greater degree than you might suspect. The combination of standardized testing procedures, transforms, and communication with the Academy serves to allow for IDT's to be developed that yield a representative ACES interpretation of that particular camera's output. One needs to keep in mind that the objective is not to make everything look the same. It is to put all cameras into a common color space that is less limited than the common current ones, allowing for processing that does not involve further transforms and does not require that each camera be treated differently in the final color pipeline. The characteristics of each camera still remain, even though they are all being put through a common rendering transform (the RRT). A Red image is still going to look different than an Alexa image because the capture characteristics are not the same.