Ask Mike Most Anything

[Elsie N]

I had hoped someone who knows Mike personally would start this thread because they would be able to list his accomplishments and give those of us who only know him from this forum, a better idea of where he is coming from with his answers. He did agree to do this forum thread when he answered a poster on another thread:

i'd be willing to do it provided people don't expect answers within minutes or even hours, as I work all day and don't have a lot of ability to be spending time here. As for David, well, David and I are pretty good friends and i've actually tried to model my on line behavior on his to some degree. I haven't always succeeded....

I thought about titling the thread "Ask Mike Most Anything (about workflow), but that would have been using only part of his knowledge base because as someone who works deep within the industry, he knows many things that go beyond workflow.

To the newbie, I do not know if he will "suffer fools", like you and me. If you have thick skin, ask your question. The Q & A will find it's own level and even if we do not get our questions answered to our satisfaction, then maybe someone more knowledgeable will strike up the conversation that gets us going in the right direction.

[Elsie N]

1st question and this may be way below your pay grade.

Mike, I want to go all digital from camera to exhibition for a low-budget festival movie. No VFX work, and except for some fill lights I will be shooting using only available light (outside daytime for the most part... some twilight). I'm a jack of all trades and a master of none, but I have confidence that I will be able to do my own post work via Redcine and Adobe CS5.5 and output to a digital format that most Festivals would accept. I would welcome any thoughts on pros and cons of this undertaking.

[Elsie N]

...as someone who a few years ago also advocated a film targeted path, I would say that a number of things have happened that have changed my mind. First, color science has improved considerably, both on Red's side and on the side of labs and vendors such as Cinespace that has allowed for an inverse LUT approach for the film deliverable to be a practical and sensible path. In addition, on the practical side, film is not the primary delivery target going forward, particularly in North America, but in other parts of the world as well. So the need for the film version to be the model for all other versions is not what it used to be.

Now, on the subject of exactly what is different, well, even in the early days of digital cinema cameras we struggled with trying to fake those images into a film color pallette, because they just aren't the same. So we would use a film print preview LUT, and take the original video targeted images and use a transform to put them into "quasi Cineon" space, at least as far as the values themselves are concerned. But the colors never really worked, and we were faced with having to clean up a lot of things - cyan in the lowlights, artificial feeling flesh tones, pink highlights, that sort of thing. We were also forced to accept a film knee and shoulder that would minimize some highlights and low end detail that we actually had in the digital version, which helped the film print (which retained those highlights and lowlights because it is recorded without the preview LUT) but really hurt the digital deliverables. More importantly - and I'm speaking as a colorist now - the constant use of secondary correction to clean up the color anomalies I already mentioned has the immediate effect of destroying a lot of color purity. By that, I'm referring to the colors that result when the original colors are allowed to stand on their own and the delicate balance of all colors in the scene is based on an overall change of balance, much like the printer lights in a film lab. When you don't go in and change individual hues, saturations, and the like, the original colors are maintained and separation is achieved. This leads to more apparent detail, more apparent sharpness, more apparent depth, and more purity of color across the brightness spectrum. This is all a very long winded way of saying that the image looks richer. Now, even in past years, this was always apparent, but because of the dominance of film in the distribution chain, as well as a relative immaturity of inverse LUTs, we were compelled to go with film as the primary deliverable target. That is no longer the case. Nor is it the case that film is the only acceptable and/or most desirable look for modern motion pictures and genres. I feel that if anything has become apparent in the last 2 years or so, it's that film has a look that is the product of its components - chemistry, silver halide crystals, a negative-to-IP-to-print process, and other things. And digital cameras have their own look - similar, but different, and different from each other, just as film stocks are subtly (and sometimes not so subtly) different from each other. And one doesn't have to look like the other, especially when one (film) is becoming less and less prevalent, and the other (digital) is becoming more prevalent and is clearly the primary target going forward.

For a digitally captured production to look its best, to look like it was intended to look, and to get the most out of the camera being used, a pure digital path is a much better way to achieve that. If one feels that film is the only acceptable aesthetic, or even the most desirable one, then one should try to find a way to shoot film. Manipulating digitally captured images in a rather destructive way in order to fake it is not, to my mind, a productive exercise, especially in a world in which film is not the future. But that's only one opinion


This was posted by MMost in another thread and is what prompted this thread.

[David Mullen ASC]

I actually have a question, which is about transcoding for TV work. I remember that you (Mike) said something to the effect that since most TV shows are still being edited on AVID's, there is a transcoding step involved even when shooting ProRes on an Alexa, but is there a time or cost difference in terms of transcoding from 4K HD R3D files to whatever deliverables are needed for TV post versus transcoding from ProRes to, again, whatever is needed for TV post? All the delivery formats get a bit confusing. In the end, are most TV shows finishing to a 24P HDCAM-SR tape master, 4:2:2 or 4:4:4? Color-corrected to Rec.709?

[John Edwards]

Hi Mike and David,

Firstly, An honor to have you here and have this opportunity, thank you, I'm a big fan of your work and process.

I'm a colorist and I have a few questions for you about the RED workflows you have used with your colourists/DI ;

- Have you always graded directly from RED R3D files or have you ever worked from DPX (log) files converted from the RED RAW files?
- If you worked from 'RED' DPX, do you know what log settings were used in the conversion (Redlog, PDlog685 etc), and what LUT's were used in the grading chain.
- Did you ever monitor your RED-shot footage with a print-LUT applied or was it all P3 until you went to print and did tests?
- Who is your favourite post-supervisor and DI supervisor and why?

THANK YOU!

Thank you!

[M Most]

1st question and this may be way below your pay grade.

Mike, I want to go all digital from camera to exhibition for a low-budget festival movie. No VFX work, and except for some fill lights I will be shooting using only available light (outside daytime for the most part... some twilight). I'm a jack of all trades and a master of none, but I have confidence that I will be able to do my own post work via Redcine and Adobe CS5.5 and output to a digital format that most Festivals would accept. I would welcome any thoughts on pros and cons of this undertaking.

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.

[M Most]

I actually have a question, which is about transcoding for TV work. I remember that you (Mike) said something to the effect that since most TV shows are still being edited on AVID's, there is a transcoding step involved even when shooting ProRes on an Alexa, but is there a time or cost difference in terms of transcoding from 4K HD R3D files to whatever deliverables are needed for TV post versus transcoding from ProRes to, again, whatever is needed for TV post? All the delivery formats get a bit confusing. In the end, are most TV shows finishing to a 24P HDCAM-SR tape master, 4:2:2 or 4:4:4? Color-corrected to Rec.709?

The files that are usually needed are Avid MXF files (usually 1920x1080, DNxHD36), as the vast majority of TV shows are cut on Avid. The time to transcode varies because of different dailies approaches. Some shows do a direct transfer through a single LUT and leave it at that, although I know of a number of shows that have tried that and ultimately abandoned it because they were seeing too many inconsistencies during editorial and found that distracting when viewing cuts. Others have a DIT assign LUTs on a per scene or per shot basis and then have the DIT create the daily deliverables, although in "high end" network level series, this is not usually the case. A variation on that approach is to have the DIT create LUTs (or full color corrections), but have the processing done at a back end facility, which is the approach taken by FotoKem in L.A. and Bling in Canada. And many, if not most, shows still go through a "traditional" dailies workflow in which the files are sent to a facility and the material goes through the hands of a colorist in a traditional grading environment. All of these approaches work, but as you might suspect, they all "transcode" using different approaches, different equipment, and at different points in the process. In most cases, going from R3D to Avid MXF is not going to be much faster than real time given a single machine. Going from ProRes can often be a bit faster than that, but as you know, there is more to dailies than simply transcoding. There is color correction, double system sound synching, logging, renaming (usually the editors want scene and take, not file name), QC'ing the results, creating Avid bins, etc. All of those things have to be done regardless of what camera is being used, so the impact on the cost of the dailies process based on the camera should, at least in theory, be rather minimal. But in a facility model, many different approaches are used, some of which involve more people, and some of which involve more automation, and how those work flows are designed can and often does determine the cost factors.

As for finishing formats, nearly all network and cable shows these days are mastered to HDCam SR, usually at 23.976, some in 444, most in 422, it usually depends on the studio involved. Color correction is essentially always targeted to that environment (Rec709). Network delivery elements are usually HDCam SR at either 60i (for 1080i networks, such as CBS, NBC, USA, HBO, TNT, etc.) or 60p (for 720p networks such as ABC, Fox, ESPN, FX, etc.). As recently as last season there were still standard def deliveries in addition to HD for at least one network, but I believe this season is HD only for just about every network. There are, as with almost anything these days, variations. For instance, Warner Brothers wants shows that are shot on Red (they don't have very many at the moment, but they do have some) to be archived as 4K DPX files on LTO4 tape. We've already supplied that on a number of shows for them (3 seasons of Southland, the first season of Shameless, and the last 6 episodes of ER). And we're finally seeing the beginning of file based deliveries in specific areas (TNT promo gets ProRes files, for instance). That's a trend I believe will continue.

[M Most]

I'm a colorist and I have a few questions for you about the RED workflows you have used with your colourists/DI ;

- Have you always graded directly from RED R3D files or have you ever worked from DPX (log) files converted from the RED RAW files?

I've done both, but the majority of the time I've had the opportunity to work directly from the R3d files, so it's something I'm used to.

If you worked from 'RED' DPX, do you know what log settings were used in the conversion (Redlog, PDlog685 etc), and what LUT's were used in the grading chain.

That has changed as Red's color science has changed. At first, I usually used Rec709 gamma and Camera RGB color, because that yielded the most neutral starting point, and in the early days of Red, that was an advantage. When Redspace appeared, I tried it, but usually found that the Redspace gamma was a bit too "pushy," so I stayed with Rec709 gamma but adopted Redspace for the color matrix. RedGamma was a good improvement, much less harsh than Redspace, so a lot of productions started using that for monitoring and I adopted that when it was appropriate, but continued to use Rec709 gamma when it was more appropriate. Last year, for instance, I was doing the grading on both The Gates and Covert Affairs. On The Gates, I used Rec709 gamma almost exclusively because it was a dark show and looked better without the "push" that Redgamma provided. Covert Affairs, on the other hand, was a show that required brightness (after all, it is on USA, the "Blue Sky Network..."), and so Redgamma was a much better starting point, and was also what the cinematographer (Jamie Barber, an old friend) was using for monitoring. Soon after Covert's season ended, Redgamma2 was released, which I probably would have preferred if it had been available. This year, the situation has changed again with RedlogFilm and the Redcolor color matrix variants. I primarily work on a Baselight these days, and Baselight happens to have a very well developed tool set for working with log formats (it was originally designed primarily for film DI work, which is almost always done in Cineon log space). So for me, the combination of RedlogFilm, Redcolor (I prefer it over Redcolor2 for most of my work, but your mileage may vary), and a log based color grading system has proven to be a vast improvement over anything I had even a year ago. I recently did the pilot for the SyFy show "Alphas" using that approach, and the show Memphis Beat is essentially done that way as well (it's converted to DPX files for grading on a Lustre, but it's converted to that using RedlogFilm and Redcolor).

One thing about working with Red material is that you have to keep up. You have to be willing to embrace constant change, be willing to experiment, and be willing to customize your approach for different material depending on the nature of the images and where they're going. That's true of any format, but it's especially true with Red, as they move very quickly and that can be a bit maddening if you don't keep up. But if you do, it can be rather interesting and exciting. I've always chosen the latter.

- Did you ever monitor your RED-shot footage with a print-LUT applied or was it all P3 until you went to print and did tests?

I've done it both ways, and once again, it has been an evolution. A few years ago, any Red DI work I did went through a film targeted path (i.e., viewed through a print preview LUT) because it was the best way to ensure that what I was seeing was what the print would look like. More recently - i.e., the last 2-3 years - it's been in a P3 environment because inverse LUT color science has improved immensely, and because the digital deliverables (DCP, and all video formats) have taken on a greater importance. And that's a good thing, because as I stated in another thread, you get a "purer" result by staying in the digital space and accepting the images and their colorimetry for what they are rather than try to force them to look like another medium.

Who is your favourite post-supervisor and DI supervisor and why?

The one that understands the process and trusts that I do too.

[Bruce Allen]

Mike, I'd love to know: if you were in RED's shoes, what would you prioritize?

You're one of the RED's best friends - because you like RED productions and understand all of the new RED magic - but also deal with the pragmatic realities of the (sometimes entrenched and curmudgeonly, but also sometimes just very sensible) industry. How can RED help you win over all of the other folks ? Or should they just double down on image quality and value and eventually the haters will come around to the RED way because it is so far ahead in price / performance?

Keep the focus on R3D grading... or offer an Alexa-like onboard ProRes / DNxHD workflow as well? What would you work on most in future sensors / imaging pipelines - Colorimetry? Latitude? Sensitivity? Medium Format? Frame rate? Price? And how about that 4K laser projector and RedRay?

I'd also love to hear what you think Blackmagic, AJA, Avid and Adobe should be doing. Though if they were smart, they'd ask you privately of course! Maybe they have! I don't want to drag this thread into some kind of religious format war, so please feel free to disregard :)

Thanks again for all of your thousands of wonderful posts and good advice.

Bruce Allen
www.boacinema.com

[Harcharan Singh]

Hi Mike,
This is a great post..
We are using Assimilate Scratch for grading and following RECOLOR2 and REDFILMLOG for grading with a film LUT applied. The 3D LUT has been made by profiling the HP dreamcolor monitor and then making LUT depending on the stock it will be printed .
Do you also use a print LUT when using REDCOLOR@ and REDLOGFILM or should we grade on HP monitor color space of P3 and use an inverse LUT?
Thanks
Harcharan