S16mm DI options?

[Michael Collier]

I am looking to produce a 10-20 minute short adaptation of a feature length that I want to shoot in S16mm.

 

First question I couldnt get a decent answer to is just how good will it look if I scan the negative into 2K uncompressed and the laser scan that to 35mm? I have heard everything from having the apperance of slightly grainy 35mm footage to Clerks. (I am planning on shooting Vision2 500T stock)

 

next I am planning on shooting between 4,000ft to 8,000ft (10:1 shooting) This obviously would cost a lot to telecine out in 2K and edit online. Is it economical to just get a mini-DV work telecine if timecode is kept in tact (I plan to use an Aanton camera system with timecode). Afterwards can I hand-splice the film (from the negative or positive print?) for full 2K telecine or does it have to be scanned by the reel? My ultamate goal is to have a 4:4:4 2K 10bit file that I can do modest color work in Adobe After Effects

 

After color work I want to be able to conform the 2K scan with the miniDV work-cut and render out again to uncompressed 10bit to be scanned to 35mm (also DVDs, HDCAM masters etc will be output as well)

 

Will this workflow work? What do I need to watch out for? Does anyone know of a film-scan lab that can do S16 to an uncompressed 2K scan?

 

I am also worried about getting a telecine with NTSC colorspace. I am looking for a full 2K DI. anyone who has done this before if you could help it would be much appriciated

[Eric Steelberg ASC]

Congratulations! I look forward to watching it soon, hopefully - is it getting distribution in England?

 

 

Your best best is to do a one light transfer to SD (dvcam for instance) and edit that. Afterwards, go back with your EDL and do an HD telecine to HDCam SR (4:4:4 10 bit). No point in going to 2k...only about a 100 pixel difference.

 

Doing a full 2k scan seems to be overkill. I wouldn't worry about the color space so much because it'll get converted prior to your filmout in the linear-log conversion.

 

You could just do one flat 'neutral' transfer to SR instead of dailies and have that be it and do your color in the computer or post facility. Just keep in mind the sotrage space and computing power needed to so all this at that resolution.

 

There are many different workflows that can be used. The best thing would be to consult with the post facility who will be doing your filmout.

[David Mullen ASC]

You can transfer selects to HDCAM-SR or 2K using the EDL, from the camera rolls, and do an autoconform to create an edited digital master. Or you can transfer cut S16 neg but I suggest it be cut with generous frame handles on each side of the actual footage, and then digital edit out the handles.

 

Grain in the digital blow-up should be similar to the grain in the original; one of the advantages of doing a digital blow-up is that you aren't adding MORE grain because of duping. But you still have the original S16 grain of the 500T stock; while it is possible to digitally "degrain" the image, this can lead to artifacts because the image will probably have to be electronically sharpened after it has been degrained, otherwise it would look soft. So the end result may look video-ish if you are not careful.

[Michael Collier]

Doing a full 2k scan seems to be overkill. I wouldn't worry about the color space so much because it'll get converted prior to your filmout in the linear-log conversion.

 

Im not concerned about those extra 100 pixels, I'm worried about the compression induced by going to an HD format. They are set up for NTSC colorspace. I am well-familiar with NTSC colorspace (I have been knee-deep in it since I was 13.) Even if there are LUTs to aproximate the same colorspace, I am still loosing data in the shadows and hightlights, IE it will look more like video than I am willing to accept. If I can handcut the neg from my EDL I can then telecine out only 5% extra.

 

David, thanks for the advice about keeping the handels long. I usually do that when autoconforming a project to save hard drive space, so I was planning on doing a 1-3 second overscan in the 2K scan. I always tweek projects hours before sending them off, so a little play room helps.

 

Oh, and FYI this project, while funded will keep in my production workflow. My company (Random-Acronym) doesn not outsource work we can do. Right now we have enough time to do the post in house (I started with editing and ENG shooting when I was 13, so these proccesses are second nature to me. Data-centric workflow developed as I developed as a film-maker) So once scaned it will go to hard drive (not to an HDCAM-SR deck, useless unless you are shooting a cine-alta or viper, IMHO) and imported directly to adobe premiere, where it will be conformed to the DV and then imported into Aftereffects for colorwork (I have always prefered after effects to any other software, I have tried others and if you know what you are doing its all the same)

 

Once this is done the resultant 10bit uncompressed file will need scanning to laser printer. I plan to do no degraining as David mentioned was possible. No offense, but as I see it that will introduce the video look into what I want to look almost photokem (reason for not doing photokem, I mean come on I am a product of the digital revolution. I would loathe seting up opticle disolves, syncing magnetic sound tracks to film and working with IP and IN workflows. I need instant control over very fine points of my image, and need to see the final output very very quickly.)

 

 

My mindest for the production (Because I have worked in SD and HD, but not much in film, I think a certain way) I want to maintain data from the beginning. Film is a way to capture, but nothing else. All I want from film is the exposure latitude, the color rendition and the resolution that everone expects to see in a fullscreen movie. I would love to be a rodriguezz (sp?) and do everything digital, but I am a realist. With my experience shooting HD and my experience in the theater I know I need film.

 

The other modivation for this project is two-fold. And Mr. Mullen I look to you to see what you think of this idea, I respect your advice in these forums more than anyother (partly because I have enjoyed the look of all your films I have seen to date, but mostly because you exibit the kind of encycopedic knowledge that I hope to have one day, if you have an opening for a protoge (sp again) let me know) First I have extensive demo reel covering dialouge shots, landscape, still, spot news, and extreme shooting but have very little film to back it up. I have good knoweldge applicable to any medium (including extensive knowledge of film and film equipment I have only read about but never used) and want to have a piece of film that shows that my abilities are not limited to any one format. Also the guy who wrote this short really wrote it as a feature length. The idea is to produce the short for 7500-10K and use that at festivals to try and find funding for a 3-5mil feature project, ala Napolean Dynamite.

 

so obviously I want a 16mm or 35mm reel that not only demonstrates my commitment to the frames I create, but I want a short that will help forward my career a little bit (also I am making my way from Anchorage, Alaska,where I am a fairly big fish in a little pond to Orange County, CA where I will be hard to pick out of a crowd of students and ametures. I want to stand out and show my eye, because film is all I have thought about since I was 11)

[Michael Most]

Im not concerned about those extra 100 pixels, I'm worried about the compression induced by going to an HD format. They are set up for NTSC colorspace.

 

HDCam SR, in 4:4:4 mode, supports RGB colorspace. There is no technical difference between the information on an SR tape, in 4:4:4 RGB mode, and a scan made at the same resolution and encoded to a Cineon or DPX file. There is no color space conversion going on at all.

 

So once scaned it will go to hard drive (not to an HDCAM-SR deck, useless unless you are shooting a cine-alta or viper, IMHO) and imported directly to adobe premiere, where it will be conformed to the DV and then imported into Aftereffects for colorwork (I have always prefered after effects to any other software, I have tried others and if you know what you are doing its all the same)

 

It is not "useless" because telecine, a real time process, is likely to cost you quite a bit less than scanning, a per frame process (unless it's done on a Spirit 2K or 4K). As far as color correction, if you "know what you are doing," you're not using desktop compositing programs to color correct, especially when you're dealing with film files. You're using devices that are built for the purpose, such as DaVinci, Pogle, or at the desktop level, Lustre or Final Touch, along with proper lookup tables to yield a reasonable preview of the film result. If you want a "professional" result, you should consider letting "professionals" help you in the areas that are not your concentration. If you are a good shooter, that's great. But I know of very, very few people who are both great shooters and great colorists. If you're putting together a reel, you should consider having someone help with some of the finishing steps who does that for a living. The result will likely be far better than what you can achieve on your own using After Effects on a desktop PC with an uncalibrated monitor.

[Michael Collier]

Thanks for the info on HDCAM SR. I had heard things about it being 4:2:2 in NTSC colorspace, but I suppose thats just one option of HDCAM

 

My reason for not wanting to go to HDCAM is I dont want to rent an HDCAM Deck to fly the data in my computer. I am planning to use a Black-Magic codec (10bit 4:4:4 lossless compression) so that I can edit at home

 

Why does everyone assume color is so difficult? Yes traditional shooters left that up to a colorist, but I think in the modern era there is no reason a DP cannot color his own material. I started when I was 13 on video and have color corrected EVERYTHING since then. My plan is to calibrate the monitor with a digital eye (similar to what a colorist would do) then I will see RGB data with accuracy. I just have to get the RGB to where I want then let the printout handle the LUT. I am not talking about a huge coloring job, just adjust the contrast of the image and reduce saturation to get the look im going for.

 

Why should I pay a colorist to do that? A daVinci provides a few tools After Effects does not have, but in general After Effects will provide a great result. I am from the digital, rodriguezz era. I am simply trying to conform my old workflow to a film one, until HD advances enough to drop film.

 

Also for the most part I am interested in data-reduction coloring. Contrast and Desat both loose information, so digital grain isnt much of a concern (unless I push the contrast too much) and the film out will be a low-grain stock aimed to look identicle to the RGB i will be coloring from.

 

Does anyone in here have any real experience with this? I know its possible, I know it will look good, I am looking for tips and advice to make this work flawlessly, not naysayers.

[Andy Sparaco SOC]

Michael:

 

I think if you look at Discreet Combustion you will find a much more capable and powerful set of color correction tools then those in AFE. Combustion is the same technology in the Flame/Fire/Smoke discreet products which is also closely related to the Discreet Lustre product.

 

In regards to the colorist to each his own but there is a learning curve you will never match while at home . A full blown Xfer suite is a powerful image manipulation cave. You will have far fewer mis-steps if you benefit your project with the talent of an expeirenced colorist. You will get better results faster and they will be cleaner in the end.

 

When is the last time you filled a cavity in your own tooth?

 

Ihave been working with AFE since it was Cosa AFE in 1992 and it is a great product but you have the to resist the impulse to try and do it all your self. You will find it will be more cost effective in the end.

 

Re-inventing the wheel is expensive and time-consuming. I know cause I have a couple-penny wise pound foolish

[Michael Most]

I am from the digital, rodriguezz era.

 

Well then, you might consider that your hero Robert Rodriguez has never done post production color correction on his own, ever. All of his pictures have been finished in properly equipped DI theaters by highly professional colorists. Robert actually does admit to some limitations, and enlists help where needed.

[Michael Collier]

What do I loose if I try it myself and fail? The scan is still on HDD, so I can still take to a post house to clean up my mistakes or entire start over. If I get it right tho I save thousands of dollars and move closer to the workflow I would like to use for all my films. If in 5-10 years I not only get to DP the projects, but also do the colorwork I would feel like I have more control over the image. Even if it were a monitored session I would be dictating to the colorist what I want. Is it wrong to have a strong idea of what you want and the experience to get it yourself?

 

The point is im not reinventing anything. I am doing what I have always done. Yes color correction is complicated, but what I am trying to get accross to you is that I have been there before several times. I color everything I do. In HD you have to correct keeping in mind the losses a CRT monitor will put on the image, that I'm confortable with. But with a 2K or HD uncompressed file all I have to do is apply the LUT for the release stock I plan to use. I know I can probably get the LUT from kodak, and if after effects doesnt have a way to apply that right off the bat (I dont know that it does) it is quite possible and actually quite easy to write a script to apply it. (I mean its not a global thing, its pixel for pixel if this then that routine)

 

The idea is that I see on my calibrated screen exactly how it would look projected. Thats what I'm most concerned about. How to get to that point.

 

Also keep in mind I want to keep it simple. I will get the aproxamate look I want in camera. the most I will do is some desat and up the contrast.

 

I have worked extensivley with combustion, but dont own a copy and never found it much better than After Effects (especially considering the relativley simple corrections I want done)

[Eric Steelberg ASC]

What do I loose if I try it myself and fail? The scan is still on HDD, so I can still take to a post house to clean up my mistakes or entire start over. If I get it right tho I save thousands of dollars and move closer to the workflow I would like to use for all my films. If in 5-10 years I not only get to DP the projects, but also do the colorwork I would feel like I have more control over the image. Even if it were a monitored session I would be dictating to the colorist what I want. Is it wrong to have a strong idea of what you want and the experience to get it yourself?

 

The point is im not reinventing anything. I am doing what I have always done. Yes color correction is complicated, but what I am trying to get accross to you is that I have been there before several times. I color everything I do. In HD you have to correct keeping in mind the losses a CRT monitor will put on the image, that I'm confortable with. But with a 2K or HD uncompressed file all I have to do is apply the LUT for the release stock I plan to use. I know I can probably get the LUT from kodak, and if after effects doesnt have a way to apply that right off the bat (I dont know that it does) it is quite possible and actually quite easy to write a script to apply it. (I mean its not a global thing, its pixel for pixel if this then that routine)

 

The idea is that I see on my calibrated screen exactly how it would look projected. Thats what I'm most concerned about. How to get to that point.

 

Also keep in mind I want to keep it simple. I will get the aproxamate look I want in camera. the most I will do is some desat and up the contrast.

 

I have worked extensivley with combustion, but dont own a copy and never found it much better than After Effects (especially considering the relativley simple corrections I want done)

 

Respectfully...

 

You don't just get a LUT from Kodak. LUTs are all proprietary to the post house you do your film out at. They're set up to the film recorders, projectors, film emulsions, and even take into account the chemistry of different labs. Then you have to make sure the lab hits the LADs. You shouldn't expect to do your own correction and that be it. At the very least you'll have to go into the lab and do some timing to get your print to match what you inended. You don't have the equipment to see on your monitor what it will look like on film.

 

And by the way SR is 4:4:4 and can use RGB color space. Furthermore, I'd be really surprised if you told me you could tell the difference between an SR HD telecine vs 2k files. Transfer it flat and you'll get Cineon type footage. Don't make your project more complicated than it needs to be.

 

Keep in mind if it was as simple as you are assuming it might be, places like EFilm wouldn't exist.

[Dominic Case]

do an HD telecine to HDCam SR (4:4:4 10 bit). No point in going to 2k...only about a 100 pixel difference

This is dangerous nonsense.

 

If it was only a matter of 100 pixels difference, there wouldn't be any advantage in going one way or the other. The truth is:-

1. 2K datacines have just 960 pixels per colour, and 1920 for luminance. Right away you can see that is a long way short of true 2k. It's not even 1k in colour.
   
2. HD is compressed. You will have artefacts working agaisnt you.
   
3. The colour space is limited, and therefore you have to make some grading decisions as you scan. This restricts your ability to do a decent digital grade after conforming, which is the whole point of going via a DI
   
4. You absolutely should worry about the colour space: converting it to log before filmout is far too late. You need to get it right on the way in, not on the way out.
   
5. Talk to a lab that uses real scanners (eg Imagica, Northlight or the new Arriscan or Oxberry) not datacines if you want true, uncompressed 2k. It will be worth your while.
   
6. By the way, 2k for super 16 is the equivlent (rougly) of 4k for 35mm: so it's the best way to go.
   

The best thing would be to consult with the post facility who will be doing your filmout.

Well you have redeemed yourself with this advice. Start with the end of the process, and plan back for the best way to get there. Use experts.

[Michael Most]

[*]2K datacines have just 960 pixels per colour, and 1920 for luminance. Right away you can see that is a long way short of true 2k. It's not even 1k in colour.

 

This was true of the original Spirit datacine and its immediate successor. I do not believe it is true of the more current Spirit 2K and Spirit 4K.

 

[*]HD is compressed. You will have artefacts working agaisnt you.

 

Technically, this might be true. Realistically, I have yet to find anyone who has found any artifacts, on any level, using HDCam SR, especially in 4:4:4 RGB mode, but even in 4:2:2 YUV. Even after multiple generational tests.

 

[*]The colour space is limited, and therefore you have to make some grading decisions as you scan. This restricts your ability to do a decent digital grade after conforming, which is the whole point of going via a DI

 

This is, once again, perhaps technically true. Once again, though, realistically it is not particularly significant in real life situations.

 

[*]You absolutely should worry about the colour space: converting it to log before filmout is far too late. You need to get it right on the way in, not on the way out.

 

You can do this using either the Spirit 2K or 4K without any real problem. They can both output 10 bit log directly, derived from either the internal 12 bit linear (on a Spirit 2K) or 16 bit (on the 4K).

 

[*]Talk to a lab that uses real scanners (eg Imagica, Northlight or the new Arriscan or Oxberry) not datacines if you want true, uncompressed 2k. It will be worth your while.

 

I don't necessarily disagree with this. However, the alternatives are not as limited as they used to be, even as recently as 6 months to a year ago.

 

I would also add that Dominic is one of the more knowledgeable people on the planet with regard to all of this, and I respect all that he says, here and elsewhere. However, I think it's important to consider the current state of the art in terms of the alternatives to all-out frame by frame scanning, especially when there are budgetary considerations involved. The alternatives are much more viable than they once were.

[Dominic Case]

I would also add that Dominic is one of the more knowledgeable people on the planet with regard to all of this, and I respect all that he says, here and elsewhere.

Well thank you Mike. :) I'll say the same for you.

 

It's true that many of the issues I mentioned may be only minor in a lot of instances (though not always, and people vary on their views of what is "minor" to them) - and every new format reduces the problems (though sometimes at the expense of other ones).

 

But I think it doesn't do any harm to bring the concerns out - especially when the original poster was adamant about wanting to go full 2K, but spoke of telecines: I think it's important to distinguish between telecine transfer and scanning: they have different purposes and so make different compromises.

 

And while the newer Spirits certainly have more pixels in their CCDs, (strictly, photosites, not pixels) it is important to know that your work will go on a machine that actually has that head fitted, if that is what you believe you are getting. Marketing doesn't always disclose the details.

[Marc Wielage]

how good will it look if I scan the negative into 2K uncompressed and the laser scan that to 35mm?

It can look great, provided the 16mm negative is shot well. My usual advice to DPs is to overexpose the negative about 1/3-1/2 a stop, to provide as thick a negative as possible, and avoid underexposure whenever possible. I think this is generally good advice for shooting for TV as well, provided you understand that the colorist can always add contrast and density when necessary. But brightening underexposed shots is often problematic at best.

 

 

Is it economical to just get a mini-DV work telecine if timecode is kept in tact (I plan to use an Aanton camera system with timecode). Afterwards can I hand-splice the film (from the negative or positive print?) for full 2K telecine or does it have to be scanned by the reel?

I'd advse that you avoid using mini-DV for post and instead use DVCam, because your timecode options are generally more flexible with the latter. I know of very few LA post houses that can handle telecine transfers going to mini-DV. DVCam is no problem.

 

Transfer all the film in standard-def, using a best-light setting, and capture all the keykode, video timecode, and audio timecode information. Have the facility imbed the visible timecode and keykode numbers into the video (preferably in the black matte area outside the picture area) as a backup. Once all that's done, do your offline edit, and then have your editing system generate a "negative cut list," providing all the keykode numbers for the specific frames needed. (This is a standard feature of the newer Final Cut Pro and Avid systems, and there are workarounds for other editing systems.) The Aaton code will only facilitate your audio syncing in telecine and has no bearing on scanning or editing per se.

 

At this point, have your post facility just do 2K scans the specific takes required, based on the keykode numbers. I generally tell people to give us at least 10-frame handles before and after each shot with 16mm, to be on the safe side. You can try to conform all of this yourself, but you'll need a lot of storage and expertise in order to do it.

 

There are pros and cons to all scanners out there. I've had experience with Imagica scanners, Northlight scanners, and Cineon scanners, and I've scanned several miles of film myself on Spirit scanners (old Spirit and the newer Spirit 4K's). It's possible to get good results from all of them. I think the Spirit 4K is hard to beat for the best combination of speed, picture quality, and economy, particularly for S16mm.

 

I would strongly advise that you avoid trying to do any color-correction yourself unless you've done lots of tests with the company doing your filmout recording. No two DI companies in LA use the exact same LUT, so the color-corrected files you create may not yield predictable results. Expect to do some tweaking at the facility before the filmout is created. Bear in mind you'll also have the chance to do some minor touch-ups at the lab if necessary for the final prints.

 

Note that although I work for Technicolor, all of the above is strictly my opinion. Your miles may vary, etc.

 

--Marc W.

[Michael Collier]

Thanks for the tips. As expected the project ballooned with more investor intrest.

 

Now its a 100minute feature with the same basic goals.

 

Film out is not a concern anymore. It will go straight to HDCAM for festival play(or HDCAM SR, assuming there are festivals out there that will accept that). I am assuming this film may not ever see a film out, given its quite an underdog, being a $30K feature shot independantly in Alaska (though the landscape will give us free production value)

 

That aleviates a lot of concerns with the film out, because I have done color correction for high def for several years. If a distrabutor wants a theatrical release, they can pay for a full 2K DI with proffessional colorist with a 35mm film out. If it ends up on HBO, I would be out a lot of money if I paid for film out, assuming that its above the TV market.

 

Now my other concern: For money reasons we have dropped shooting on the Aaton XTR in favor of a crystal sync camera that does not include timecode. What is the best way to deal with this situation? Obivously we will have to slate every take. Will it make things any easier if I get a slate with timecode LEDs and a recorder that can jam to the slates timecode? My thinking is that if I am looking at a bunch of wav files with no real names, just timecodes, It would be much quicker to find the audio clip based off of the timecode displayed, and would allow for quick syncing.

 

Also are there any tips on how to calculate stops? for the most part I have been memorizing the 1/3 stop increases my SLR shows. (160, 200, 250, 320, 400 etc) is this accurate, or is there a formula that will calculate it quicker than just memorization? This will be important to make sure I retain the highlights I want and blow-out where I need highlights to hit the upper end of the toe. (its film noir and includes several visually exciting scenes that will need to be very specific. ei darkend interogation scene, exorsism (which I am still trying to figure out how to make that look different than the exorsist and constantine.) lots of good stuff, but all needs to be very dilliberate to make sure I get my intent accross (this is not a movie I want to find in post)

 

What issues do you think there will be with a 4:1 shooting ratio? Obviously I will have to plan out EVERYTHING. Most of the film will play in wider shots as apposed to the tradition shot/reverse shot style. actors will be heavily rehearsed, but even with all those accomodations, does anyone see any obvious pitfalls to this approach?

[Michael Most]

Also are there any tips on how to calculate stops? for the most part I have been memorizing the 1/3 stop increases my SLR shows. (160, 200, 250, 320, 400 etc) is this accurate, or is there a formula that will calculate it quicker than just memorization?

 

 

Ummmmmm.....

 

You list yourself as a Director of Photography and you're asking this question??

[Michael Collier]

I'm mostly a DoP in digital/High Def. This is my first film shoot (other than still) and just want to tripple check everything I have read over the years. I have a good amount of knowledge that I aquired from reading tons of books, and experience with video, but in video if I want to see how many stops over something is, I iris down until zebra lines are on it, then I count how far that is from the exposure I am aiming for, given the stops listed on the lens barrel. Its a work around, but I have really never needed to know asa stop conversions.

 

I just want to tripple check everything u know? I ahve been shooting for going on 12 years now, ever since before I could drive people have been paying me for editting and shooting services, plus a million short and feature projects I know I can do this, I'm just paraniod about falling into what should be obvious pitfalls. These pages have been a great help and I love checking my assumptions/ideas with pros who have been with film since before video was born (after all in the new cinema era, a young DoP who can shoot both video and film would be very attractive hire.)

[K Borowski]

I'm mostly a DoP in digital/High Def. This is my first film shoot (other than still) and just want to tripple check everything I have read over the years. I have a good amount of knowledge that I aquired from reading tons of books, and experience with video, but in video if I want to see how many stops over something is, I iris down until zebra lines are on it, then I count how far that is from the exposure I am aiming for, given the stops listed on the lens barrel. Its a work around, but I have really never needed to know asa stop conversions.

 

I just want to tripple check everything u know? I ahve been shooting for going on 12 years now, ever since before I could drive people have been paying me for editting and shooting services, plus a million short and feature projects I know I can do this, I'm just paraniod about falling into what should be obvious pitfalls. These pages have been a great help and I love checking my assumptions/ideas with pros who have been with film since before video was born (after all in the new cinema era, a young DoP who can shoot both video and film would be very attractive hire.)

 

I think you need to do some more rehearsing with your SLR. You're mixing up F-stops with shutter speeds or film speeds here. F stops are numbered like so: 0.7, 1, 1.2, 1.4, 2.0, 2.8, 4.0, 5.6, 6.7, 8.0, 9.5, 11, 13, 16, 22, 32, 64. I may have missed a half-stop or two in there. I am just illustrating the point that those aren't F-stops you're talking about. The smallest F-stop (and illogically biggest number) you would realistically have to worry about in S16 would be F16, maybe 22 or 32 on a few lenses. You might be referring to a feature on some 35mm SLRs that allows you to manually override film for pushing or pulling, or just if you wish to rate it differently. Film speeds go in thirds of a stop, and sometimes this is interchanged with shortening exposure an amount comparable to stopping the lens down 1/3 of a stop. Anyway, you don't have control over ANY OF THAT in standard, 24/25 fps cinematography. The shutter speed is SET. All you have to deal with is the F-stop. There are some cameras that allow you to alter the size of the rotating movie shutter, generally done for in-camera fade-ins fade-outs in the olden days, but since movies use a slight blur in each frame to blend the frames together, closing the shutter smaller than about 1/45 of a second starts to show signs of strobing, a stuttery flicker that is normally only seen when you have light hitting the lens from the right angle at a standard shutter speed. I don't know what your camera has, but I think you are best off leaving shutter speed alone, except for affect. Since you're shooting a 500 ASA stock, you'll probably be shooting around a 5.6 or above, in a conventionally lit indoor shot. Outdoors, you'll need to shoot at F16, with a neutral density filter slapped in front of the lens. Since you're talking about either blowing this up to 35mm or digitally projecting in a theatre environment, I'd recommend you consider a finer-grained stock. The grain in S16 is fairly evident even in SD with such a high-speed film. Outdoors, use either EXR-50D or the new Vision2 50D for a very fine-grained result. The entirety of the film Pulp Fiction was shot on the slow EXR-50D stock so as to minimize the grain in the blowup. In any case, it is generally best to overexpose color neg. Some peoople will overexpose by as much as a stop (i.e. you'd shoot at F11 with a meter reading for F16). Usually I'll go for 1/2 to 2/3 stops of overexposure. Since film has about three stops of overexposure latitude and one under, it helps to put the image information right in between over and underexposure, and it also moves the image detail into the area of the exposure "curve" that has finer-grained halides. This is again good for something that you intend to project or show in HD. While you're a child of the digital "revolution", please don't feel you're beyond reading books on the subject. Get a book on cinematography. Read it, repeatedly. Watch films like Pulp Fiction, Easy Rider, probably two of the more famous films shot on 16mm. Find out how they did it and see if you either like or dislike the look of these films and what you intend to do the same or differently as they were done. Most importantly, shoot tests. You seem to sound as if you have a decent budget, so this probably already occurs to you as a matter of course, but film is not video. You have the potential to ruin every foot of film you have shot through the camera if something is out of whack, so make sure to fire off a few hundred feet of test film with your camera not only to check it out but to make sure you have your exposures where you want them, and also to make sure you have the look in the negative that you want.

 

Regards.

 

Karl Borowski

[Michael Collier]

While you're a child of the digital "revolution", please don't feel you're beyond reading books on the subject. Get a book on cinematography. Read it, repeatedly. Watch films like Pulp Fiction, Easy Rider, probably two of the more famous films shot on 16mm.

 

I have a stack of books taller than me I have been reading since I was 11 on through today. I read them every chance I get. Most are falling apart from wear. I have seen both movies and am a fan of both.

 

I think you need to do some more rehearsing with your SLR. You're mixing up F-stops with shutter speeds or film speeds here. F stops are numbered like so: 0.7, 1, 1.2, 1.4, 2.0, 2.8, 4.0, 5.6, 6.7, 8.0, 9.5, 11, 13, 16, 22, 32, 64.

 

I think you confused two parts of what I was saying. The first was I was looking at my SLR that has a feature to rate your film (It automatically scans the canister for speed, but this can be altered) it goes in 1/3 stops. The other part was explaining why as a DoP (of digital) I have yet memorized ASA values, and its relation to stops (I know how to convert in whole stops, but not in 1/3 or 1/4 or 1/2, and logrithmic math is not my best subject.)

 

The point I was making was that I have developed a trick to use my betacam ENG camera as a light meter, which I dont know logarithmic conversions. I read my cameras 'light meter' in stops, and in video its less important to light to a certain exposure level, becuase you can quickly tell how close you are by looking at the output.

 

(BTW, the trick, and maybe I havent described it well, but I have my zebra lines set to IRE 70, so if I wan to tell the exposure latitude between say the midtone and a window, I set my iris to put the zebras in something I consider midtone (usually skin tone if its a human subject) I note the F-stop its on and then Iris down until the zebras just show up in the area I want to meter. If you count the number of full stops difference on the lens barrel (I have a fujinon 7.5x14 2/3"), you get a good aproximation of how hot something will be. Its a nifty trick for all you ENG shooters out there.)

 

 

The shutter speed is SET. All you have to deal with is the F-stop. There are some cameras that allow you to alter the size of the rotating movie shutter, generally done for in-camera fade-ins fade-outs in the olden days, but since movies use a slight blur in each frame to blend the frames together, closing the shutter smaller than about 1/45 of a second starts to show signs of strobing, a stuttery flicker that is normally only seen when you have light hitting the lens from the right angle at a standard shutter speed. I don't know what your camera has, but I think you are best off leaving shutter speed alone, Since you're shooting a 500 ASA stock, you'll probably be shooting around a 5.6 or above

 

I think the camera I have is something around 172.5 degree shutter. (Halfmoon type) It does not have a variable shutter, nor does that matter to me. I do not particularly like the saving private ryan look. In my line of work shutter tends to get over-used (especially it seems in my market) with a flip of a switch most shooters in town can go between 1/60 sec to 1/2000 of a second. Add an interlaced 60(feilds) a second it looks terrible. This is a slower film than what quick shutter is good for, and I dont have access to enough light to shoot a film on such a slow film.

 

I'm not shooting 500 speed. I am shooting a 200speed fuji color stock rated at 160. Most of the film will be desaturated to a black and white in the HD finnishing post stage. parts will retain color, but even those will mostly be muted colors, exept for a few key scenes. Also it is in a film noir style, so contrast in the film will be maintained, and a bit of grain is allright (as long as its nothing close to clerks)

 

 

 

Most importantly, shoot tests.

great advice. I have budgeted 400 feet of film for tests and its the only film whos dailies will be in film (the rest will be in mini-DV or DVCPRO) Edited February 28, 2006 by Michael Collier

[Dominic Case]

I'm mostly a DoP in digital/High Def. This is my first film shoot

A stop is a stop, regardless of the capture format.

 

One stop refers to a doubling or halving of the actual exposure. On your still camera (SLR) you can control this by shutter speed as you have listed, or by aperture (which is what people mean when they say "what stop are you shooting at). As Karl said, you really only have control over aperture when you go to motion picture (though shutter angle does allow you to change exposure time, but it's not usually in order to control exposure - let's not go there yet!).

 

Since one stop is double/half the exposure, your shutter speeds will halve or double for one stop increments/decrements. As fractions of a second, that's

 

30, 60, 125, 250, 500 etc.

 

For 1/3 stops, the increment is the cube root of 2 (or 1.26)

30, 38, 48, 60, 75, 95, 125, 160, 200, 250, 320, 400, 500 etc.

 

The pattern is repetitive once you have a range from one stop to the next (just double the numbers for the next range. (The approximations jump a bit when you go from 60 to 125, as it's not exactly double.)

 

You will see that the higher range of numbers are the typical EI ratings (or ASA) for film emulsions.

 

But what you are really asking though, is the range for f/stops (or t/stops).

 

If you double the diameter of the lens aperture, you get 4 times the light. So single stop increments (for just double the exposure) go up as square roots of 2. Each stop number is 1.4 times the previous one.

 

1, 1.4, 2, 2.8, 4, 5.6, 8, 11, 16, 22 etc.

 

If you are going up in 1/3 stop increments, then the multiplier is the sixth root of 2, which is about 1.12

 

starting at f/2, it goes

 

2, 2.25, 2.5, 2.8, 3.2, 3.5, 4, 4.5, 5, 5.6, 6.3, 7, 8, 9, 10, 11, 12.5, 14.2, 16 etc

 

The numbers from one stop to the next are simply double the ones for the previous stop. So if you memorise (for example) 4, 4.5, 5, 5.6, and the full stop range, you can work out the whole series. (roughly, as we always approximate to pretend that the numbers come out rounder than they really do.

 

But if you are good with cube roots (or even sixth roots), then that is the formula you asked for.

[Michael Most]

Watch films like Pulp Fiction, Easy Rider, probably two of the more famous films shot on 16mm.

 

Pulp Fiction was shot on 35mm, anamorphic. Easy Rider was, of course, 16mm.

[David Mullen ASC]

Pulp Fiction was shot on 35mm, anamorphic. Easy Rider was, of course, 16mm.

 

"Easy Rider" was also shot in 35mm. According to the imdb, the New Orleans scenes were shot in 16mm though.

 

"Pulp Fiction" definitely -- in fact, it was shot on 35mm Panavision anamorphic almost entirely on 50 ASA 5245, about as far from 16mm as you can get short of shooting in 65mm.

[Michael Collier]

But if you are good with cube roots (or even sixth roots), then that is the formula you asked for.

 

Thanks, exactly what I was looking for. I knew some math involving stops, but mostly it focused on whole stops, and faced with calculating (with absolute reliable results) in 1/2 and 1/3 stops, I found a hole in my education. Im not very good with cube roots, but if modivated I can learn anything, so I guess I will begin lessons so that I know more than the theory behind the concept, I want to be a walking calculator (it will make things easier for me)

 

 

Pulp Fiction was shot on 35mm, anamorphic. Easy Rider was, of course, 16mm.

 

I thought it was 35mm, I saw a picture of them filming in Jack Rabbit Squares, and there was a panavision on what looked like a pee-wee. I couldnt imagine panavision making a 16mm film camera that big (I didnt know if panavision makes 16mm, with film gear im the ranking ameture.)