Welcome to our community

Be a part of something great, join today!

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

Diffusion filter to emulate Alexa on Scarlet or C300

M

Matt W.

Well-known member
Joined
Oct 3, 2009
Messages
1,350
Reaction score
1
Points
38
This is a tricky question since there's camera system subtext, but let's stay out of that. This is pure curiosity.

A camera is Alexa. Shooting at 800 ISO. To prores.

B camera is either Scarlet or C300. Shooting at 800/850 ISO. External recorder capturing prores.

The cameras are meant to match and there is time for a grade but no time for a beauty pass to emulate the Alexa's built-in halation simulating filter. I want to put a filter on the B camera to emulate the halation filter on set. Two problems present themselves:

A filter in front of the lens should be stronger at 200mm than at 18mm. The on-chip filter should not. So I might need a set of varying strengths, and this is a pain constantly switching filters. Will I need varying strengths? How can I tell what strength at 18mm seems like a different strength at 85mm or at 200mm?

I dislike filters that have large enough particles that they are visible in the bokeh (for this purpose, at least). I want a very fine-patterened filter.

Would an ultracon set work? What strengths should I buy? Thank you all for your help. Please don't discuss camera systems except tips for matching between them. Thanks.
 
Last edited:
David Mullen ASC said:
An UltraCon or DigiCon set would work fine though you didn't say whether you were shooting the Alexa and Canon in log mode.
Click to expand...

Alexa in log. Canon in c log, which is apparently "fake" log.

David, have you shot with both cameras? I have not but have a bit of experience with both in post and find them to have a similar "look" but the Alexa has much better highlight retention and a truer log curve. The Canon is harsher with a sharper, more "digital" image that is not as pleasing and its color rendition is worse. Would rating both at 800 ISO and shooting in log mode be your advice? To me the big issue is the canon loses as much as two stops (conjecture) of highlight detail and the feel is a bit less organic.


I'll probably get an ultracon set, as I am familiar with the filters. But I can't afford every strength. Maybe just a 1/,2, 1, and 3. Would those make sense for a range of lenses between 11mm and 200mm or is that much too strong?


This is mostly unrelated, but what diffusion filters is Richardson currently using? And can you stack them with ultracons or is that pointless?


Thanks!
 
Last edited:
I haven't shot with the Canon, log or not, but I assume that C-Log mode is the lowest contrast image you can record tonally, so if it's not quite as wide a range as with the Alexa, the UltraCon or DigiCon might help a little (but honestly, a filter like that may help add 1/3 of a stop of extra shadow detail but that's about it, the rest is just a lifting of the blacks.) The filter would help more in just creating a look that feels more similar.

Just get the 1, 2, 3 UltraCons, the heavier ones might do more harm than good. Just be aware how the UltraCons work, by reacting to the ambient light hitting the glass, so don't overly shade the matte box but also watch out for panning into something very bright because you will get a veiling effect. Schneider DigiCons may even be better for you since they have a black element added to their UltraCons that darkens highlight exposure slightly.

Richardson hasn't been using diffusion filters as much these days but some of "Django Unchained" used a net over the back of the lens. I think he's also used some very light Black ProMist for some shots in other movies lately, intercut with unfiltered shots.

Sure you can stack two diffusion filters if you watch out for double reflections around bright edges (which is one reason I like the simple Panavision 4x5 mattebox with the anti-reflection filter (ARF) tray... very low-tech, two filters are stacked together with no gap, and tilted slightly by use of a metal wedge in the filter slot. Since the UltraCons will soften a little, compensate by using a lighter grade of diffusion than what you were planning, and keep in mind that both may be lowering contrast. In fact, if using a "mist" type diffuser like regular ProMist or White Frost, I wouldn't bother with the UltraCon since the mist particles will be doing some contrast lowering already. But if using something like a black net as diffusion, then an UltraCon would still be working to lower contrast.

Ultimately be mild with your filtering on the Canon since you don't want to lower resolution too much compared to the Alexa. Ultimately you'll be doing most of the matching work in post, probably partly by adding some contrast to the Alexa image.
 
In another thread on a similar discussion (I'm really interested to trying low con filters but haven't been able to yet), David suggested that its very easy to go overboard on look/mist/low con filters since the strength of the effect is often missing on preview monitors/evfs or even your average display. 1 and 3 might be too strong.

Other things to consider (and again, I have yet to test) is that I believe UltraCons and DigiCons were designed to try to avoid highlight halation/bloom/glow, which is typical of the Alexa look. So you may want to test/compare before buying a few hundred dollars worth of filters.

That being said, if/when you do test, it'd be awesome if you post some .r3ds on reduser (assuming you try Scarlet.)
 
Yes, if you actually want some light halation just around a point of light, rather than just contrast lowering, then something other than UltraCons would be better -- like the lightest LowCon, White Frost, etc. But I don't think there is that much of a halation effect with the Alexa that you actually have to worry about adding a diffusion filter just to match that, the bigger issue is matching the look of Alexa's wide dynamic range.
 
Thanks! The lowcon might be it. I might try a strength of 1 to make the image flatter and soften it a bit, too. The Alexa does seem to have bloomier highlights than the Red, but it is subtle:

http://gunleik.com/tests/alexmx/RED1_MX_and_Arri_Alexa/Full_latitude_2.html

Surprised to learn Richardson is using less diffusion. I suppose I owe it to myself to see Django Unchained.
 
Matt W. said:
The cameras are meant to match and there is time for a grade but no time for a beauty pass to emulate the Alexa's built-in halation simulating filter. I want to put a filter on the B camera to emulate the halation filter on set.
Click to expand...

I think this can be done in post to some degree, using various forms of highlight defocus effects, either as plug-ins or as options in Lustre / Resolve / Baselight / etc. I believe Bob is doing most of his D.I. post over at eFilm, which is using a customized version of Lustre. His films have a lot of post color-correction work going on, which he carefully supervises.

I think the only way to get two non-identical cameras to sorta/kinda match is to do a lot of tests beforehand and see how far you can push them in post. Filtration is not the only answer or method to get there. I don't think anything will get them to match 100%, but you might be able to get more than 50% of the way there in terms of the flare you're talking about. Color is relatively simple, provided they're both well-exposed and well-balanced (and of course, the scene is well-lit).
 
Mark K. said:
On top of any filtration you decide to use, I'd strongly suggest pointing the B-cam at lower contrast portioning of your set/lighting, it'll help mask the difference in DR.
Click to expand...

I like this advice. Simple but pragmatic. Thanks!
 
From David...
"Sure you can stack two diffusion filters if you watch out for double reflections around bright edges (which is one reason I like the simple Panavision 4x5 mattebox with the anti-reflection filter (ARF) tray... very low-tech, two filters are stacked together with no gap, and tilted slightly by use of a metal wedge in the filter slot."

David...I'm intrigued by your ARF system. This problem of double reflections has been an issue with me every now and again. I know what the issue is but always thought there was some alignment issue with my matte box. Now if I hear you correctly it just may be my matte box is not out of alignment enough? Typically how far "wedged" out of alignment do you push the filters and do you have to take into account any panning or tilting during your shot? Is it just a matter of throwing the double reflections out of your shot with the wedge? I would love to hear more about your solution and whether I could apply it to my own non-panavision matte box. Thanks in advance
Taylor
 
The main reason you can avoid double reflections with the Panavision 4x5 matte box is that it has a wide slot that fits two filter trays in contact with each other... In fact there is room for three filters without the ARF wedge. ARRI matte boxes separate each tray with a metal strip... It's the gap between the glass that allows a double image to occur with a bright highlight. The ARF wedge in the Panavision matte box just allows a flare to be tilted downwards or upwards hopefully out of the actor's face. In some cases you can remove the ARF wedge and the metal trays and just tape the two filters at a more extreme angle in the slot, the tape being necessary to keep the filters from falling out of the matte box. This trick will send a flare even lower or higher out of the shot.

The problem with a gap between flat pieces of glass causing a double reflection is one reason I hardly use a Pola filter anymore since even in a Panavision matte box, the round Pola sits at the back separate from the 4x5 filters.

Of course for movies where I don't use diffusion on everything this situation is less problematic since I generally only have an ND filter in the matte box, not two filters at the same time.
 
Thanks! How often do you find yourself using diffusion and which kinds recently? I just watched Lincoln and really liked the look but it seemed less hazy than some of Kaminski's earlier work... I'm wondering if heavy diffusion is going out of style in favor of lighter grades.

For guerilla and micro-budget (Scarlet/C300/etc. with EF mount zooms) do you think there's an argument to be made for threaded filters (zooms with 77mm threads), or do they reflect light worse than when using a mattebox?
 
Diffusion in general has been going out of style since the 1980's but there have been notable exceptions -- Robert Richardson and Janusz Kaminski, for example, or some of Seamus McGarvey's movies ("Atonement", "Anna Karenina"), some of Chris Doyle's work, though like Richardson he has been using less and less diffusion since the 1990's. However, for television work, diffusion has been more popular than for theatrical features, partly to offset the harshness of digitally-shot close-ups on HDTV monitors. For theatrical presentation on a large screen, you have to be a lot more subtle with diffusion. I use a lot heavier diffusion on "Smash" than I would use for most features.

I think with the greater clarity of digital images, diffusion has made a comeback to some extent, though again, for features the use has been subtle. For television work shot digitally, diffusion has been commonplace for flattering actors, and with the short color-correction and post schedules of episodic TV shows, if you don't do it in camera, it may not get done in post. It's one thing to have one or two shots that need an effect added in post, but it's another to have every shot affected in some way when time is short in the color-correction suite.

As for saving diffusion for post, I think there are arguments both ways -- for a theatrical movie that is generally supposed to look sharp and unfiltered, then it's probably best to shoot clean and if a particular close-up is so sharp as to be unflattering, you can fix that in post once you see it in the D.I. theater. But in the case where the whole movie is supposed to look filtered, particularly when using organic and unpredictable diffusion like nets, there is an argument for doing it in-camera because the results will be more organic optically. They may even affect framing/operating, such as the scene in "Atonement" where Keira Knightly is looking into the mirror, heavily backlit, the camera shiftily slightly side to side to bring the halation in and out of frame.

Also, I've found that if all you want is a tiny halation around points of light, the lightest filter such as a 1/8 Black Frost will create that look with less affect on sharpness than digital tricks that involve blurred overlaid luminance keys, etc. I find it hard digitally in post to get that halation effect without also blurring some fine detail in bright highlights compared to a light optical filter on the camera which mostly just scatters some light right around the point of light in the frame.
 
Amazing, thank you! You should consider writing a history of cinematography textbook (granted, there is probably no market). Your insight into the process is even more valuable for your vast knowledge of cinematic history and understanding of both form and commerce.
 
David Mullen ASC said:
For television work shot digitally, diffusion has been commonplace for flattering actors, and with the short color-correction and post schedules of episodic TV shows, if you don't do it in camera, it may not get done in post. It's one thing to have one or two shots that need an effect added in post, but it's another to have every shot affected in some way when time is short in the color-correction suite.
Click to expand...
Actually, it can be done pretty fast. I had a sitcom pilot just a few months ago where the producers were very concerned about the age of the lead actress. I had one node dedicated solely to defocusing her face, and I just automatically dragged that one into every medium shot and close-up that featured her. I think that only added about one minute per shot. In a show like Smash, it could be tough simply because it's a very "cutty" show with a lot of edits. If it were an overall, and I was working in resolve, I'd do the defocus/glow as a final layer applied to the entire sequence -- not per shot. But I agree that in general, this is tough to do on a TV schedule, and for a lot of reasons, it might be best to just "bake in" the look for time and money reasons.

Also, I've found that if all you want is a tiny halation around points of light, the lightest filter such as a 1/8 Black Frost will create that look with less affect on sharpness than digital tricks that involve blurred overlaid luminance keys, etc. I find it hard digitally in post to get that halation effect without also blurring some fine detail in bright highlights compared to a light optical filter on the camera which mostly just scatters some light right around the point of light in the frame.
Click to expand...
I'm a huge fan of 1/8 and 1/4 black promist (and frost) filters, and it's definitely a different look than what we can create in post. As I've often said, there are certain things for which we don't have any knobs; Tiffen Glimmerglass is a good example. For that effect, you gotta do it in the lens.
 
You must log in or register to reply here.
Back
Top