Home of Professional Cinematography since 1996

3 perf (2)

class="Body">Some advice on 3-perf that comes from working on 3 network series.

class="Body">1. Having enough 3-perf cameras on hand is a challenge. Mixing 3 and 4 perf is a pain for post production. However, sometimes it can't be helped. For example, I had 2 3-perf cameras on a show but had a 4-perf steadicam on occasion. Also, when a camera goes down, the rental house has to fix it, not replace it. We would then get a 4-perf camera for our back-up .

class="Body">2. You can't see film dailies for lens tests because there are no 3-perf projectors. The best thing I found was to go to the transfer house and view it there digitally on a high-res monitor

class="Body">3. The AC's biggest pain is the almost ZERO TOLERANCE FOR HAIR IN THE GATE! It's very much like anamorphic that way. I would let go little bumps of dirt or hair that didn't go in TV. I suffered no repercussions from post for that.

class="Body">4. After all is said and done the advantages are great. 1000" running time is in the neighborhood of 14 minutes, 400" around 5 minutes. (my charts are at work) Thus fewer reloads. Very helpful at magic hour and when the actor is in a tense scene.

class="Body">5. This dramatically lowers film cost, using 25% less stock. @ $. 50 a foot just for raw stock, let alone processing and production wins big time. That's why Paramount TV is all done 3-perf.


>I'd like to take up the idea of a new route for film aquisition.

class="Body">Since the DI process can handle 3perf without any "trade off's" it is becoming a viable alternative to s16 and HD. You can use faster film than s16 and save on the lighting budget/time and get a better result than HD. One argument is that it is complex to produce dailies on film or to project it. Now if dailies are done to a compressed (low cost) HD format like DvcProHD and projected they would represent the material more precicely than optical rushes. A "pos conform" can be done from the rushes as well.

class="Body">Some productions are following this route allready (The Human Stain + 2 more, Lakeshore ent. LA). Conspiracy of Silence was shot 3perf, but without HD rushes thugh.

class="Body">Any thoughts?

class="Body">Kris, DFL.


class="Body">We too believe that 3perf has a great future for film origination vs HD video. But considering that all Kinoton projectors can run 3Perf prints at the flip of a switch, it's no longer a 'passage obligŽ' to make an SD or HD tranfer, contact printed rushes save precious time and money.

class="Body"> http://www.aaton.com/about/preferences/3perfandscope.html

class="Body">--Jean-Pierre Beauviala/Aaton

class="Body">How about MaxiVision 3perf at 24fps?

class="Body">Steven Poster ASC

class="Body">On the link to the printing chain on the Aaton link, pertaining to 3 perf, at the top of the page the filmstrip on the right indicates a 3perf super 35 1.85 ratio image.

class="Body">At the bottom of the page,the optical printing part of the printing chain, shows an unsqueezed super35 1.85 image on the interpositive but the internegative and the print appear to be squeezed anamormic elements.

class="Body">Why would anyone squeeze a 1.85 ratio ?

class="Body">Is there a mistake in there? if the 1.85 IP is recropped to 2.35 then converted, it does makes sense, but then you'd still be looking at 1.85 rushes on the Kinoton

class="Body">Can you shine some light on this?


class="Body">Francois Miron

class="Body">Filmgrafix Productions/Montreal

class="Body">Rune Ericson* was a strong advocate of 3Perf for 24fps cinema theater distribution, arguing that projector modification costs would be quicly paid for by the distributors themselves on savings 3Perf prints would bring them. (that was long before the Maxivision guys patented the format invented by the Goskino in 1961... ;-)

class="Body">But the 3Perf release concept never took off because of the resistance and flames against it from some big processing labs and film manufacturers. (one day I will tell you how in 1986 I desperatly tried to convince a Rochester tech V.P. to adopt a 4Perf/3perf compatible edge number barcode... with sub-zero success).

class="Body">And who would now have the clout to shake-out the print distribution chain considering that 4K telecine/scanners are almost there and that 4K Digital projectors are around the corner?

class="Body">On the other hand -origination side- I hope that film manufacturers will be smart enough to promote the 3Perf-Super35 format vs HD24 since there is no big revolution here: all modern 35mm cameras offer this pull-down option (nothing else to modify); most recent SD and all HD telecines are 3Perf-able within seconds; keycode readers, editing machines, cut list softwares etc. all can handle 3Perf aftermaths.

class="Body">* Rune wrote an article in an American Cinematographer issue circa 1982/1983

class="Body">Jean-Pierre Beauviala/Aaton

class="Body">However, I still think there is a mistake in the 3 perf Aaton printing diagram on their webpage, they talk about 1.85 super35, in the flowchart at the bottom, the IP, IN conversion shows an anamorphic squeeze on the IN at the 4 perf conversion stage

class="Body">On that same flow chart we see the advantage of looking at 1.85 3 perf rushes on theKinoton

class="Body">Why would you squeeze 1.85? that does not make sense, you squeeze 2.35

class="Body">Am I the only one that sees this?

class="Body">Francois Miron

class="Body" Filmgrafix/Montreal


class="Body">Yes, it's sort of a mistake. The 3-perf Super-35 frame, which is roughly 1.85 : 1, should end up as a hard matted (i.e. letterboxed) widescreen image when converted to 4-perf 35mm for standard matted 1.85 projection (where the projection matte would hide the hard-matting in the image).

class="Body">The diagram instead shows (vaguely) a final 4-perf 35mm frame which is not hard matted, but where the original 1.85 image fills the 4-perf frame vertically, but still has a soundtrack area on the left. In other words, an anamorphic (scope) image, not a hard matted 1.85 spherical image.

class="Body">The diagram's final image would only happen if you composed the 3-perf Super-35 frame for cropping to 2.35 : 1, and then cropped & stretched that to 4-perf 35mm anamorphic.

class="Body">David Mullen

class="Body">Cinematographer / L.A.

class="Body">3 Perf post production

class="Body">I have a related 3-perf question. I recently worked on a couple of pilots that were shooting for tv and protecting for 16-9. Both shot in 3-perf. If you're working for television where you wouldn't have to worry about a blow up for projection, is there a down side to 3-perf? And secondly, is the portion of the frame that goes to 1.33 the same size in super 35 3-perf and in standard 35 4 perf?

class="Body">Marty Mullin

class="Body">Yes, but minor. You have to be that much more vigilant about gate checks (similar to 16mm, because the frame lines are so narrow), and you have less latitude in post for reframing. Of course, this is sometimes a moot point because if you're protecting for 16:9, you have no room left on the sides in 4 perf, either. A while back, the movements were a bit noisier as well, but this has largely been addressed.

class="Body" >And secondly, is the portion of the frame that goes to 1.33 the same size in super 35 >3-perf and in standard 35 4 perf?

class="Body">If you're shooting Academy 4 perf and protecting for 16:9, the 1.33 extraction is actually a smaller image area than with 3 perf. Of course, it's unlikely you would shoot Academy for 16:9, and when both the 3 perf and 4 perf are set up with S35 centering, the 1.33 image area is almost identical. So I guess I changed the question a bit, but I think this is the answer you're actually looking for.

class="Body">Mike Most

class="Body">VFX Supervisor

class="Body">IATSE Local 600

class="Body">The image sizes are as follows (biggest to smallest):

class="Body">Super 35 4-perf 1.33:1 Big TV

class="Body">.912 x .684 = .624 square inches

class="Body">Normal 35 4-perf 1.33:1 TV

class="Body">.792 x .594 = .471 square inches

class="Body">Super 35 3-perf 1.33:1 TV

class="Body">.680 x .511 = .347 square inches

class="Body" Normal 35 3-perf 1.33:1 TV

680 x .511 = .347 square inches

class="Body">As you can see, the image area for Normal 35 4-perf TV is about 36% larger than 3-perf TV. And Super 35 Big TV is about 80% larger than 3-perf TV.

class="Body">It's also interesting to note that the image area for 3-perf TV is the same whether you're shooting Normal 35 or Super 35 because the height of the image does not change. The difference is that you get more room on the sides in Super 35. It's much like the comparison between Normal 16 and Super 16.

class="Body">Samuelson's Hands On Manual contains a lot of useful information about film formats.

class="Body">D.A. Oldis


class="Body">That's only true if you're not protecting for 16:9 using a common center approach, which I believe the poster mentioned he was. Once you're locked into a 16:9 frame, the additional image area in 4 perf is an "overscan" area, useful only for vertical reframing.

class="Body" Mike Most

class="Body" VFX Supervisor

class="Body" IATSE Local 600

class="Body">Oops! Looks like I had a little typo. The above data was for 4-perf rather than 3-perf. The comparison should have read:

class="Body">Normal 35 4-perf 1.33:1 TV (extracted from 1.78 rather than Academy)

class="Body">.595 x .446 = .265 square inches

class="Body">That's about 23% smaller than:

class="Body">Super 35 3-perf 1.33:1 TV (extracted from 1.78)

class="Body">.680 x .511 = .347 square inches

class="Body">D.A. Oldis


class="Body">Perhaps I didn't fully understand your question the first time I responded.

class="Body">Obviously if you were shooting Super 35 3-perf 1.78:1, a 1.33 image would be extracted from within the 1.78 image. That's because the 1.78 image is already using the available vertical negative area, so the only way you could obtain a 1.33 image in 3-perf would be to crop the sides.

class="Body">In my earlier posting about image sizes my assumption was that if you were instead going to shoot Normal 35 4-perf 1.78:1, you would compose for 1.78 and protect the top and bottom of the frame for 1.33:1.

class="Body">If, however, your question referred to shooting Normal 35 4-perf 1.78, then extracting a 1.33 image by cropping the sides, never using the top and bottom of the negative, the usable image size would be as follows:

class="Body">Normal 35 3-perf 1.33:1 TV (extracted from 1.78 rather than Academy)

class="Body">.595 x .446 = .265 square inches

class="Body">That's about 23% smaller than:

class="Body">Super 35 3-perf 1.33:1 TV (extracted from 1.78)

class="Body">.680 x .511 = .347 square inches

class="Body">If you're shooting HD, Super 16, Super 35 3-perf, or any other inherently widescreen format, it makes sense to compose for the widescreen aspect ratio and extract a 1.33 image if needed. But with Normal 35 4-perf 1.78, cropping the sides to get 1.33 is a substantial waste of negative.

class="Body">For comparison's sake here's the data for:

class="Body">Normal 35 4-perf 1.33:1 TV

class="Body">.792 x .594 = .471 square inches

class="Body">D.A. Oldis


class="Body" > If you're shooting HD, Super 16, Super 35 3-perf, or any other inherently widescreen >format, it makes sense to compose for the widescreen aspect ratio and extract a >1.33 image if needed.

class="Body">I agree, but that's exactly what's done for a very large part of current television production (set up for S35, of course). The common feeling is that in the end, it's simpler and more straightforward to use a common center approach than a common top and sides approach, even though some feel otherwise. Quite frankly, it amazes me that with the current requirement for 16:9 protection we're not all shooting 3 perf for television.

class="Body">Mike Most

class="Body">VFX Supervisor

class="Body">IATSE Local 600

class="Body" > A note on 3 perf. Whereas I believe 3 perf will become more prevalent once 16x9 is >accepted by advertisers for tv, so far there are limitations if you wish to transfer via >telecine.

class="Body">Definitely not the case in Los Angeles. Practically every facility in town has multiple rooms that will do 3 perf, on either Rank or Philips equipment. Not surprising, as we have always shot most prime time television programs on 35mm, and have been using 3 perf on and off since the mid to late 1980's. In the mid 90's, it became the preferred shooting format for multicamera film sitcoms (2000 foot mags = 25 minutes per load), and a number of single camera productions have used it as well. It seems to me that the majority of film shot for television in Australia and Europe is 16mm/S16mm, which would likely contribute to the lack of support for 3 perf 35mm. Is that not the case?

class="Body">Mike Most

class="Body">VFX Supervisor

class="Body">IATSE Local 600