The Large format look doesn't exist

Jerome

All stills are made with vintage s35 lenses in combination with a high quality expander projected unto the full 36X24mm Kinefinity MAVO Edge 8K sensor.

Large format as a look doesn’t exist, a sensor is a 2d plane. But throughout history we saw the seek for larger formats in order to increase analogue resolution by creating finer grained images by means of enlarging the image acquisition format. With the digital this is still kind of true, but in a more complex way. As an example, one can have 2.5k resolution with an ARRI Alexa classic ( a bit more but lets keep it simple), and 4k with an Alexa lf, while the pixels stay the same (same um). But since we blow up the image to the same size in the cinema the Alexa LF pixels end up to be smaller on the screen than the classic, and thus the image is finer grained.
We can talk hours about pixels but this is not the point right now

We could also simply design a sensor which remains s35 but has more pixels installed to get to the same result, however smaller pixels gather less light, and thus more noise etc, but soon in a few years these circumstances belong to the past.
So we end up with the discussion about large format (135 photo format mostly 36X24mm), vs s35 or s16.
One point that’s being made is that you can get shallower depth of view on large format lenses, because depth of view is a combined enterprise of focal length / f stop (and ofc relativity in terms of foreground and background and angle of view)and on large format one can use a longer focal length but get a wider angle of view than on a a smaller format. In terms of shallow of depth, we know that most well known cinematographers mostly shoot at f2.8-5.6.

Ironically Roger Deakins (who got most probably paid by some large format company), said that he loved the large format look and separation, while he also said he was shooting at f2.8-f4.0 and compared to s35 you can simply match this by finding a equivalent focal length with same angle of view and shoot f2.0-f2.8
For some directors who favor very wide angle of view and tiny depth of view, a lens like vista vision 18mm t1.5 might be interesting (or the newly released 15mm t1.8 by zeiss supreme), in order tp squeeze a bit more shallow depth out of these lenses in comparison with s35. But ironically from a lens designers perspective, large format and thus bigger sensor format means that you have to scale up the whole lens design in order to have such a big image circle. And if there was more momentum for the s35 format (which is already crowded with so many vintage offerings, and ofc cookes, ultra primes master primes) we would easily see these same lenses in the same size as large format but made for s35. if you can make a 18mm large format lens t1.5 you can also make a 12mm for s35 with a T1.0 rating or something and you would end up with the same angle of view and depth of view, as an example the Master prime 12 mm T1.3 gives you almost same depth of view and angle of view on s35 as the vista vision, but with much better image quality and 600 gram less weight.

So this is the most interesting part, lens design, and of course momentum, right now it makes the most sense to design large format lenses, because that’s what everyone talks about. But if you look at sweet spots for lens designers then s35 remains the golden standard, and I am 100% sure one can release every large format lens for s35 and get exact the same image quality if not better, with the same depth of view and angle of view, one could of course use a focal reducer which is a bit of a compromise optically, but even then we already see the result, same angle of view on a smaller format and same depth of view and one stop faster. You can keep on theorizing and apply the same theorem to s35 vs s16, but it seems both from a sensor size point of view and lens design point of view that things are bottlenecking if you go too big or too small, and s35 sits nicely in between.

Now the other way around: what if I want to use my s35 glass on the edge 8k 36×24mm sensor and benefit from more pixels, thus smaller pixels when projecting or viewing on a monitor? its not as easy, but after a big investment and having tested every expander in the whole wide world I found one expander that can adapt to about any s35 lens in the world (except a few weird retro focus s16 zooms) , and is optically way better than all the other expensive offerings and has a shim less backfocus control. I would say the holy grail. I used a lomo superspeed 28mm (10-28-1) mostly wide open, that’s T1.5 (f1.2) , and when expanding it becomes a 40mm and will cover large format and has a t stop of around t2.0 and a f stop of around f1.7 (f stop determines dof not t stop,). So still blazing fast and tiny dof compared to other 40mm large format lenses. the expander is superb, and I see no diminishing results, maybe there are but because I gain so much more resolution to spread this s35 image circle over a large format sensor with 8k, i get a lot more smoothness out of this lens, so I gained more than I lost. So one could rehouse this lens and include a expander and sell it as a t2.0 40mm lens, with unique characteristics, and sharper than most low budget offerings like dzo or meike etc.
eventually I want to use this lens as a 28mm t1.5 on its intended image format, but right now i get more out of it projecting it on this 8k large format sensor, but it shows that large format is only a temporary hype because s35 sensor tech is not yet there to life up to large format in terms of sensitivity and mega pixels, but in a few years this might be different and we will hopefully all migrate back to smaller sensors.
Smaller sensors also means smaller camera, and lower power consumption.
Conclusion, one can get exact the same depth and angle of view on super 35 format as on large format or 645 medium format. It’s even more extreme : it would be easier to achieve this on s35 compared to large format.
Many photo lenses made for 135 and used on large format have the disclaimer: corners only sharpen up around f5.6-f8.0 and also Corner illumination is drastically reduced when shooting wide open, which basically means that they are s35 lenses with a generous image circle. 🙂

Kinoliker

Jerome I agree that you don’t get a shallower depth of field inherently from shooting a larger format. But where I partially disagree is the title of the post, as if one were to boil the large format look purely to shallow depth of field, then yeah that’s dumb, a lens is a lens, you’re not going to get objectively shallower depth of field as the quality of blur is the same, but to me, the large format look is more about perspective distortion from the combination of being closer to a subject and longer focal length lenses used to achieve similar field of views to shorter focal lengths on smaller sensors at similar distances which can give the perceived effect of a shallower depth of field when it’s not. Also what the 35mm equivalent model misses is the effect of longer focal lengths on an image relative to equivalent field of views, here’s an unscientific example with an 85mm 1.4 lens and a 400mm 2.8 lens with similar field of views. And could you tell me what expander you were using?

iiiNelson

To play devils advocate… and I responded with this on your Facebook post but I’ll place it here as well.

Doesn’t physics dictate that once depth of field is “normalized” across sensor sizes that the optics eventually become nearly the same size given equal flange distance and aperture sizes? There are examples of smaller lenses that can cover a FF sensor without issue on the photo side and extremely large ones like the Zeiss Otis or most Sigma Art based lenses. One of the reasons that FF lenses on the photo sizes have began to get larger is to optimize them so that the projection is much larger than needed to get optimal performance and avoid the edges of the lens where performance is worse. Consequently this is why many modern lens design projections will also cover the cropped medium format sensors. In some ways the cinema side is doing the same sense there is no agreed upon digital standard for what is FF/LF. Whether one wants to use 36×24 (which ironically is considered FF/small format in the photo world and not to be confused with half frame 18×24), or go with the Arri LF sizes or the Red VV or the Alexa 65 sizes… there’s broad terminology for “LF.” So essentially manufacturers have taken a deliberate step to make lenses cover more than they need to and the byproduct is that modern lenses are larger than they need to be. We see more lenses advertising image circle projections of 46.5mm where traditional FF requires about 43.3mm. We see even with S35 that companies are pushing the image projections out to over 31mm when traditional S35 would be covered with a little over 28mm I believe.

Now if this trend continues, remains to be seen for larger sensors when/if computer aided designs begin to enter the cinema world to electronically correct optical flaws. In any case there’s a market for both FF and S35 sized imaging sensors. Pixel pitch is a thing as is imaging area. Sensor generations are a thing too in a way that modern Micro 4/3 sensors probably outperform FF sensors from 10 years ago in both light gathering sensitivity and color retention. People get too caught up in the idea of shallow DoF and that’s fine but light gathering and larger sensors are the best reasons to choose a larger sensor - given that we are talking about like for like generationally. Generationally, all things being equal, one can push a larger sensor a bit farther and have a cleaner image with less noise - if that matters. Also businesses are in it to make money and there seems to be a huge growth opportunity in a largely shrinking (by the numbers) camera market.

So I don’t care either way… I do t see myself buying a S35 camera ever but. Can see myself cropping my Edge 8K to the 6K or 4K S35 mode in order to use certain lenses if the mood strikes me.

Jerome

iiiNelson just a thought experiment. Take a 50mm large format lens amd put a focal reducer behind it 0.7x. now the same lens is about a 35mm with a s35 image circle, and one f stop wider open + one t stop. So exact same angle of view as on a lf camera and sane depth of view, and twice the amount of light to compensate for the smaller pixels, so you can shoot on lower iso that matches smaller photosites. Now once your done with the thought experiment, let me know. Also think the opposite way; What happens when you put an expander on a s35 lens in order to match a lf camera? these are the things that matter. and this is how many lenses are designed, for instance the ultra primes, have base lenses, like the 28mm and are optically expanded to longer end (32,40, 50 are same base as 28 but extended , 28 is limited in f stop to make 32, 40, 50 still same t stop). expanders and reducer sits in a lot of high end cinema glass already. can show you many optical designs, also present in the beautiful book by holben.

iiiNelson

So I don’t own any focal reducers/expanders personally… and I don’t own APS-C/S35 lenses any longer either… and honestly haven’t for a long time. I utilize a combination of hybrid cameras and the Edge 8K so I’m coming at this from a point of FF shooting. It’s not to say that I have anything against anyone that wants to shoot in S35… I know plenty of people that happily choose Fuji X system in photo/video for this (of many reasons for balance) but the point of the physics remains that when talking about resolving like for like and matching or normalizing aperture/iris for equivalent depth of field that the S35 lens is going to grow in size. Now one can argue that perhaps the happy medium is to create highly corrected f/2.8-5.6 FF lenses if size is a concern…. But when they start normalizing F/1 or larger aperture S35 lenses then the size differences won’t be all that much different.

Jerome

It’s all about scale. If you scale up a sensor one need to scale up the lens design. Photoglass for 135 was always a compromise because the target was a small physical form factor and low weight. Most 135 glass sharpens up in corners or has normal corner illumination when stopping down. Many 135 photoglass was being rehoused for that reason for s35. Lens design is about physics. If you can make a 645 f2.8 lens, you can make a 135 lens with f2.0 and s35 f1.4. Having the same equivalent focal length with matching angle if view. These are hard facts. Reason we wanted to have bigger formats was because of grain. Now with the digital we can almost make s35 sensors that are as good as lf sensors. Once the differences are absent (looking forward to the new arri camera) we will see s35 becoming mainstream again.
About expanders and reducers : an amazing amount of lens designs have such optics build in. I can take a master prime and put an expander (musashi are really good but super expensive) and turn it into a T2.0 large format lens. If you want a technical good t1.4 large format lens and kind of beat the dof on existing s35 lenses you will end up with a very heavy and very expensive lens. Scaling up an image circle always comes at the expensive of costs and weight or else compromises were made like the vista visions which are very mediocre wide open and still super heavy to be carrying around.

iiiNelson

Ehhh… I’m much more excited about the next Arri Mini LF Camera… I don’t care about the S35 at all personally.

Jerome

iiiNelson I am wondering if there is a good reason to be more excited about a lf camera than a s35. Many many s35 lenses exist and about any lf lens performs better on s35 than lf. Also lf cameras consume more power and are bigger in size 🙂. The reason to be excited about a new s35 arri camera os that because the only reason they designed the large format camera was to get past netflix certification. Now if they can make a s35 camera with superb specs it will be very popular. They already made a 65 camera and the lf was only rushed out after the sudden burst out of 4k as bottom line for acquisition

iiiNelson

Jerome because I like a consistent sensor size between my photo and cinema cameras for me. Also on the wide end there are just better option IMO for FF cameras. Like personally I like to have at least a 21 on the wide end and with S35 I’d need a 14 which exist but wide angles are hit or miss.

You are correct that there are more S35 lens options available but there are a lot of great FF lenses being developed… so again. I like the idea of what the next Arri LF could bring hence my excitement there.

Martin

Who doesn’t enjoy working a lighter rig. (not too light though) I agree with your points. Also the fact that all this nice glass is available, to discard it over market trend is bananas and kinda sad in my opinion. But to your point Jerome, having the option to migrate from one lens format to the other is interesting. And who knows, in the future we might be swapping sensors out of our cameras as well, just like OLPFs.

I would be curious to know more about the expander you really like. I know Laowa recently released a 1.4X expander at an affordable price, Kipon has a 1.7X option as well. I’ve been using the Kipon 0.7X for Mamiya lenses on s35 Epicdragon sensor recently for a project with beautiful results. The amount of light that thing lets onto the sensor is impressive!

Jerome

iiiNelson thats fair enough, I feel that in a certain way we agree as much as we disagree, which is nice, and yes as you know i love the versatility of large format, on the edge i can simply shoot 6k s35 and 8k lf 🙂

Jerome

Kinoliker hey welcome on our forum. Perspective is not a proprietary effect of a lens, instead a lens has nothing to do with perspective. Perspective comes from the relative distance from the camera to the object etc. A lens+ camera combination can only crop, any lens + camera combination has a so called angle if few measured in degrees for vertical size or horizontal size .
I guess you mean that if the magnification of a lens is less you are likely to go closer to the subject to get the intended magnification of the object (framing). You could do the same thing on a smaller format like s35, take a wider lens and go closer to your subject and Perspective changes. Perspective is not a camera related subject. A lens cannot change Perspective, it can only change certain things like bending light (barrell or pincushion), or other defects like chromatic abberations, and depth of view as a combination of f stop and focal length and distance to object in focus. A lens is converting light rays and reproducing these by focusing them on a canvas. So before we overcomplicated lenses we first have to consider what a lens effect is and what’s not . I recommend when doing a lens test to keep the distance from object to lens at the same distance so one can study the lens effect, otherwise we study things that lay outside of the capabilities of a lens, which is interesting too, but Perspective should never be part of a technical proprietary lens conversation. Hope this helps.
My point is that you could take a s35 lens which has the same angle of view as a large format lens. Put it at the same given f stop as the large format lens, the s35 lens will have deeper focus, but technically from a lens design perspective one could make that s35 lens a stop faster (f stop) when keeping the Size and weight the same as the equivalent large format lens, in that case the dof (depth of view) would be the same when matching angle of view of both lens+ respective camera format combinations.