Thinking about aspect ratio

[John Jaquish]

They choose it for its compositional opportunities, and yes, because it feels "cinematic", all the more so now that television has gone 1.78 so now 1.85 no longer feels different enough. And yes, it's an awkward shape to compose in, which is why few paintings were made in that aspect ratio, most opted for the more elegant 1.61 golden rectangle. But that's not necessarily a defect of 2.40 -- it's unusually long shape allows you to emphasize visual imbalance, negative space, etc. In a way, I think of it as a "modernist" shape compared to the "classical" shape of 1.66 to 1.85.

 

Great point about modernist/classical viewpoints. Another reason (or maybe the reason) I like 1.66 -- the golden rectangle.

 

Incidentally, 2.40 is close to the "silver ratio:" http://en.wikipedia.org/wiki/Silver_ratio

Edited February 12, 2015 by John Jaquish

[Dom Jaeger]

.. it's an awkward shape to compose in, which is why few paintings were made in that aspect ratio, most opted for the more elegant 1.61 golden rectangle.

 

Many painters have used the Golden Section ratio within their work, and it's a common assumption that painters also chose a Golden Rectangle frame to compose within but the framing aspect ratio most commonly used in classical painting turns out to be closer to 4:3.

 

This statistical study of over 500 paintings by various Masters found the average aspect ratio to be 1:1.34, quite far from the 1:1.618 Golden Section:

http://arxiv.org/pdf/physics/9908036v1.pdf

[Dennis Couzin]

This statistical study of over 500 paintings by various Masters found the average aspect ratio to be 1:1.34, quite far from the 1:1.618 Golden Section:

http://arxiv.org/pdf/physics/9908036v1.pdf

 

Thanks to Dom for discovering this curious paper from Romania. Does the "statistical study" hold water? Why did Mr. Olariu choose these 9 artists: Bellini, Caravaggio, Cezanne, Goya, van Gogh, Delacroix, Pallady, Rembrandt, Toulouse-Lautrec?

 

Mr. Olariu found 34 Goya paintings with average proportions 1.04 ±0.04. (The proportion is the long edge divided by the short edge. The ± must mean standard deviation.)

 

Did Goya really paint mostly square or almost square paintings? Google "Goya images" and you will see very many famous Goya paintings. Almost none of them are nearly square. Google images is a kind of random sampling weighted by popularity. What random sampling did Mr. Olariu do to find the square Goyas? The paper says:

"The paintings considered in this statistics have been selected from the specified references, where the sides of the paintings have been indicated."

Well, does this mean that Mr. Olariu "selected" the samples? Based on what? Or does it mean that he used every painting in the references that included its dimensions. The references are just two:

1. Peter B. Norton, Josph J. Esposito, The New Encyclopaedia Britannica, 15th Edition, 1995

2. Nicolas Pioch, WebMuseum Data Base

 

I looked in the second reference and found just 8 Goyas. All have their dimensions included. Their average proportion is 1.45 ±0.19.

 

The cited statistical study appears to be garbage.

 

[Leon Liang]

Is there any mathematical or artistic reason that 1.85:1 was chosen as a standard? I mean, 1.33 was the result of the original 4-perf 35mm without the soundtrack, 1.66 was close to the golden ratio, 2.2 was chosen because it was the most that could be taken from 5-perf 70mm and Technirama and 2.4 was chosen because of 2x anamorphic on Academy 35mm, and so on.

 

I just can't seem to think of any real reason why 1.85 was chosen and has lasted as a standard for so long.

[Carl Looper]

Is there any mathematical or artistic reason that 1.85:1 was chosen as a standard? I mean, 1.33 was the result of the original 4-perf 35mm without the soundtrack, 1.66 was close to the golden ratio, 2.2 was chosen because it was the most that could be taken from 5-perf 70mm and Technirama and 2.4 was chosen because of 2x anamorphic on Academy 35mm, and so on.

 

I just can't seem to think of any real reason why 1.85 was chosen and has lasted as a standard for so long.

 

From what I've read it's the ratio you get when you extend 3 perf into the soundtrack area. 3 perf Super35.

 

http://www.cinematographers.nl/FORMATS1.html

[Leon Liang]

But 1.85 was originally cropped from Vistavision (6-perf horizontal 35mm) and then cropped from 4-perf academy 35mm before the development of Super 35 (35mm without the soundtrack) in the 80s and 90s.

[Carl Looper]

But 1.85 was originally cropped from Vistavision (6-perf horizontal 35mm) and then cropped from 4-perf academy 35mm before the development of Super 35 (35mm without the soundtrack) in the 80s and 90s.

 

Yes, from what I can tell, Vistavision originally allowed anywhere between 1.66 and 2:1, in order to support theatres with different screen sizes.

 

Vistavision recommended using 1.66:1, 1.85:1 or 2.00:1

 

Looks like 1.85 was chosen as roughly half way between 1.66 and 2.0 (to the nearest "5").

 

I guess with the advent of 3 perf Super35, that kinda cemented 1.85 as a happy sweet spot.

 

C

[David Mullen ASC]

3-perf is 24.9 x 13.9mm, which is 1.79 : 1.

 

1.85 dates back to the 1950's. Studios recommended different amounts to mask 35mm 4-perf to for projection, Paramount suggested 1.66, Disney 1.75, Universal and MGM (I think) wanted 1.85, which was considered the most you could get away with... Anyway we had all of these various amounts of cropping, it just sort of settled on 1.85 over the years, probably because theaters didn't want to keep switching their mattes for different movies.

[cole t parzenn]

Why did 1.66 stick, in Europe? (Or did it?) Is it so hard to switch between 1.66 and 1.85? If 1.85 was considered the most drastic crop you could get away with, how did 3- and 2-perf 2.35 come into use?

[David Mullen ASC]

Well basically people disagreed on how much you could get away with cropping, but 2-perf came out in 1960 and was considered a low-budget option for scope, and was helped (grain-wise) by the way Technicolor did the blow-up with the dye transfer process.

 

3-perf took off in the 1990's for TV and wasn't used much for features until the 2000's.

[cole t parzenn]

So, was 1.66 used more for technical reasons than artistic, then?

[David Mullen ASC]

Not sure that question makes sense to me so I'm not sure how to answer it. 1.66 was chosen by the people who preferred it over 1.85, for whatever reasons they felt made sense to them. For the studios that proposed it, I'm sure it was more of a technical issue of avoiding over-cropping, but it may have also been an aesthetic reason to, they just preferred the look of less cropping. How much cropping is OK to use comes down partly to taste in the designers of the format.

 

Trouble with theaters switching to a 1.66 mask from a 1.85 mask is that the picture would get taller if the projector lens remained the same and perhaps the screen masking (black borders around the screen) would not allow a taller image. Some theaters have adjustable borders but some don't.

 

Paramount proposed 1.66 for 4-perf 35mm print matting, and they also came up with the least widescreen of the 1950's large negative formats, VistaVision, which has a native 1.50 : 1 full aperture. So whoever made up their board of tech advisors seemed less enthusiastic about very wide aspect ratios.

[David Mullen ASC]

I actually think 2.4:1 (or 2.35/2.39, whatever you want to call it) doesn't have the vastness and scope and/or spectacle of 2.2:1, which to my eye looks absolutely natural - but that's only my eyes.. So in my opinion 2.4 is actually too wide and filmmakers have been using it either because they've wanted something wider than 1.85 and haven't had access to 70mm, or, as Mr Hayward said, because it's a trend.

 

That's a bit of hyperbole. I can understand liking the slightly less wide shape of 70mm 2.20 : 1 but to say that 2.40 doesn't have the vastness, scope, and spectacle of 2.20 is crazy, they aren't different enough to create such widely different visual effects or reactions in a viewer.

 

I drew a 2.20 and 2.40 rectangle here:

 

Is the difference so great that one creates vastness, scope, and spectacle but the other doesn't??? A theater screen masking or curtains could easily trim 2.40 to something closer to 2.20 without most people noticing it -- I doubt that if you brought in the screen curtains by a foot or so on each side, suddenly a 2.40 movie would be radically altered and feel much more vast and spectacular.

[Leon Liang]

I know the difference is quite miniscule, but for some reason I personally feel like 2.2 just works "better" than 2.4 because of that small difference. That's just me - I think it's because the frame of 2.2 is higher than 2.4 if you keep the horizontal axis the same, and in my eyes that extra height seems to be better for framing. I repeat: that's just me.

 

They're both undeniably great aspect ratios nonetheless.

 

I mean, it's the same with 1.66 vs 1.85 - the difference between the two is the same as the difference between 2.2 and 2.4, but some people think 1.66 works better while some think 1.85 works better.

[Brian Drysdale]

I noticed that a Super 16 to 35mm blowup is noticeably less grainy when projected as 1.66 than 1.85. Although, that may involve a projection lens as well as a mask change.

[Carl Looper]

The origin of 1.66:1, like many aspect ratios, would be due to the ancient predilection for integers. The 1.66:1 ratio can be expressed in integer form as: 5:3. Just as the ratio 1.33:1 can be expressed in integer form: 4:3. The aspects would have originated in this integer form.

 

4:3, 5:3 and next comes: 6:3

 

Or when normalised, by treating all heights as 1, we get the same expressed as:

 

1.33:1, 1.66:1, 2:1

 

But lets return to the previous integer series

 

4:3, 5:3, 6:3

 

Now lets suppose that, for whatever reason, 5:3 isn't wide enough, and 6:3 is too wide. Doesn't matter the reason. We also note that the difference between 5:3 and 6:3 can be considered significant. Now the answer to this problem of one being not wide enough, and the other being too wide (for whatever reason) the answer is to choose an aspect half way between the two:

 

11:6

 

Which, when expressed in normalised form ,and "rounded" to the nearest 0.05, comes out as: 1.85:1

 

Now one might say there's not much difference between 1.66 (5:3) and 1.85, but the reason for this is that one has made this so. One has selected a ratio that is half way between 5:3 and 6:3. The difference is made smaller than the alternative difference: that between 5:3 and 6:3.

 

C

[Leon Liang]

A theater screen masking or curtains could easily trim 2.40 to something closer to 2.20 without most people noticing it -- I doubt that if you brought in the screen curtains by a foot or so on each side, suddenly a 2.40 movie would be radically altered and feel much more vast and spectacular.

Do you know of any instances of theatres actually doing this? I remember Vittorio Storaro once said in an interview that some cinemas actually bring in the curtains because they feel that 2.40 is too wide, which I find ridiculous. But Storaro was promoting Univisium, so that probably explains it.

[David Mullen ASC]

Sure I've been in theaters that cropped 2.40 all the way to 2.00 because of the curtains and screen borders, usually old second-run theaters before digital projection.

[Leon Liang]

Sure I've been in theaters that cropped 2.40 all the way to 2.00 because of the curtains and screen borders, usually old second-run theaters before digital projection.

Here in Sydney when I watched 'Boyhood' in theatres they did the opposite - they cropped it from 1.85 to 2.40 and when I watch some of the scenes in 1.85 on YouTube it seems like there's a lot of "dead space" on the top and bottom of the frame, as if they filmed it with the crop in mind. I've only seen that film once in theatres so I'm not sure if other cinemas have done the same thing.

[John Jaquish]

 

Now lets suppose that, for whatever reason, 5:3 isn't wide enough, and 6:3 is too wide. Doesn't matter the reason. We also note that the difference between 5:3 and 6:3 can be considered significant. Now the answer to this problem of one being not wide enough, and the other being too wide (for whatever reason) the answer is to choose an aspect half way between the two:

 

11:6

 

Which, when expressed in normalised form ,and "rounded" to the nearest 0.05, comes out as: 1.85:1

 

And of course, incidentally, the integer ratio between 8:6 (1:1.33) and 10:6 (1:1.66) would be 9:6 or 3:2 or 1:1.50.

[Carl Looper]

Origin of 16:9

 

If we treat the height as 9, then this allows the difference between 5:3 and 6:3 to be subdivided into 3 parts (instead of 2):

 

15:9 (5:3)

16:9

17:9

18:9 (6:3)

 

C

[cole t parzenn]

deleted

Edited February 15, 2015 by cole t parzenn

[Dennis Couzin]

Can you imagine a museum in which all the paintings are between 1.78:1 and 2.4:1? That museum would offend anyone's love of pictures. There can be a room in the museum where all the paintings have similar format, just as there were editions of woodcuts all one size and shape, but a pictorial medium demands freedom of shape.

To call a spade a spade: cinema is at an aesthetic disadvantage because its underlying technologies -- of shooting and display -- limit pictures to a few aspect ratios. And cinema has painted itself deeper into the corner with its embrace of wide-screen.

Painters have always enjoyed composing for wide-format. The painting is best wide when the subject is wide. A famous example is Leonardo's "Last Supper". Thirteen men (almost) abreast suggest the painting's 1.91:1 proportions. But Leonardo wasn't straitjacketed into that format. Most of his paintings have vertical formats.

Thus cinema wide-screen leads to paucity of subject matter or, just as bad, to forced composition.

Another famous wide painting is Picasso's "Guernica", 2.23:1. This is not a case of wide subject fitting wide format. It's in the genre of multi-action panoramas, where the different parts, while comprising a whole, are meant to be examined separately. Only the bravest experimental cinema allows multiple actions across a screen, with each action demanding our attention, and then it's one short step to multiple screens, the breaking up of wide screen cinema.

The suggestion that lively movement commends wider pictures is unconvincing. Was there not enough motion in Vertov or Leger or MacLaren or Lye in their 1.33:1 format? Humans do locomote horizontally, and our heads do rotate better than they tilt, but that's exactly why the cinematic image field doesn't need to be wide. Watch "Spacy" by Takashi Ito. It would be less dynamic (and spatial) were it shot in a wider format than its 1.33:1. "The Ideal City", one of the widest paintings of the Renaissance, is 3.54:1 and perfectly still, anti-cinematic.

David Mullen's remark (in post #16) that wide composition "allows you to emphasize visual imbalance", while hopeful, is off the mark. Visual imbalance has been explored for centuries in pictures of all shapes. Look at the intentional left-right imbalance in Dürer's vertical format engraving of "St. Jerome in his Study".

 

Eastman invented rollfilm and set the 35mm width, but it was when the Lumiere bothers perforated the film that an aspect ratio was set.

They chose 1.33:1. The choice was a bit influenced by the 35 mm width, since a much wider format would have been too grainy and a much taller format too expensive, but I suspect that the choice was mostly aesthetic. 1.33:1 proved a good compromise format for the cinema medium's start. Otherwise it would have been quickly supplanted. The later evolution toward wider cinema format had more to do with theater geometry and movie marketing than with picture aesthetics. A specially wide screen aesthetics evolved, making the practitioners in the niche feel big, when they really weren't. I just watched Greenaway's "Eisenstein In Guanajuato", 2.4:1. The cinematography is the most advanced I've ever seen, the camera in gyrations to keep the frame filled, until it can't and the frame just breaks into three. Yes, if cinema tries hard enough it might shatter the very idea of aspect ratio.

 

Alternatively, cinema can ditch the movie theaters and TV screens for newer, happier and more flexible display modes. Then cinema can finally be pictures with freedom of shape.

Edited February 16, 2015 by Dennis Couzin

[David Mullen ASC]

I was referring to the greater ability of 2.40 to emphasize negative space and imbalance compared to the other choice today for cinema, 1.85, not to every painting ever made. In my opinion, the farther the frame gets from a perfect square, the more negative space you can put to one side of the subject. If you put a close-up of an actor all the way to one side of the 1.37 Academy frame, there won't be a lot of excess space on the other side compared to doing it in 2.40.

 

Obviously that would be true as well for elongated vertical frames but that's not much of an option for cinema. And my comment was in response to the original article that started this thread, which complained about the awkwardness of composing within an extremely wide rectangle - my point is that this awkwardness can be turned into an advantage, you don't have to perfectly balance everything within the 2.40 frame.

 

David Mullen's remark (in post #16) that wide composition "allows you to emphasize visual imbalance", while hopeful, is off the mark. Visual imbalance has been explored for centuries in pictures of all shapes. Look at the intentional left-right imbalance in Dürer's vertical format engraving of "St. Jerome in his Study".

[cole t parzenn]

This all ignores theater design, video formatting, and where you're sitting, of course.