Matte Painting.

Matte Painting foi uma matéria bem aleatória. As aulas começaram muito bem, e no meio do processo eu peguei birra de alguma coisa que ainda não sei – a idéia, o conceito, a execução, meu trabalho, não faço idéia – mas fui deixando de lado, enrolando e fazendo esforço mínimo até agora de manhã. Hoje é o último dia para entrega de trabalhos. Tinha uma porrada de coisa pra fazer e tirar o atraso, outras que a gente tinha feito em aula, semanas atrás, e eu não lembrava direito como reproduzia, enfim, só emoção!

Pra começar, o que é, exatamente, esse assignment? A idéia aqui era construir uma imagem 2D no Photoshop, quebrada em camadas bem definidas, seguindo o conceito de “city on the edge”, e depois usar essas camadas já divididas pra criar paralaxe e sensação de movimento tridimensional através de projeções no Maya. Para a minha, achei que seria legal ter uma coisa no deserto, com tons quentes e fugir do tradicional montanhas, castelos e neve.

O primeiro passo era fazer um documento-referência com imagens que seguissem o estilo que a gente queria alcançar no final do projeto. É um arquivo gigantesco, tá aqui embaixo.

A partir dele, cada um tinha que fazer cinco sketches bem simples, pequenos, em preto e branco, com variações de composição para a cena.

Depois, com o feedback recebido em aula, escolher dois desses e acrescentar cor na mistura.


Nesse passo eu mandei os meus pra pedir feedback nas cores e o Tom mandou de volta essa outra composição, misturando elementos variados de tudo que eu já tinha apresentado até aqui. Gostei MUITO da imagem dele, e acabei seguindo com essa como base até o final.

Em seguida, começamos a empilhar fotos por cima dos desenhos até chegar na imagem final, que seria convertida em 3D. Minha evolução foi em grandes passos, como dá pra ver abaixo.



Comecei a mexer no Photoshop pra fechar a composição da minha imagem e recortar todos os elementos direitinho – pelo menos esse pedaço eu já tava mais encaminhado. Não que não tenha mudado um monte, mas eu já sabia o que tinha que fazer. Fui começando a me divertir fazendo a parada, e quando chegou a hora de jogar tudo pro Maya, que as pecinhas foram se encaixando, fiquei decepcionado comigo mesmo por não ter começado a fazer esse assignment mais cedo.

Do Maya eu voltei pro Photoshop pra arrumar um mooonte de detalhes, a fim de deixar o movimento mais fluido, e fiquei nesse dueto Maya-Photoshop por algumas horinhas. Por fim, terminei de montar o quebra cabeça, criei uma depth pass rapidamente, já usando a estrutura existente – ao invés de usar o padrão de depth pass, que ia foder meus planos todos e mandei renderizar. Então, com as layers separadas bonitinhas, no Maya cada uma vai para um plano diferente e elas são espaçadas entre si para criar a sensação de movimento. Tá aqui uma imagem angulada de como é essa estrutura.

E aí vem a animação, pro resultado final:

< < BACK TO CHAPTER IIIB - ISCORAMAS (1.5X)

DISCLAIMER: This article was written way before the announcement of the FM system. I haven’t had time to study it properly, so some information here is probably outdated already.

C – DOUBLE FOCUS (2X)

Entry point for many anamorphic users, these lenses are cheaper because they present a lot of challenges to work with. All of them are projection lenses, have an awesome 2x stretch, and placed on top of a taking lens. As they’re projection lenses – and the screen is NEVER close to the projection booth – it’s quite common that their minimum focus ranges between 6 to 9m.

The name of the chapter already introduces on of their challenges: double focus. In order to have a focused image it’s required to focus both the taking lens and the anamorphic block to the very exact distance. If one of them is slightly off, image quality degrades quickly.

Taking this into account, it’s easy to figure that camera movements and rack focusing are very complex operations to be executed during a single shot. The shooting style has to fit the lenses.

It’s a practical thing to split them in two main categories: 35mm projector lenses – large, heavy and, sometimes, colorful – and smaller projector lenses, like 8mm or 16mm, used in airplanes, military applications or film clubs.

The 35mm anamorphic projector lenses are still in production by Schneider Optics. There are vintage models made by Sankor, Bausch & Lomb, Carl Zeiss e Möller. Their main advantages are high image quality and sharness from edge to edge of the frame. Because of their minimum focus distance, diopters – close-up lenses – are essential extras to use these lenses in narrative films (6m to achieve focus is pretty far for anything with a low budget). Besides, due to their weight, a proper support system is required.

The taking lens has to be considerably longer than the ones used in the previous examples, 85mm is the widest you can go on a full frame sensor to avoid side vignetting. When converting the focal length, this comes close to a 45mm on the horizontal axis.


Henri Chrètien Hypergonar Hi-Fi 2 and its support system.

Another obstacle: none of these lenses have filter threads, preventing proper use of diopters to improve minimum focus. Some have clamps to hold filters, but in most of them we have to improvise a solution, like the Hypergonar Hi-Fi 2, to which I glued a 95-86mm step down ring to the front of the lens, allowing the use of filters with both diameters.


95-86mm step down ring glued to the front of the Hypergonar Hi-Fi 2.

Now getting to the more interesting part of this chapter: lenses for smaller projectors. Manufactured almost exclusively in Japan by Kowa and Sankor during 1970-80 and shipped out to the whole world, these lenses come in different shapes and sizes but they mechanism and performance are almost identical. Minimum focus is set at 1.5m and properly designed clamps and rings to attach filters onto either Kowa or Sankor.

The most wanted versions are Kowa 8-Z and Sankor 16-D that used to be the high end versions of the smaller ones. It’s quite common to find these very same lenses with different labels – Eiki, Elmo, Vidoscope, D.O. Industries, Singer, the list keeps going – but their manufacturer was always the same, it’s just the name that changes.


Sankor 16-D plus clamps and accessories.

The 2x stretch makes the anamorphic bokeh very pronounced when using these lenses, but 16:9 video ends up with a ludicrous 3.56:1 proportion – a little image strip in the middle of a black screen. The best workaround for this issue is to shoot 4:3 as RED cameras in Anamorphic mode, or the Alexa 4:3, Canon DSLRs through Magic Lantern and, even more recently, the Panasonic GH4. This way the resulting image has a 2.66:1 aspect ratio.


Strong anamorphic flares and bokeh.

One of the differences between 35mm projector adapters and the smaller ones is the coating used on the glass. The 35mm ones are multi-coated since they’re modern lenses and this kills the anamorphic flares. On the other hand, the smaller ones are single-coated and present warm orange flares that can’t be achieved by any other way.

One of the most dreaded features introduced by projection lenses are the “anamorphic mumps” which first came to sight when Hollywood started using anamorphics and it’s still a common feature among the vintage models. David Bordwell explains it way better than I ever could:

“(…) the ‘squeeze ratio’ of Chrétien’s lens design varied across the horizontal axis. These optical tics created distortions and patches of soft focus. The most embarrassing flaw, created by faults in magnification and the uneven compression of the visual field, made central figures look oddly bloated. In close ups, the result was ‘CinemaScope mumps’. Not all of the films were shot with Bausch & Lomb lenses, but other brands of anamorphic lenses tended to cause the same problems.” (BORDWELL, 2007, p.288)

As Bordwell describes, the cause of this quirk is the uneven glass curvature related that creates the stretch factor. Around the edges of the lens the compression is stronger while in the middle the image tends to be more spherical-like. Stretch is not constant across the horizontal axis and also changes according to the focus distance. For objects that are further away, closer to infinity, stretch is 2x, but when the subjects come closer to the lens, this number can go as low as 1.7x. When this footage is stretched back to its “proper” aspect ratio whatever subject is in the center of the frame becomes wider than the real world version of itself, thus creating mumps.


Mumps on a close up shot using a 135mm taking lens + Kowa for Bell & Howell.

There are also exceptions and special cases, as double focus lenses with 1.5x stretch, designed for amateur 8mm and 16mm cameras, such as Bolex-Möller or Yashicascope. Because of their size, they’ve been kindly named baby anamorphics. As they exist in smaller numbers and from different manufacturers their optical quality vary greatly.

Taking these exceptions into account, the few double focus adapters that can be compared to Iscoramas, when it comes to image quality, are Kowa for Bell and Howell and the Bolex-Möller. The first ones, from Japan, were made specifically for an european projector company (Bell & Howell) that demanded extensive quality tests, so this brand of Kowas can be set apart from the other ones.


Kowa for Bell & Howell with Redstan clamps.

Bolex-Möller are germans and, as the name indicates, they come from a partnership between the camera maker Bolex and Möller optics, which means these are lenses for shooting, not projecting. Their minimum focus is at 0.5m. There are two different versions of Bolex-Möllers: one for 8mm and the other for 16mm, considerably bigger. After a while, the partnership ended, but Möller still made some more anamorphics on their own. Even though the markings are different (they don’t say “Bolex” anymore), the optical quality is identical.


Bolex-Möller 8/19/1.5x.

2x stretch is only achievable through thicker glass elements. Because of this extra thickness, another problem comes to life: while using DSLRs, depth of field can be so short that it doesn’t cover the whole thickness of the anamorphic glass. In this situation, the center of the image is in focus, but there’s a mismatch between vertical and horizontal fields on the edges of the frame, reducing image quality.

Taking this into account, in order to shoot good quality footage it’s better to work with smaller apertures which grant enough depth of field to cover not only the center of the adapter. Apertures between f/4-5.6 tend to have great results. The issue makes it harder for this kind of lens to have decent performance while shooting in low light conditions.

PROS:
- Very strong anamorphic bokeh;
- Lower prices than the other anamorphic adapters;
- Minimum focus between 0.5m and 1.5m (for older lenses);
- Strong flares (for older lenses);
- Great image quality from edge to edge (for modern lenses).

CONS:
- Double focus: it’s necessary to focus both the taking lens and the anamorphic at the same distance so the image is sharp in the camera;
- Too big and heavy (for modern lenses);
- No filter threads;
- Minimum focus is too far (for modern lenses);
- Taking lens aperture must be between f/4-5.6 for better performance;
- Camera must be fixed because racking focus is a nightmare;
- Mumps and distortions across the frame;
- 3.56:1 aspect ratio, way too wide for anything practical.

TO CHAPTER IIID – CINEMA LENS (2x) >>

< < BACK TO CHAPTER IIIA - FOCUS THROUGH (1.33X)

B – ISCORAMAS (1.5X)

Back in 2009, an Iscorama could be bought for little over a hundred bucks. Currently their price ranges between US$2500 and US$3500. It takes a lot of effort, patience and time to get better deals.

Iscoramas were made between 1970 and 1980, by ISCO Optics of Göttingen, in Germany, and work in a unique way. They’re adapters – in the sense that they fit in front of a spherical taking lens – and their high prices come from a great number of qualities. There are three main models, 36, 42 and 54, and their names come from the diameter of the rear anamorphic element. The weight difference between the three is drastic. While the 36 weighs around 400g, the Iscorama 54 goes over 1kg.

Stretch factor is 1.5x, which leads to an aspect ratio of 2.66:1, slightly wider than the CinemaScope standard. The widest one can go using an Iscorama on a full frame sensor in order to avoid vignetting is 50mm. Anything longer than that is safe.

Differently from focus through and double focus systems, which have improved performance between f/4-5.6, Iscoramas shoot properly focused images no matter what’s the aperture on the taking lens. Their sharpness is very very high.

Their biggest quality: all focusing is done based on the Iscorama’s focus markings, which is very accurate. The process consists in focusing the taking lens up to infinity and then working with the Iscorama, which has a minimum focus of 2m. It’s exactly the same thing as shooting with a regular lens, but keeping all the advantages of the anamorphic format. The way its focusing works is patented, which has never been bypassed, so these are the only adapters with single focus.

A couple of years after their original release, ISCO noticed that their adapters were being used paired with other lenses than the standard 50mm that came with it, so they took this chance to improve the glass and release them as individual pieces, without a default taking lens. That’s when the Iscoramas 36, 42 and 54 were born. They’re more modern and multi-coated, so less prone to flaring.

These lenses were designed for photography – not solely projection – and so they have standard filter threads (72mm, 82mm and 95mm for Iscoramas 36, 42 and 54 respectively), besides killing the need for special clamps on the rear element as well, since they also have regular filter threads (49mm, 62mm and 77mm respectively).


Iscoramas 36 and 54, Iscomorphot 8/1.5x and Isco Widescreen 2000.

Their mechanism to align the stretch is also simplified, directly on the lens through the use of side buttons that must be pushed in so the user can spin and align the stretch axis.

As any other anamorphics, Iscoramas also have their exceptions. The first of them is the Iscomorphot – also known as baby Iscorama -, a version for 8mm movie cameras which shares the same optics as its bigger brothers: focusing done only on the Iscomorphot and taking lens set to infinity. This little gem can focus down to 0.5m without diopters, but its image has very little sharpness and contrast, with reasonable results from f/5.6 or smaller apertures. Combined with APS-C sensors, it’s vignette free from 60mm on. Doesn’t work well with full frame cameras.


Iscomorphots 8/1.5x duo.

The second exception to the Iscorama family is the Isco Widescreen 2000, which works just as a focus through adapter with improved optics and 1.5x stretch. With this one, focusing is done on the taking lens and the adapter is also fixed between 4m and infinity, requiring diopters for closer shots. It was used as a projection lens, so needs clamps and mods in order to provide regular filter threads.


Isco Widescreen 2000 and accessories.

Not taking these two exceptions into account, here goes the pros and cons list for the Iscoramas:

PROS:
- Single focus: patented design, focusing is done only on the Iscorama;
- Simplified aligning process;
- No need for clamps;
- Strong flares (only with the Original Iscorama);
- Anamorphic bokeh;
- Great sharpness, even at larger apertures;
- 1.5x stretch which leads to an image close to the CinemaScope aspect ratio;
- Regular threads for diopters and filters.

CONS:
- VERY high prices;
- Minimum focus down to 2m (can be solved through diopters);
- Multi-coated (Iscoramas 36, 42 and 54).

A brief note about the Iscorama 36: since the lens body is made of plastic, it’s quite common that as time goes by they are more prone to damage. Aiming at this valuable lenses market, british company Van Diemen developed a new housing for the Iscorama, fully metal made, with follow focus gears and a minimum focus modification that drops it to 1.2m instead of the original 2m.

The process costs £850 and has an estimated turnaround of 90 days to be completed plus mailing time. The conversion’s results are impressive but there are a lot of customers’ complaints about waiting a much longer time than advertised to get their lenses back.

There’s also a homemade modification to improve the lens’ minimum focus by unscrewing the front element and removing the original stopper. It’s risky if you’re not used to opening lenses, but works perfectly.

I’ve sent my Iscorama 36 for Van Diemen’s rehousing and wrote a full review of the process.

TO CHAPTER IIIC – DOUBLE FOCUS (2x) >>

< < BACK TO CHAPTER III - LENS RESEARCH

A – FOCUS THROUGH (1.33X)

The most subtle lenses, when it comes to stretch factor are also the newest ones. Between 1999 and 2004 the standard screen size switched from 4:3 to 16:9 and, the same way it happened in the movie industry, would be too expensive to replace all the camera equipment currently in use – specially for those who had just bought an expensive mini-DV camera.

The adapters’ manufacturers were scarce and their target market was composed by users of high-end cameras like Sony PD170, VX2000, Panasonic DVX-100 or Canon GL1 and GL2, that shot only at 4:3 aspect ratio. Panasonic made their own adapter, the AG-LA7200, one of the biggest 1.33x, expensive until this day. Sony and Canon were covered, mainly, by Century Optics (which ended up as part of Schneider Optics) and smaller brands (also less reliable when it comes to quality), as Optex and Soligor.


Optex 1.33x Adapter

These cameras had very different (and non-standard) mounts: some used bayonets, others had regular filter screws and different sizes. Each adapter had a specific back for the camera it should work with. Another important fact is that most of them were only partial-zooms, which means they only worked well within a specific focal range on the camera’s zoom range. From 50mm onward, the image would be completely blurry.

Focus was also fixed between 4m and infinity, so there would be some serious blurriness issues when the subject was closer than that. This design’s big advantage is that all focusing is done on the taking lens. It’s like focus went “through” the squeeze, easy to understand their name now, I think.

Both these negative qualities – partial zoom and fixed focus – reflect on their modern reuse, paired with DSLRs. Lenses over 50mm and subjects closer than 4m both require diopters to be in focus. Diopters are treated like filters but they’re actually auxiliary lenses and also go by the name of close-ups.

The biggest advantage of this type of adapter is that they allow the use of wide-angle lenses – between 28 and 35mm on a full frame sensor – without any vignetting from the internal walls of the adapter. Besides that, all the focusing is still done with the taking lenses, simplifying set workflow.

Their build is simple: just two cylindrical optical elements aligned with each other. All 1.33x adapters have a screw on the side that, when loosen, allows the user to rotate the glass and properly align the stretch direction.

Low quality on the edges and corners is a common issue with 1.33x adapters. It’s easy to notice strong chromatic aberration, loss of sharpness and barrel distortion when the taking lens is a wide-angle and you’re almost vignetting. Flares are also very strong, two thin lines easily recognizable by their strong blue tint.


Flare, lens distortion and chromatic aberrations on a single picture.

As many of these adapters have bayonet mounts, they require some modifications using step-up rings and pliers (still to be translated). Another challenge is attaching diopters in front of them in a safe way, since there are no threads for additional filters. There are clamps that work with the smaller models – Optex, Century Optics and Soligor – but the Panasonic AG-LA7200 is always a challenge due to its massive size.


Century Optics with clamp for 72mm threaded filters.

Panasonic’s greatest advantage, and the reason it’s still so valuable, is his optical quality that’s better than the smaller models and the fact that it allows much wider shots through 28mm lenses on a full frame sensor without any vignetting.

The downside is because of its size, it’s pretty hard to find properly sized diopters. With a great deal of help from a friend, Bruno Nicko, starting with a 3D model using the exact measurements taken from the front of the lens, we made a unique aluminum piece with 105mm threads.


Panasonic AG-LA7200 with custom 105mm thread clamp.

Without this piece, the only diopter choices for this lens would be from the rare russian LOMO Foton-A, or 4.5 inch filters. These diopters are big enough to cover the Panasonic AG-LA7200, but they still require some mods in order to attach safely to the front of the lens. There’s also a legend about a 105-86mm step down ring that no one ever found on ebay.


LOMO Foton-A +1 diopter.

Near the end of the DV fever, when many cameras were already capable of shooting 16:9 digitally through the pixel aspect ratio feature, Century Optics released the only full-zoom adapter for this dying market. It’s a big lens (almost the same size as the Panasonic) that, besides the alignment knob, also has a focus ring. This allows the user to have sharp images of subjects up to 0.5m (opposed to the original 4m for the partial zooms) and also use lenses longer than 50mm..

Its performance is much superior when compared to the other adapters – even considering its double focus aspect: it’s required to focus both taking lens and the adapter in kind of the same distance. Not nearly as precise as double focus needs to be (we’ll get to that in a couple chapters). It’s still possible to rack focus using only the taking lens, as long as you plan it well and set the adapter’s focus ring to cover the target range.


Century Optics WS-13-SB.

As you can see in the pictures of the adapters so far, the optical elements in these lenses aren’t circles, but more of a square shape, which means that the oval shaped highlights – anamorphic bokeh – won’t show up when this kind of adapter is used, first because the 1.33x stretch is too subtle to be noticed on the bokeh and second because the square aspect of the glass doesn’t affect the light shape. To achieve this effect it’s necessary to use a modified taking lens (see chapter IIIH – Faking the Look). On the bright side, when you have a 1.33x stretch applied to 16:9 footage – such as today’s standards – the resulting image follows a 2.36:1 proportion which means CinemaScope aspect ratio straight out of the camera.

A problem that didn’t exist when these adapters first came out is, due to DSLRs large sensors, getting a properly focused image using focus through adapters can be challenging. Mini-DVs have a 1/3″ sensors, which caused a larger depth of field. When you pick a sensor that’s almost thirty five times bigger than that, the best way to get a sharp image is to close down the iris on the taking lens. Apertures between f/4 and f/5.6 are usually enough to achieve good looking, good focus results. Diopters also help a lot when it comes to improve the quality of the images taken with focus through adapters.

A brief comparison between pros and cons for focus through adapters:

PROS:
- They’re focus through. All focusing is done on the taking lens;
- Relatively low prices;
- Allows using wide-angle lenses;
- Strong flares;
- Natural CinemaScope.

CONS:
- Strong lens distortion and chromatic aberration around the edges;
- Needs diopters when the taking lens is longer than 50mm;
- Needs diopters for subjects closer than 4m;
- Taking lens aperture should be between f/4 and f/5.6 for best performance;
- No oval bokeh.

TO CHAPTER IIIB – ISCORAMAS (1.5x) >>

Far Cry 4.

Um ano e meio atrás eu postei sobre a diversão que foi jogar Far Cry 3. Agora estou de volta, pra falar da edição seguinte. Lançado pouco menos de um mês atrás, Far Cry me manteve entretido por mais de quarenta horas. Jogar um shooter no PS4 é ridiculamente mais difícil que no PC, mas mesmo assim sobrevivi. Explorei todos os cantos, cavernas e montanhas de Kyrat. Em termos de gráficos, a diferença entre FC3 e FC4 não é lá grande coisa. A história tem similaridades também, e missões parecidas, mas há mais opções de escolhas e o jogo todo é muito consciente de si mesmo – o vilão constantemente questiona suas escolhas e ironiza o que acontece no mundo fictício.

De uma forma bem sutil, o jogo faz piada com seu próprio gênero, o que acho algo genial porque faz o jogador pensar, nem que seja um pouquinho, sobre o que ele tá fazendo ali.

Anyway, a destruição é fantástica e eles realmente acertaram no funcionamento dos explosivos dessa vez: carros e coisas voam primeiro, ao invés de entrarem em combustão espontânea com uma humilde granada, como era em FC3, o que faz a dinâmica de combate ficar muito mais caótica (por exemplo jogar um carro pra cima e passar por baixo a toda velocidade, enquanto dirigindo), dá pra criar barreiras e se esconder atrás, dá pra explodir TUDO (tem um lança granadas semi automático) – falando nisso, o arsenal do jogo é enorme. São 64 armas, sendo que a maioria delas tem customizações, lunetas, cartuchos maiores, silenciadores epor aí vai. O mapa é colossal – ao longo dos primeiros dias, achei que tinha andado pra caralho. Aí abri o mapa, e tava em menos de um décimo do espaço total!

Chega de rasgação de seda, isso não é um review, é só um comentário animado sobre ter terminado 100% hoje, e um print screen exagerado com TUDO que completei nas montanhas de Kyrat.

Lista de Chamada.

Entre todos os meus anos de escola, faculdade e afins, sempre tem uma coisa em comum que me desagradava: a ordem alfabética na chamada. Ok, se fosse só pra chamada era tranquilo, mas era também pra apresentações, entrega de prova, entrega de trabalho, então duas opções eram cenário comum e as duas envolviam expectativas baseadas em terceiros. Ou todo mundo ia bem, e eu ficava preocupado de ter ido mal, ou todo mundo ia mal e eu ficava preocupado de ter ido mal também, ou era absolutamente aleatório, e uns muito bem outros muito mal, e não tinha parâmetro do que esperar.

Acho que foi por isso que eu parei de ver trailers de filme: expectativas destruindo tudo. Eu adoro trailer, mas se o filme me interessa de verdade, eu não vou ver trailer de história, só teasers, que mostram planos legais, mas totalmente fora de contexto. Depois de ver trailers e filmes o suficiente, dá pra antever 95% da trama, inclusive com as revelações bombásticas, só assistindo o trailer. Se o filme é baseado num livro, ou qualquer coisa assim, menos mal, porque o fim já tá lá, e muita gente sabe, mas se é roteiro original, costumo evitar. Duas experiências recentes me provaram errado nessa teoria, porém. A primeira delas foi Edge of Tomorrow, que postei aqui uns meses atrás, que eu achei que ia ser só ação desenfreada e nada demais, mas o filme é absolutamente incrível, e depois Gone Girl, que vi com a May e o trailer te encaminha numa direção, mas o filme faz as coisas totalmente de outro jeito, e a sensação de surpresa, de não-esperado, é absolutamente maravilhosa.

Enfim, como e por que eu comecei a escrever sobre isso? Porque aqui na VFS a lista de chamada não segue ordem alfabética. Não sei qual é a ordem que eles seguem, talvez de matrícula, sei lá, mas sei que sou um dos primeiros, quase pelo meio, num lugar bem sossegado. Querem mais desvantagens de estar no fim da chamada? Enquanto seus coleguinhas Albertos, Alexandres, Andrés e similares são chamados primeiro, eles já começam a conversar, e você tem que ficar lá, prestando atenção na conversa e na chamada ao mesmo tempo, pra não tomar uma falta ou perder o nome na hora e mandar um grito depois de “ô professor! não ouvi meu nome! tô aqui!”, tipo aquele que a gente lança na hora que o ônibus tá lotado e você precisa descer, mas ainda tem seis infelizes entre você e a porta, e você corre se espremendo e se acotovelando enquanto o motorista começa a andar, aquele grito de “peraê motõ, que eu ainda vou descer!”.

Tá, falei mais um parágrafo inteiro e não expliquei como cheguei nesse assunto. Pelo menos falei da VFS, que era a origem do tema. Durante as apresentações de ontem, eu fui o penúltimo de todos. Ser penúltimo, previsto pra feedback às 2h30 é uma coisa, mas quando a história toda atrasa mais uma hora e meia, ficar esperando e pensando em tudo que você fez errado ou podia ter feito melhor começa a dar nos nervos. Nessas situações eu sempre acho que o resultado vai ser pior do que eu quero que seja. Não sei por que, talvez algo relacionado a “sempre se preparar para o pior”. Engraçado que essa sensação é pra quando as coisas estão terminando. Quando estão começando, eu sempre vou todo “ah, é tranquilo, vai ser moleza!”, e quebro a cara repetidas vezes – porque quase nunca é moleza mesmo.

Enfim, falei um monte de groselha que não é relevante pra ninguém. Vou pensar em um assunto ainda menos produtivo pra escrever.

Mudança de Formato.

Ontem tivemos o feedback e revisão dos nossos projetos de demo reel na VFS. Tá melhor explicado nesse post aqui, mas uma coisa é tão peculiar e relacionada com esse blog e minha evolução no caminho da fotografia que PRECISAVA ser postada aqui.

A GRANDE alteração no meu projeto é que o formato da tela agora vai ser QUADRADO. Proporção 1:1. Achei bastante curioso – e até engraçado – depois do grande surto e todas as experiências com anamórficas e tudo mais, meu primeiro projeto de verdade não é nem 4:3, é QUADRADO! Claro, o quadrado tá muito bem justificado na história e até na divulgação, mas ainda estou impressionado com essa série de acontecimentos.

Dezembro de 2013: “Nunca mais vou filmar em 16:9, widescreen normal é chato, vou vender todas as minhas lentes, agora é só Cinemascope, anamórfico, yeah yeah!”

Dezembro de 2014: “Foda-se o que eu disse, vou fazer um filme QUADRADO!”

A vida apronta cada coisa com a gente, né? Em termos muito positivos, agora vou aprender a força de composições quadradas, médio formato, visibilidade em celulares, redes sociais, e essas outras coisas loucas que a gente do cinema tende a achar que é muito previsível (eu pelo menos tendo a pensar assim), além de poder aproveitar as habilidades e conhecimentos de minha irmã recém-formada!

< < BACK TO CHAPTER II - INTRODUCTION

LENS RESEARCH

After the first weeks of research, reading, watching tons of test videos and getting to the core of the anamorphic squeeze pros and cons, there were still plenty of questions unanswered. My curiosity got the best of me and so I decided it was time to invest in buying lenses. My starting point was Andrew’s Anamorphic Shooter’s Guide, from there I went on refining my decisions and ended up buying four lenses, all of them way different from each other and that will be better explained along the following chapters. They were Kowa for Bell & Howell, Panasonic AG-LA7200, Hypergonar Hi-Fi 2 and a 50mm LOMO Squarefront.

More experienced users reinforced the idea that I should use simpler taking lenses with less coatings to make flares pop. This meant no modern glass, no Canon L-Series, no zoom lenses. I ended up choosing vintage primes, from 1970-80, and was pretty happy with russian designed M42 mount all manual
(focus and aperture) lenses. The kit included a 37mm f/2.8 (Mir-1B), 58mm f/2 (Helios 44), 85mm f/2.8 (Jupiter 9) and 135mm f/2.8 (Tair 11). With these lenses I covered the standard focal lenghts in a prime basic kit and could figure out all my framing and composition using them and moving the camera slightly forward or back.

To flare out better those anamorphic features, it’s good to pair them with vintage, pre 1970′s taking lenses, years before the development of multi-coatings (MC), a physical high-vacuum process of metalization (I’m pretty sure this is mistranslated) that protects modern lenses from glass damage, lessen flaring effects and increase light transmission. This means that lenses made before this thing was invented are non-coated or single coated, which affects greatly anamorphic’s more proeminent quality: lens flares.

I’ve picked mine as M42 mount as a matter of personal taste, but there are plenty of other mounts that can be easily adapted to Canon EF. Even though adapters are quite common, before buying any lens, be sure to check if it works properly with the camera you’re using. A simple search is good enough to save you from a lot of headache involving flange distance and things being out of focus when they should be sharp. This also helps you to choose exactly which model you want and what features you want to accentuate on your work (I mean, there are countless types of 50mm around, which one suits you best?).

There was still an important unanswered question: how to connect properly both optical blocks, anamorphic in front and the taking lens behind it? LOMOs aside, all anamorphics can be considered adapters added to the taking lens, as a filter that you screw on top and gives you a very specific result.

In the very beginning of my searches I came across lots of posts stating that for working perfectly safe, you should use clamps to connect both lenses. Clamps are, in its essence, metal rings that attach to the back of the anamorphic adapter, mostly using screws, and end up on a regular filter thread that goes onto your taking lens. This allows the closest possible distance between the two optics, which saves us from the trouble of light loss when working with anamorphic lenses (usually the loss is so small it doesn’t even get to a third of a stop).


Redstan Clamp for Kowa Bell & Howell, 62mm thread

Many anamorphic adapters have threads on their backs already, but these aren’t standard sizes and we usually have things like 83mm or 73mm, that aren’t used nowadays. Clamps serve to fix this issue as well as to deal with stretch alignment. It’s possible to loosen the screws, without the adapter toppling over, and rotate the anamorphic block so the stretch is properly aligned with the horizontal axis. Screwing an anamorphic onto a taking and having it totally skewed in the camera is the first thing that happens when you switch lenses. It’s fundamental to align it properly or the final image will be irreversably skewed. Check the pictures below to see what this means, exactly.


Misaligned anamorphic block

Aligned anamorphic block

The main advantage of purchasing a clamp that was made specifically for a lens model is that both fit perfectly – it never gets loose or puts any of your lenses in danger of falling -, feels perfectly safe and it’s super easy to align. There are very few real manufacturers and a LOT of improvised workarounds. Redstan, in the UK, is quite famous for his very specific clamps of the most common lenses, and Vid-Atlantic, an american company with cheaper options if you don’t want to spend too much money on a piece of metal. Redstan’s quality is unbeatable, but he’s quickly out of stock because they take a while to build and people are always buying.

As I said in the beginning of this chapter, each one of my lenses worked differently from the others, had different stretches, image quality and years of manufacture. I’ll split them according to stretch factor because it’s the easiest way to group them.

TO CHAPTER IIIA – FOCUS-THROUGH (1.33x) >>

< < BACK TO CHAPTER I - WHO AM I AND WHAT IS THIS ABOUT

INTRODUCTION

What are these so called “anamorphics”? The dictionary states “Optics. having or producing unequal magnifications along two axes perpendicular to each other”. In our case, a lens or adapter that distorts only the image’s horizontal axis. This is called “stretch factor”, which ranges between 1.33x, 1.5x and extreme 2x. That means that the resulting image will be horizontally squeezed and, in order to bring it back to normal proportion, you need to unsqueeze (or stretch) it following these ratios. A different name for stretch would be pixel aspect ratio: a virtual number that defines the width of each single pixel when displayed (instead of the standard 1×1 ratio).

And what’s the point of having something that looks totally weird in camera and then NEEDS to be post-processed so it looks right? Well, anamorphic lenses were “born” for the movies around 1950, when Twentieth Century Fox bought a patent for Henri Chrètien’s (a french scientist) system of bent lenses developed for astronomical research.

Years before, the movie industry had already got to the conclusion that wider images are more appealing to human eyes. There was already a process to achieve such results, called Cinerama, in which movies were shot with three simultaneous cameras and showed in a similar setup with projectors using a curved screen that was way bigger than the current standards. The downside was triplicating the use of celuloid for each single movie during production, and even larger amounts during distribution, which wasn’t being so profitable.

Around the same decade, three different methods of anamorphosis competed in the industry. The first one was based off prisms, developed and used by Panavision, the second was achieved through the use of bent mirrors, Technirama and the third one is our all-time favorite, the subject of this work, bent lenses that compress the captured image, originating the famous – and desired – CinemaScope, 2.4:1 aspect ratio, meaning that for each height measurement, there’s a 2.4x measurement in width.



Celuloid frame and different methods for a widescreen image.
Letterbox on top and anamorphic on the bottom.

In Chrètien’s design, the image compression would be directly related to the glass’ curvature. His anamorphic lenses were made up, roughly by two optical blocks. The front block was responsible for the anamorphosis and the rear block was a regular spherical lens, both put together inside a single lens body. Simpler setups work the same way until today (we’ll get to that) and you have to focus both spherical and anamorphic separately.

From here on, the spherical block will be referred as “taking lens”. It’s an important concept that will be used over and over. Don’t get confused.

Once the image is shot and squeezed onto the film roll – there was no digital intermediate back then! – another anamorphic lens was required when projecting the final motion picture in order to de-stretch the image. These projection lenses are still out there, flooding ebay and every other online market. They’re bulky, colorful and heavy. Another downside is their really far minimum focus distance since the screen and the projection booth is usually somewhere between 9 to 15m apart.

Meanwhile, across the world – the Cold War was going on and the USSR was also a major technological center – LOMO anamorphics were coming out, russian lenses made by their main optical developer, responsible for all camera and movie gear in the country conglomerate. These lenses were made solely for motion pictures and, differently from the american/french design, anamorphic and spherical blocks were split by default, but could be connected through a simple bolt, which made sure focusing on both blocks was identical and solving the issue of having to focus each lens individually. These were pretty strong builds and spread mainly across Asia. As the american standard, LOMOs had a 2x stretch.

At the beginning of the 1960′s, ISCO Optics of Göttingen, german manufacturer hires the man behind LOMO’s designs, deliberately replicating his previous work with LOMO, ISCO releases a series of anamorphic lenses – the Iscoramas, with a stretch of 1.5x – aimed at the “rich amateur photographers” niche. The series consisted of a front anamorphic block attached to a 50mm f/2.8 cheaper taking lens (with Exakta, Minolta, Nikon F or Praktica M42 mounts). Image quality matched the market: sharp from edge to corner.


The Original Iscorama – 1.5x stretch

Almost twenty years later, in the 1980′s, Iscorama’s users found out it was possible to split both parts of the lens and use the anamorphic block paired to other taking lenses. ISCO takes advantage of this practice to release updated versions of their products, getting rid of third-party manufacturers by discarding the 50mm taking lens and selling just the anamorphics. Those were the incredibly famous Iscoramas 36, 42 and 54.

The optical engineer behind these lenses used a design based on variable diopters, in which the taking lens is focused to infinity and all the focusing is done on the Iscorama. By registering this patent, ISCO Optics of Göttingen killed any other manufacturer’s dream of replicating its solution. This caused a drastic loss in terms of similar workarounds until this day, making Iscoramas a unique species.

What we’ve seen so far: based on the need of creating wider-looking images and avoiding redesigning and making every single motion picture camera currently in use, scientists develop anamorphic lenses that attach on top of regular spherical lenses and compress their images. Russians aren’t far behind and develop their own kind of lenses. Germans steal the russians’ formula and improve it, releasing the Iscoramas, which hold single focus patent.

Following these events, japanese companies start making their own anamorphic elements (Kowa and Sankor), with a 2x stretch, and even Henri Chrètien through Societe Technique Optique de Precision puts out some lenses for cameras and projectors, the Hypergonars S.T.O.P.

Market keeps evolving and, while some of these lenses were pretty good with 8mm and 16mm film, with the change from film to video, and later, to all-digital formats, amateur anamorphics are less and less useful and end up forgotten at the bottom of “used gear” boxes. Videocamera’s lenses aren’t interchangeable and still photographers were never too fond of the squeezed images. Around 2008 almost all these anamorphic gems could be grabbed off eBay for less than US$200.

The last additions to the anamorphics list came out after mini-DV cameras became popular. These are focus-through adapters and their name comes from the fact that they don’t have a focus ring, just a simple screw for proper alignment – so the stretch is applied on the correct axis. Focusing is done on the taking lens, attached to the camera.

At the top of its game, Panasonic’s DVX100 – the camera that brought the power back to the indie moviemaker – got its own anamorphic adapter, the Panasonic AG-LA7200, with much greater size and quality than this generic-branded competitors.

These adapters goal, however, isn’t to achieve CinemaScope aspect ratio. Most mini-DV cameras, and even the DVX100, shot with 4:3 aspect ratio, as any default TV at the time. These adapters came out to allow older cameras to shoot in the new and widescreen 16:9 “modern” proportion, much like what happened in the the movie industry: it’s cheaper to buy an adapter than a new camera. Unfortunately these weren’t a big hit because they were already pretty expensive and not many people were willing to pay extra for them. Were also forgotten and discontinued.

We’re finally at 2009 and all these lenses can be found on eBay for laughably low prices. Canon 5D Mk II came out the previous year and it was revolutionizing digital video’s world. Then, a curious dude (or gal) like you and I decided to check if these old, weird and (at the moment) pretty cheap lenses could work with DSLR’s video capabilities.

I can only imagine his/hers surprise when the thing worked and, not only that, the resulting image was amazingly wide and unique. Results were posted online and some more people got curious. Among these it’s easy to name Andrew Reid – responsible for EOSHD and author of some of the books used during my research -, Edwin Lee – famous for being a pioneer in anamorphics and DSLRs – and Alan Doyle, also known as Redstan – a great source of knowledge about motion picture anamorphic lenses.

Shortly after processing the boost in width, the new anamorphic users realized a couple other features inherent to those lenses and adapters that were pretty hard to achieve if not using them, which had a lot to do with the images’ cinematic feel. The first of these features was the different bokeh, out of focus highlights, which remained oval even after proper unsqueezing. With regular spherical lenses, bokeh has always more of a circle shape.

Another unique feature that quickly drove anamorphics up in popularity not only among the indie productions but also on big budget feature films was the anamorphic lens flare. Lens flares aren’t always bad, specially if used while developing a unique look. Anamorphic flares are long, streaked and (usually) blue. They’re totally different from spherical flares and have a lot of personality. Using J.J. Abrams as a reference (easy one, I know), anamorphic flares are a constant through his work and we can easily spot his signature over a couple brief seconds.

From these various tests, reviews and experiments published over YouTube, Vimeo, EOSHD and other online forums, many others got interested in the game and so began the rush for anamorphic glass worldwide. In less then a year, Iscoramas that went for US$200 now reached US$4000 on eBay auctions. Many other lenses were also sold for crazy-high prices.

Following this raise in value, many anamorphic owners decided to sell their beat up old lenses, increasing the availability and lowering overall sales prices. Now, a couple years after the initial outbreak, an Iscorama goes for about US$2500 on auctions and a little more in direct sales.

LOMOs’ prices are widely apart since many of the already have cinema mounts and quality standards, so can easily go over US$4000, but it’s important to keep in mind that these were never cheap in the first place and we’re brought into the DSLR world through the use of adapters. I won’t risk writing about focus-through adapters and projection lenses because their prices don’t seem to follow such strict rules.

Over the last months the anamorphic community and enthusiasts rejoiced with announcements of brand new lens designs and releases to come soon. Among them we have SLR Magic – Hong Kong based – and Letus35 – in the US. Both promise a 1.33x stretch and focus-through adapters with custom controls for situations such as close up shots. Price ranges between US$800 and US$1300.

TO CHAPTER III – LENS RESEARCH >>

My graduation work at University of São Paulo revolved around shooting with anamorphic lenses, Magic Lantern RAW and Canon cameras, keeping the budget as low as possible and experimenting as high as I could.

The resulting document was written in portuguese and I’m now beginning to translate it, so more people have access to my conclusions and don’t have to take so much time going through the process. I’ll try to translate one chapter per week so hang tight, we’ll get to the end, eventually. In portuguese, this thing is around a hundred pages long, so it’s thorough.

Keep in mind this was originally finished by December 2013 and most 4k consumer cameras weren’t out yet, as well as Black Magic was just starting to ship their first units.

A lot has changed in terms of cameras, I used (and still use) a Canon 5D3 for my research and the camera is the least important part of this thing.

DISCLAIMER! If you see something really weird or stupid with my english, please leave a comment. I’m Brazilian, so english is not my native language, and I want this to be easily understandable for everyone.

To encourage anyone willing to go through all this ammount of text in the search for answers, here’s the end result of the process, a two-part episode for a webseries, almost totally developed by myself (I wrote, directed, shot, edited and did all the post-production work). Both parts are absolutely 100% anamorphic and will be carefully analyzed over the following chapters. Oh, be sure to enable english subtitles!


WHO AM I AND WHAT IS THIS ABOUT?

One of the main things for me in photography is the strong connection between technique and creativity one has to make to achieve good looking results. Knowing how the camera works isn’t enough. We need to understand human vision, composition, colors, light, and find/develop our own style.

When I started to attend to Film Production (Curso Superior do Audiovisual, at USP), in 2010, I was already taking pictures for and studying photography for a year and a half. It was my main hobby and entertainment while not studying the subjects that would get me into the university.

Coming from a very math-based background (Computer Science at Federal University of Bahia, UFBA), numbers were already pretty exciting to me. Through my four years at Film Production I started to develop and emphasize the creative part of the photographic process as well. Be it a single frame, be it moving pictures at 24 frames per second.

During summer break (winter break, for the north hemisphere), mid 2012 I was feeling bored and needed to learn something entirely new to me. I decided it would be a testing and experimenting year, very different from my standard, production-like work of the previous three years. Randomly roaming the web, came across a subject that had already caught my eye before: anamorphic lenses. That summer felt like the perfect time to go deeper into them. It was time to move from 16:9 TV standards to something more movie like as the 2.4:1 CinemaScope aspect ratio.

When I say I had already come across the subject before, it was around 2010 and, again randomly, read some articles about such lenses, that would allow shooting CinemaScope straight from camera instead of cropping the top and bottom of the frame. It was interesting at the time, but I lacked the technical knowledge to fully understand what was being said on those articles. Time went by and I ended up forgetting all about until 2012′s summer.

In the meantime whenever I considered shooting CinemaScope pictures, I always felt bad for discarding almost half the image through cropping. Such a waste of money and production value, right? Using the crop method for CinemaScope you either go ultra wide angle so you can frame your characters properly and end up with a ton of depth of field, or go with regular lenses and hope nothing too important ends up cropped off in post-production.


Without and with anamorphics aspect ratio comparison

After my initial readings I found out that with anamorphic lenses the image is squeezed before it reaches the sensor/film. Its horizontal correct measurement is determined by the image’s default width multiplied by a stretch factor (which may vary from lens to lens). You don’t discard parts of your frame, quite the opposite, you add more into it, you add stuff that would fall outside of a regular spherical lens angle of view.

From 2012 until now I’ve read countless articles, threads, forums, posts, ebay auctions and even a book entirely on the subject in order to better understand these lenses inner workings. It’s not all pretty and wonderful as it seems in the beginning, but they achieve a great deal of what I expected. There are various complications and quirks, but nothing that’s absolutely impossible to merge onto set work procedures.

I reached out for other anamorphic users in Brazil, lenses for sale, trade or testing, any kind of answer for my questions in my own language. Didn’t find anything at all, just a bunch of curious people like me. With the internet’s connecting power I went virtually abroad and started researching and experimenting. During this process, way over twenty different anamorphic lenses went through my own inspection and analysis.

This series of experiments were followed by ongoing posts on my personal blog (this one right here), aiming at introducing the subject in portuguese. I was pretty much obsessed with these lenses already and had to share my knowledge with whoever was interested, specially because I had a ton of questions at the beginning and couldn’t find any answers until I had the lenses in my hands and looking through the camera.

Research has been an amazing experience. During the process I ended up selling ALL my spherical and modern glass, keeping only vintage russian lenses and anamorphics. I don’t intend to leave them, it was a point of no return.

My main goal with this project is to share as much as I can of all I discovered and tested with these exotic and rare lenses, be it in terms of origins and history, as well as practical use in real shots.

I mean, amazing CinemaScope pictures in a camera that costs less than US$4k, when could I dream of this?

TO CHAPTER II – INTRODUCTION >>

« Older entries