It seems like technology updates and advances are almost an everyday occurrence these days with yet another new camera, tablet and phone model vying for our attention. It’s easy to glaze over whenever another tech announcement is made. Rarely do we see major shifts like we did with the release of the CD, DVD and iPhone. For the independent film makers and videographers of this world, nothing shook up the status quo quite like the release of Canon’s 5Dm2 camera in October of 2008. Suddenly, and seemingly out-of-the-blue, the look-and-feel of cinema was affordable to the masses. The shallow depth of field of its full frame sensor enabled anyone with a vision to shoot their story cinematically. It’s a well-known part of our industry’s history, but i feel its worth reflecting on again here. That is because i feel that the camera we have had the privilege to shoot with over the past couple of months reminds me of the last “tectonic tech shift” we saw four years ago.
For a long time we battled with all the limitations that DSLR video brought with it – lack of controllable audio and a way to support and stabilize it to name just a couple. We learned to overcome these obstacles because of the incredible payoff provided in the footage we were now able to capture. I remember shooting one of the first DSLR weddings with a 5Dm2 back in early 2009. The resulting film was short and gained lots of adulation, but the resounding voices thought up all the reasons it was impressive, but by no means a realistic alternative to “professional” video. The costly, and small capacity media, lack of time code, handling problems and lack of full manual controls were all problems people focused on. Now just look where DSLR video production is today.. countless commercials, independent films and even features have been produced on DSLR. Like many of you I have looked up to the work of people such as Vincent Laforet, Shane Hurlbut, Konrad Czystowski and Phillip Bloom just to name a few. I think the thing that is common to all these innovators is where many people just saw problems, they focused on creating solutions and using this new technology in ways people hadn’t thought of.
Enter the 1DC.
Ever since my first visit to the annual NAB conference in Las Vegas (where I first laid my hands on the 5Dm2) I’ve taken time each year to revisit. With the mammoth amount of information online during and after this Mecca of Film, TV + Broadcast expo, many colleagues asked why spend the money and effort to travel from Sydney Australia to go? Anything you’ll see at the show will be online and in forums almost instantly wont it? Well, Yes. But apart from being a great chance to let whatever-hair-i-have-left down, there’s nothing quite like getting hands on and meeting the people at the cutting edge of tech in person.
The 2012 conference seemed to be the dwindling of the 3D craze replaced by the upsurge in 4K resolution offerings. After hearing Shane Hurlbut discusses his remarkably impressive “the ticket” presentation, I shuffled slowly behind the Canon stage to where the new 4K 30 inch displays were being showcased. Suddenly my foggy pre-midday-Vegas-brain became alert – What a picture! Seeing 4K resolution video projected on a screen is a familiar look, however, seeing crystal-clear vision of a steam train chugging through a ravine only inches from your face is completely different. For those who haven’t seen 4K resolution on a display for themselves, it simply is hard to describe. Its true what they say, it is like looking through a spotlessly clean window to the scene beyond. It’s truly remarkable. What was so impressive with Canon’s new 4K camera is the fact that it is a DSLR. 4K resolution (and higher) has been around for some time now through camera’s like the RED Epic, Scarlett and Sony’s F65. But never have we seen this kind of resolution in a DSLR format. I couldn’t wait to get my hands on one.
Just a few short months later, I was luckily enough to be present when the first pre-release sample unit was being unwrapped in the Canon Australia offices in North Ryde Sydney. What was even more exciting was being told – “Do you want to take it out for a week and have a go?” Many people celebrate “Christmas in July” in Australia due to our scorching hot summers.. the Christmas spirit was not lost on me at that moment!
Along with our commercial and documentary project, untitled film works covers many weddings around Australia and abroad. The nearest shoot we had was a wedding in the beautiful Hunter Valley wine region that weekend. We decided to bring the body along and where possible shoot some test footage. Initially, apart from the extra size being a 1 Series Canon body, the look and feel of the camera was quite familiar. We rattled off several shots during the photo shoot, and apart from being super impressed by the quality of the LCD screen, we couldn’t tell in the field how the image compared to that of our 5Dm3 cameras.
A “pause” replaces shutter?!
Portraits taken in the latter half of the 19th Century show people staring expressionless down the lens. A smile was quite simply too hard to hold for minutes at a time due to the long exposure times required at the time to capture an image. When the exposure times reduced we started seeing more natural looking portraits of people. This is a great example of how a technological progression improved photography.
Pulling still frames from video is not a new concept. I have lovingly browsed through some of my great photojournalistic compendiums from the likes of Time Magazine, LIFE and the Pulitzer Prize winning images of the 20th Century. It’s amazing to discover that some images are “sequenced photographs”- which means they are still frames that originated from 35mm Motion Picture footage. Early War photography is often seen with such bylines. In fact when we look back at the development in the photographic process, we find the style of photography developing with it.
In the last few years we have seen RED’s remarkable innovations in what they call DSMC (Digital Still and Motion Camera) concepts, so the idea has been a slowly evolving one. What I think makes the 1DC such an exciting new prospect is that finally this technology is in a form that is suitable for the photographer. (we should remember that the 5Dm2’s video function was added to enable journalists to produce video clips in the same camera whilst in the field). A RED camera is simply not practical for run-and-gun scenarios that DSLR’s are designed to do so well. So with all this in mind, when we first “paused” the moving 4K footage from the 1DC, we were buzzing with anticipation.
The very first image we extracted from the footage via Quicktime’s export to still option was simply amazing. Right off the bat we had an image that was visually striking. The images held together very well in terms of colour reproduction (especially the skin tones) and were incredibly sharp. Right then we realised that a full scale shooting test was what we had to achieve.
The 1DC Specs
Physical body: What was refreshing was how familiar this new camera felt in my hands straight out of the box. Having shot with previous Canon DSLRS – the 5Dm2, 5Dm3 and 1Dx – the 1DC was a natural progression. It takes the physical weight and size that comes with being a Series 1 camera from Canon, however it still remains incredibly compact and versatile. The external chassis of the 1DC is the same as the 1DX, however, despite online discussions that it’s simply a firmware update that distinguishes the two cameras, my rep at Canon assures me that the boards, circuitry are different between the two.
Canon’s 1DC shoots 24 frames per second 4K video (4096 x 2160) as an 8 bit 4:2:2 Motion Jpeg. One of the biggest concerns raised early on was the absence of shooting in RAW. As an 8 bit Motion Jpeg, a reasonable amount of compression going on. So understandably there is far less information and manipulation possibilities in post than if you were shooting in RAW 4K. Currently, there is no other camera that shoots 4K RAW AND which has a comparable physical attributes as the 1DC. This of course perpetuates the saying “you never get everything you want from one camera”. I have no doubt of the next few years we’ll see advancements in the compression and roll out of RAW across many camera choices. Having said all that, we should remember that the 1DC does have a Canon LOG Gamma that dramatically improves the dynamic range of the motion images – 12.5 stops at 400 ISO. What records as quite a flat-image initially in playback, will enable far more detail to be retained in both the highlight and shadow areas. The rolling shutter effect and moiré, whilst still lingering, is less evident than in the previous Canon DSLRs I’ve shot with. There is a mixture of other recording formats available in-camera, however for the purposes of this blog, I’ll leave there discussion out.
The other Huge advantage that the 1DC has over its rival 4K cameras is the expanded sensitivity of the latest sensor technology from Canon. The new benchmark achieved by 5Dm3 in low light performance has been raised again here with very useable vision captured as high as 12,800 in our testing. This is a big point to note when comparing the low light nature of RED cameras and another reason this camera is the obvious choice for quickly changing environments where you have little or no time to “light”. The ability to “hot swap” our L series lenses and turn the camera on and begin shooting almost instantly shouldn’t be forgotten either.
Aspect ratio + Crop factor:
What was one of the biggest differences shooting with the camera was the new aspect ratio and crop factor. The 1DC records 4K video using the APS-H format (which is an effective crop factor of 1.3) We’ve been shooting on full frame sensors for a long time so it does take some getting used to the new crop factor when selecting which lens to go to. I absolutely fell in love with the 2:1 aspect ratio (slightly narrower than 16:9) which does (to me anyway) increase the “cinematic” look of the footage. We shouldn’t ignore several other key features such as Clean HDMI output (in HD), 18.1megapixel RAW still capability with the option of up to 12/14 frames per second depending on your shooting mode. All this is housed in an extremely rugged and compact chassis with the ability to run the camera using plug-in external power.
Whilst I was super excited to see and edit the 1DC’s footage playing on a 4K display, what screamed out to me the most to test and analyze was the camera’s ability to capture images using the motion. The prospect proving a new photographic tool was terribly exciting.
I’m a very hand-on shooter. And whilst i review the data sheets, the first thing i want to do is get into a real life shoot and put new gear, especially cameras, to the test. There is nothing like a real shoot also to gain real experience. Whilst controlled testing defiantly has its place, it can’t give you knowledge on the practical strengths and weaknesses a real-life shoot throws at you. So we planned to conduct four different shoots to test both the desired 4K images and the camera in a practical sense. These test-shoots encompassed four genres: Fashion, Wedding, Wildlife + Portrait.
Post Production – “the Wissam method”
After the completion of production it was time to take the mammoth amount of data and figure out a viable workflow. As there were no accompanying software with our 1DCs in box, and no set workflow available online to follow, we had some experimenting to do. The files off the CF card were as expected choppy at best during playback in Quicktime. Our early exports via Quicktime didn’t result in the best looking image, and we knew there had to be an easier solution. After some time we narrowed in on Adobe Premiere Pro CS6.
This updated NLE from Adobe has been taking some big strides lately and we found it simply ingested the footage via the RED 4K (2×1) project preset. (The final solution of simply using Premiere to export the still frames was suggested by Wissam Abdallah, hence the somewhat inflated recognition here) Once imported, the footage was playing back (albeit a little choppy) right there on the timeline.
Once the footage was arranged and culled to the shots we wanted to work with, we found opening each file separately in the source window and then expanding it full screen enabled us to much more clearly see the image in its entirety. Once full screen, the image could be fully appreciated. This then brings us to the most exciting and new aspect of this motion capture experiment: the selection of “the precise moment”
A photograph is all about capturing “the moment”. It is also what distinguishes it from cinematography. Photographers seek out images that best encapsulate the scene, person, place or time being examined. There are scores of different styles, techniques and processes that make up the rich world of which all falls under the one definition of capturing light. What’s always fascinated me is how many forms the art and science of photography takes. As well as the multitude of photographic techniques that are accepted and can be attributed to one person. Steve Winter’s award for “Wildlife Photographer of the year” was achieved by his remarkable image of a wild snow leopard was taken by one of his many camera traps. Even though he was not present at the time his amazing image was captured, he is still greatly recognized for the photograph – and rightly so. So how about an image, a great moment in time, acquired by simply plucking it out of free flowing video footage? Does this in any way not constitute a legitimate photograph? Does it matter?
Is this a photograph?
These are fascinating questions to me, and ones I wanted to hear people’s reactions to when looking through the motion-acquired images. What I quickly realised that some photographers held strong opinions on why not to pursue motion image capture, whilst others seemed quite excited by the prospect. The stigma of the idea of just pulling stills from video strangely gave some incredibly pessimistic views of the whole project. I don’t see motion image capture as the doom-and-gloom to traditional way of doing things. I just see it as a new tool for photographers to use (when appropriate) to capture even more precise moments in time. Technology shouldn’t hinder the creative process but further it. When we first started to “scrub” (film jargon meaning to shuttle through shots) the footage started seeing small moments in time, particularly human expressions, which were so slight they were not fully noted at the time of shooting. When the wedded couple spent time laughing and holding each other on the location shoot, the quickest of smirks or glances to each other were suddenly able to be slowed down, identified and selected. Similarly in the portrait shoot we were finding the greatest expression developed over several frames until the “choicest” of them all sat staring out at us. The ability to slow time and have all these usable frames was at the start, quite breathtaking. One reason people hire a professional photographer is that they appreciate their “taste” in recognizing, capturing and delivering the ideal moment. This post-event selection method is simply shifting some of these “recognition” skills to a later time.
Suspected limitations and pit-falls.
Our earliest tests highlighted the need for precise accuracy in focus and shutter speeds. When shooting films, we often pull focus (drift the plane of focus) as a storytelling method. You can draw the viewers eye from one part of the screen to the other. If shooting film for stills you would want to keep focus in check at all times. We found shooting with more depth of field preferable to ensure focus was hitting at all the critical times. Shooting video with the 1DC means manual focusing, which takes time to perfect if following subjects.
The other major consideration is that film is through-out its history always been filmed in landscape orientation. So what does this mean for the portrait shot? I guess it all comes down to whether you are filming first with the idea to extract stills later, or if you are just using video mode as a means to aid purely your photography. It may be harder to imagine including the vertical footage in a finished edit, however, one only has to look to the work of Alexx Henry and his inspirational “motion art” concept shoots to see the new ways in which motion and stills (both horizontal and vertical) can propel the creative process.
As discussed above, it is a motion jpeg so the need to correctly set your exposure and white balance manually is imperative. Having shot video with DSLR’s for over 4 years now, i like many others have been forced to shoot extremely cleanly. This is down to the fact that we just haven’t had the latitude later in post to correct terribly under or over exposed clips. This training looks like it may prove invaluable when shooting for motion images with the 1DC. (further testing with Canon Log will determine the true expansion of the dynamic range)
When it comes to shutter speeds, traditionally cinema has stuck to the 2:1 rule with frame rates. When in Australia we shoot at 25FPS (frames per second) we double our shutter speed to 1/50 of a second. This ratio best retains a slight motion blur within the frames and reproduces a aesthetically appealing “cinema look” to the motion. If you shoot at much high shutter speeds with a 25FPS shutter rate, the resulting playback takes on a very staccato look. (it should be noted that this high-energy strobe look can be desirable on certain occasions. It can be seen to great effect during the open Normandy beach-landing scenes in Spielberg’s “Saving Private Ryan”). Photographers rely on higher shutterspeeds to freeze the action in pin-sharp form. With the testing we have done during this project, we opted to find a compromise between film and stills. The happy middle ground seemed to be 1/100 and 1/200 shutter speeds. This mostly retained the “cinematic feel” to the motion whilst delivering more useable still frames for export. If no consideration were to be made for the video being captured, then higher shutter speeds may be valid to use. We did find many useable frames from all four of our test shoots (including the most demanding – the Sue Bryce fashion shoot with the free flowing dress) and each provided many useable stills.
As predicted, even without shooting RAW 4K, the 1DC produced buckets of data. With recording times requiring approx 1GiG of memory per 15 seconds of recording time! So a commonly used 16G card now captures just 4minutes of 4K. (at least that’s easy to remember!) So naturally, larger capacity, and faster performing Compact Flash cards are necessary. We were fortunate enough to have SanDisk provide us with two new 128G 100mbs CF cards for this project. Each of these 128G cards could hold 30minutes of 4K footage. The camera offers dual CF slots, and plenty of reasons to start saving for memory. For those lamenting the cost of high capacity media, remember cost will inevitably come down. I remember when first shooting with the 5Dm2 in 2009, 8 + 16G cards were prohibitively expensive and with time costs became much more bearable. History repeats itself.
Power-hungry processing: Anyone who has worked with 4K video will tell you of the massive grunt needed for playback and working with such file sizes. (We shot a commercial for a prominent Australian designer on the RED Epic earlier in the year and experienced the pain first hand.) Not only will larger hard drives be required, but more powerful PC or Mac’s to drive them. I should note we conducted this project using only a 17inch ACER laptop with an Intel i7 2GHz processer with 16G of Ram. From the impressive performance and playback of the footage on this laptop, I can only assume ever better results on a beefed-up desktop PC or MAC.
Retouching and Printing
Once the still had been selected from the video editing program we needed to choose an output file type. The TIFF file (around 33mb) that Adobe Premier Pro produced the best results. As the original captured footage is motion jpeg I suspect that the TIFF file is just an enlarged Jpeg. The first impression of the image out of the video environment was really positive. There were very little indicators that this was a freeze frame from video footage. The two most noticeable were the overexposed highlights and an overall sharpness that resembled an image that had been over sharpened incorrectly in Photoshop. Both of these things could be fine-tuned with internal camera settings and post production. The blown-out highlights was noticeable in only a few outputted frames and as we were using the “out of the box” picture style settings I’m confident a basic reduction in camera contrast would help to improve this. This goes the same for sharpness. Overall the sharpness gives a pleasing result, though some may find the default settings a little too over sharpened. Tweak to your pleasure but remember this isn’t a raw file so getting it right in camera will really help to getting it right on paper.
We really didn’t have to do too much to the images to get them print ready. The prints for the wedding album had only had a tiny amount of sharpening added as we wanted the test prints to really pop. Un-sharp mask filter was used but we inverted the default settings to be [Amount 1, Radius 0.2]. This gave a finer pixel contrast ratio and an overall realistic sharpness. From there a simple Auto Colour Balance and the images were almost ready to print. Basic retouching and personal touches were applied to some test prints, just like a photographer would do with regular still images.
So, the workflow from Video Still Output to Printing is exactly the same as if it were captured as a still. The key to a better image is to reduce the contrast and sharpness in the camera settings to produce a more usable image with as much data as we can get. The obvious recommendation is to use Canon Log in camera and retain as much detail and information as possible. Due to the files being so sharp, we were able to interpolate them in Photoshop and enlarge sharp prints up to 55 inches. I mean sharp prints! This size is more than acceptable for fine wall prints or exhibition. These image files would be more than usable for commercial signage and a great tool for TV commercials or movie production. Like with original photos, if it’s sharp to start with it will enlarge well too.
The Photographers’ critique.
We realized that the best and harshest critics to show these motion-acquired images would be photographers themselves. So on the 19th of December we invited some of Australia’s leading photographers to Sun Studios in Sydney Australia for a look. Reactions from each photographer varied from shock and amazement to almost disbelief. Understandably the discussion of would this negatively affect business was raised, but after some reflection everybody agreed that this is simply and exciting new tool for photographers and should not be feared. The art and skill of a photographer is still required when using a camera like the 1DC. Understanding and harnessing of light, composition and interaction with your subjects are all vital skills of a photographer and are not replaced by the idea of motion image capture. Photographers also use a variety of techniques to obtain unique looking images (like long exposure times and the use of remote flashes) these times of images would not be reproduced in video. I see the biggest step forward using motion image capture the ability to record many individual moments in time, all the while silently as there is no shutter being released. This could have great benefits in situations where you may want to remain more candid. Subjects could also feel more relaxed not knowing “photographs” are being taken.
Whatever your stance, I think most would agree there are fascinating times ahead!
Credits + Thanks:
Firstly a huge thanks to our stellar untitled editor Hayley Yeoh for putting together the film in such a short time frame before Christmas. Also the professional image makers who gave of their time to be interviewed who included: Dean Bentick (inlighten photography) Brett Purmel, David Stowe (Society photography), Ryan Schembri (Xsight) Sue Bryce, Philip Bloom, Wissam Abdallah, Graham Monro (gm photographics) Michael Martin (MM Photos). Thank you to Sun Studios + Peter Osborne for opening their doors to us for shooting and printing. Others to thank for your support and encouragement: Jay Collier and Selena Simpson, Rachael Bentick, Paul Stewart, Charles Montesin, Sam Hannaford, Steven Khalil, Helen Giovas Sotiropoulos, Pippa Walton + Ray Schembri, Erica Salmon, Kelly + James Rutty, Laekin + Chris Rose, Ben Joffe, Untitled film works crew + support: Andrew Prochuk, Turei Cooze, Kylie Lewis, Lorna-Jean Bradley, Matt Teague, Volkan Dogan, John Alten – Sandisk Australia. The biggest thank you of all goes to my darling fiance Jen who has put up with the ridiculous hours of absence leading up to Christmas working on this project.
I wish to note that this project was produced independently and no payments or influence was made from any company, including Canon. The opinions expressed in this film and blog are of the individuals. But after saying that, I wish to sincerely thank Canon Australia for trusting us with two 1DC’s for the time it took to produce the film and blog.
Facebook: untitled film works