Saturday, June 28, 2014

Shooting Full HD with an iPhone

I've been reading about efforts by professional indy cinematographers to shoot entire feature films or documentaries with an iOS device. It sounded absurd to say the least. Then, the shock came when Zacuto of USA recently sponsored an all shoot-out among the greatest camera names in professional digital cinematography with the iPhone as one of the contenders. Out of sheer curiosity I watched the entire project on YT to see pros at work and experience the difference between capturing gear of a 'toy' costing less than a thousand bucks with all those industry monsters, costing several tens of thousands dollars. Of course, to avoid shake and be able to shoot like the pros, they dressed the iPhone in all sorts of camera rigs, with matte boxes and filters and French flags. If they could mount a follow focus they'd do it too, but the iPhone is supposed to be an iPhone last I heard, so, let's be serious. The iPhone eventually managed the feat with its built-in autofocus, as expected. Anyways, the final result was more than descent, yet far below the big dogs, like Red Epic, Arri Alexa, Sony F65, and Black Magic Cinema. Footage was shot all in Full HD, ie. 1080p at the presumably cinematic 24 fps.

This morning I read another article about cinematographers getting again serious about iOS feature film productions, and I said to myself, what the heck, let's try to see what comes out of an iOS footage. I never really shot anything longer than ten seconds before, and was quick to delete it right away, after a few moments... This time however, I wanted to  live the experience on my own big-ass TV in full color FHD. So, I'd have to postprocess the footage like it was coming out of any of my other prosumer gear. I'm not a pro by far, but I can stitch together simple shots with final rendered clips that look fairly OK, technically speaking. Needless to say, I dispose of quasi pro gear, like a 5D M-III, the latest 4K GH4 and a Legria G30, even a kit of cheap Samyang cinematographic lenses, and a shedload of supporting accessories, a steady camera-like thing (Merlin), couple of motorised sliders, a Manfrotto video tripod with fluid head, a motorised pan/tilt head, and full versions of FCP, Motion and Compressor. I just wanted to see for myself how far I could go compared to my other 'prosumer' footage and clips. Especially the couple I did the last few days in 4K, the moment the Lumix GH4 marvel arrived.

I thus put my old iPhone 5 (not an S, neither a C) in my shirt pocket and came out in my backyard for a shoot. I actually shot about two dozen takes without steadycam support. Just in handheld. I tried to simulate a few dolly movements and did some pans and tilts, all handheld, trust me. Forgot to mention, I used Filmic Pro and not Apple's own camera app. Filmic is quite good as you can chose separate points for measuring light and defining focus reference points. You can also lock those, and you can separately select output resolution and frame rates. Almost like the big lads. I eventually transferred all those clips to my quad 27" iMac via the Photo transfer app, a very neat way to transfer photographic source files between iOS filmrolls and Macs.

Reading those same source files in FCP, I transcoded  them in Prores Proxy (I always do that, unless I capture in ProRes 422 with my Ninja Blade) and stitched the shots together by adding here and there dissolve transitions, trimmed a great deal, you know... the usual. I tried stabilisation on all of the shots, but some were hopelessly shaky and the FCP algorithms made a mess eventually, so I either dropped those out of the Timeline or softened further the stabilisation parameters. I eventually added a nice track from 'The Grandmaster' soundtrack (Wai Kar Wong - 2013), namely Moyou by Shigeru Umebayashi. And... Bob's your uncle. Oh, yes, before compressing to a final delivery format, I combined everything into a compound clip and I graded the latter with FilmConvert, a plugin that runs under the 'Effects' tab of FCP. Normally, to judge objectively the true quality of iPhone footage one shouldn't do too many tricks in post. But I am not a formal iPhone reviewer. IAnd never intended to be. I just wanted to see what comes out of it, if I did my regular stuff like in any other workflow I normally follow in post.

My personal conclusion is that indeed, one could actually put an iOS device in the hands of professional cinematographers/film makers to produce something worthwhile. Using purpose made add-on lenses, they could could even achieve a touch of cinematic bokeh. And with the proper grading in post one can achieve the sphere that serves script and story. With a good story and plot, lighting, art direction, camera movements and editing techniques, one could actually create a technically acceptable feature indy film that can proudly stand next to a Hollywood establishment with their millions worth big production budgets. Maybe some cheap iOS gear will offer the opportunity to realise a filmic form of Arte Povera... Like Polaroid once did. Arte Povera is still good Arte! Why not?

Monday, June 23, 2014

Video shooting adventures...

I've been shooting a lot of video, from when my kids were still young, some 20+ years ago. Video has come a long way in the meantime, and, in the last 10 years with the advent of the so-called VDSLR based videography, ie. using traditional DSLRs to shoot video (because of their large sensors and pixel size they are performing quite well in low light conditions and are capable of cinematic bokehs), low budget videographers and indy shooters have flooded YT and Vimeo with quite a few interesting projects. Another milestone in the democratization of the medium. The World is flat sort of thing.

One of the nice 'little' things that occurred in digital motion pictures is pixel resolution. I picked up this graph on Wikipedia, and even so, I can't quite grasp the size of it. What is referred to as SD (standard definition) is the sort of resolution used for broadcasting by TV networks for many-many years, and made us, amateur videographers of the late last century, drool of desire for a better amateur technology. Indeed, our quite expensive analog gear back then could hardly produce 400 lines of horizontal resolution at best. Not to mention stability of edit footage after a few cycles in analog mixing tables without genloc and timecode available (far too expensive for an amateur's pockets).

In the last ten years, however, the world has been converted with lightning speed into FHD (1080p) or 1080 lines of 'progressive' (one frame at the time) resolution, and in recent years key industry players and Standards bodies have defined additional resolution targets all the way up to 8K UHD (Ultra High Definition). Actually, anything above FHD is nowadays termed UHD! To get the point, 8K is like stacking 16 Full HDTV monitors upon each other, four in the x axis and another four in the y. In other words, if you regularly indulge yourself by watching Bluray movies in FHD, imagine the monster I just described... what it would turn your living room wall into! Already 4K is an unbelievable improvement. It's your current FHD doubled in both x and y directions. Four times the resolution you enjoy today, that is. Strictly speaking, hardcore 4K is even more than that in the x axis pixel-count (4096 indeed), but for all practical purposes let's assume the 3840x2160 frame size (that's otherwise 4x the current FHD, 2x1920 and 2x1080) under the broadly known 4K label.

I recently saw the picture of a camera shooting 8K, but I doubt there's much activity in that front yet, other than experimental. On the other hand, I read that quite a few mainstream movies being shot these days are already captured in 5K/6K and 12/14 bits of colour resolution. This sounds, spec-wise, like a videographer's wet dream... Why do them pro's do that, then? The film industry still delivers feature films and TV series in current consumer formats (FHD, Bluray and DVD), meaning they encode/downscale their UHD masters into current commercial resolutions, but, I guess, they are shooting UHD for future-proofing and the ability to remaster their work into UHD resolutions when the market is ready for it. In the next so many years... When storage petabytes become as common and inexpensive as bagels and donuts...Understandable...

I consider myself to be a frame resolution and sharpness junkie. Actually, I don't give a diddly squat about resolution per se, if you come to think of it, but I care enormously about image sharpness, richness of colours (gamut) and related bokeh. UHD resolutions are the unique available strategy in the pursue of this goal, regrettably. The prerequisite in preserving the tonal and color state of visible things into their captured image is of course the frame resolution, but not alone. Recording codecs are natural born quality killers too. You can't believe how many codecs already exist, and how many different media file containers are out there used in the video universe. Gives you practically a headache. Makes you wonder why! And they all claim to be the best invention ever since sliced bread. For amateur videographers like myself the world of codecs used for capturing, post-processing and eventually delivering the outcome to viewers is a sheer nightmare. The moment you get the impression that you 'got it', there comes a use-case experience that entirely confuses you, frustrates you and makes you start all over again.

The last few months I decided to further develop my skills as close as possible to today's State of the Art in prosumer videography. Bought some new gear to work with, rigs, accessories, even a Canon Legria FHD G30 camcorder (I wish I knew some more before making that decision... it's definitely ok, but not what I dreamed of), filters, video-mics, tripods, sliders, a pan/tilt electrical head, remote triggers, and only came short of ordering cranes and drones! A year ago I also bought a Hero3 GoPro action camera with all its usual-suspect accessories, but this must have been my worst decision ever. Among many other. The GoPro's are for a different type of videography than mine. Like they say, too old for that shite! Anyways, I normally use VDSLRs, a Canon 5D Mark III and another 5D Mark II, and maybe my old 7D occasionally, but I still dream of the day that I'll get my hands on a 4K Panasonic Lumix GH4. Pity that they only make them mirror-less MFT (micro four thirds), necessitating some sort of adaptor to be able to use my couple of dozen piece Canon and Nikon lens inventory that I built the last 30 years. You can't have it all, can you? Unfortunately, the little marvel is still out of stock in all my online supplier inventories.

I also improved my grading and NLE skills in both FCP and Da Vinci Resolve, by following specialized training (about a dozen courses the last few months alone), and I came that close to order a Black Magic 4K Production camera, which idea I eventually gave up for a planned purchase of the Panasonic mentioned earlier. Tomorrow, I'm expecting a Samyang Cinema lens kit, insanely inexpensive but fairly good obviously, at least if you believe online reviews and YT demo's. The kit contains three lenses, 14, 35 and 85 mm in Canon EF mount.

I can't shoot 4K yet, but I've still done one 4K video project in time-lapse, by capturing high res Raw stills in my Mark III, then crop-processed them in Lightroom, and imported them in a FCP 4K project, from which I eventually rendered them in the 4K format output acceptable by YT. All I had to do was export the frames from Lightroom in the genuine 4090x2160 4K resolution and compose them into a clip in FCP and encode them via Compressor. Take a look here:

Like many out there I tend to agree that although 4K is a thing for when monitors will be able to handle it in the next 3-4 years, still cameras shooting 4K today produce visibly superior quality in FHD than many regular FHD camcorder out there, regardless their robustness, perceived quality and total configuration cost, pro codec used and frame rates captured. I speak from experience. I do shoot with the best Canon glass available on the 5D, capture the HDMI signal in Prores HQ on an Atomos Ninja Blade, and proces all this footage on an 27 inch top of the range iMac with FCP X. Nevertheless, although my 1080p outputs, color corrected, graded and all look reasonably good, 4K is simply miles ahead. I'll be the first to admit that my 1080p clips are indeed nowhere compared to some of the GH4 footage I saw on YT! Goes without saying, many other factors do play quality determinant roles as well, like the dynamic range captured (per a recent DXO-Pro testimony the 5D is rather poor on DR with only 12 stops and change), ISO used, shutter angle and above all the quality of light used in a scene. However, from what I've seen, 4K is still a blast! I mean, big time!

There are moments I feel that I could give my right arm for the ability to capture quality footage like some of the high-end cameras out there. RED, Arri Alexa, the Sony F5, Phantom and why not, the Black Magic 4K Production Camera. Of course most of them use glass in practical configurations of prime and zoom lenses at a cost equivalent of high-end luxury cars. Would you rather buy a 5- lens kit or an S-Class Mercedes Benz? You could only practically rent gear like this. As a B-day present, for instance. Like many rent a Ferrari for a day! I drool watching slow-mo footage shot by RED and Phantom cameras! Gosh, I've been in the wrong job all my life. Day-dreaming of one of these 4K marvels, I keep fooling myself in the meantime with tricks like the one I used in the 4K time-lapse above, and by doing simple slideshows. See an example hereunder :

Sunday, June 8, 2014

Storehouse visual stories

I just discovered them during the recent Apple Design Awards ceremony, as Apple selected them as one of the 2014 Winners. Storehouse is a simple business model by a startup in which free members publish online stories composed with texts, photographs and short videos (max 30 secs). Stories are published online and viewers can read them and like and/or comment on them. Also share them with friends thru the usual suspect social networks. Your media sources are either your local photo libraries or a few media sharing networks like Flickr and Instagram. In other words, Storehouse is a pleasant extension to those other networks who weren't clever enough to come up with the idea. Flickr, for instance could have implemented this long ago.

The app is at its best on a tablet, for both reading and creating stories. A browser is a viable alternative to those who don't own a tablet. Everything is free for the moment but maybe freemium models are down the road. They just secured a second round of VC funding of 7 million bucks they say; so I am sure they are seriously thinking about monetisation as we speak, otherwise no VC would be that interested. Maybe VCs are attracted by the 'hub' effect and a sale down the road to a bigger fish. Who knows? For the time being all we can do is enjoy their service... for free! 

If you come to think about it, Storehouse is no more than a simple blogging facility with an attractive user experience. Nothing more, nothing less. However, the experience has still its merits and I am sure they are thinking about / or already working on possible extensions. For instance, they might offer a function of assembling books with your stories and publish them via mechanisms like Blurb/Lulu and the likes, or by themselves. They could also print them in ways that you could hung them on walls, poster-like. They could maintain privacy levels so that a published story is only viewable by those you personally select, like family and friends. A great way to share experiences. Especially with a touch of unedited video. You can say a whole lot in 30 secs, you see. Most TV ads are shorter than that. There are so many things you can do with something like this. Funny it's only Storehouse that came up with the idea. Cool stuff.

Tuesday, June 3, 2014


I just bought me a Gigapan Epic Pro and created a first stitch of a Panorama as seen from my bedroom window with 21 pics (3 rows x 7 columns). Used the 100mm Canon macro lens, a 5D Mark II body, and I post processed the images in Lightroom first. I then stitched them inside the Gigapan Stitching app and uploaded the result to their hosted galleries. It's not much of a Pan, other than the workflow worked just fine first time, and the output seems technically correct (that is, no vignette effects and no stitching artefacts). It was only a 100mm tele and a bare 21 pictures (child-play), so you wouldn't expect to find bees and butterflies on flowers and leaves from a 200m distance, would you? I'm still learning, you see... Next project, using my 70-200mm 2.8 zoom with a 2x extender (if my math is right, that would yield 400mm focal equivalent) and a few hundred high res raw frames for stitching. I am quite curious about the result... Only problem, I have no idea what to shoot!! Forrests and trees? My next door neighbour's house and garden? Climb the MAS and shoot the Metropole? The Atomium? The hills of the Flemish Ardennes? The Belfort in Ghent? The botanic gardens in Meise? Brugge die schone? Het Zwin? If you got any suggestions, you are welcome...

Update: Here's a brand new from this morning (4/6/14). Done with 120 frames in 15 columns and 8 rows. Not too bad but some problems with autofocusing. Some stitching artefacts here and there too. For instance, a ghost car at the far end of Aalststraat as it happened to pass by in the wrong moment. Can you spot a flying black bird somewhere? This time I used the 400mm focal. The shots were too overexposed, and the atmosphere was rainy. Overexposure was justified by the necessary dynamic range to capture detail in the shadow parts of the surrounding trees of the Liefmans Brewery.