The Giant Flea – Camera Test Movie IV

Below is appended my fourth camera test movie, entitled The Giant Flea. Copyright © 2017 Gorilla Film Studios. It was recorded, edited, mixed and graded by yours truly, Tomas Stacewicz. Music by courtesy of Amok Cell, one of the stage names used by my friend Jonas Olsen. The movie is 5:48 minutes long and was published on YouTube on 19 June 2017. The Giant Flea is a very basic test film with no artisic pretentions at all. It was shot entirely handheld on CinemaDNG RAW with the Blackmagic Pocket Cinema Camera and the Zenit 16 OKS 3-10-1 prime lens, using the RafCamera Kinor-16SX-2 lens to MFT camera mount adapter (with screws), and my latest DIY handheld shoulder rig based on the POOLi™ D-BASE-2 base plate and Zenit Krasnogorsk-3 pistol grip and shoulder brace and SmallRig handle. The movie was filmed on location in Kungsladugård , Göteborg, Sweden, on May 28, 2017. General camera settings: ISO 800, Shutter Angle 172.8º, 24 fps, and White Balance 6647K. General lens settings: F=10, NDx4 and f/ stopped down to approximately 8-16. General audio settings: All sound were recorded on location using the inboard stereo microphone of the Blackmagic Pocket Cinema Camera, with microphone and channel (line) levels set to 100%. Postproduction: Edited, recorded, pick-up sound mixed with Amok Cell’s music, and colour corrected (without any stabilisation) on Blackmagic DaVinci Resolve 12.5.5. The final result was rendered using the DNxHR 444 (12-bit 4:4:4) video compression codec and two cannel PCM audio codec wrapped in QuickTime and Ultra HD 4K 3840 × 2160 resolution. Exported thus, it was finally uploaded to YouTube, enhancing the image with a Saturation of +1 with no additional image manipulation to retain the flavour of the original footage.

The main subjet of the test was my DIY shoulder rig as well as the 16 OKS 3-10-1 10mm f/2.1 lens with attached NDx4 filter. I took the opportunity of caching some fleeting moments of the huge flea market that is annually arranged in my neighbourhood. The editing and grading has been consciously toned down to catch the performance of my test subjects. It is not intended to be an example of my artisitc qualities (if I have any); at least, the film wasn’t shot with that intention held in my mind. I didn’t want to distract the actual testing of the gear from any fancy filming. Thus, it is not one of my official short movies; these are recognized with the Dogma 16 sign at the beginning of each short. Still, I’m in the habit of naming all of my films, regardless of intention and ambition. I decided to name this test movie The Giant Flea as a pun on the flea market which takes up an entire sector of Gothenburg City. I consciously picket that particular composition by Amok Cell, entitled Mother interior, as it reminded me of jumping fleas. A part of me also wanted to fool anyone thinking it was a monster B-movie ☺ Sorry for that one. Perhaps I did piss a few people off, for various reasons, as the film has only received negative opinions and “dislikes”. The choice of music might have been a bad call on my behalf, as I confess that it does get somewhat enoying to hear after a while, which might have putt off some watchers. It does distract from the test footage, but at the other hand I do like the music regardless.

A screenshot of a RAW capture from the movie featuring a basket of fruits

Camera and Lens: So, as I mentioned, the main or almost only motif with this little film is to test out the optical qualities of the lens in a real life shoot condition. Though, I cannot resist also make some few remarks on the camera itself, the Blackmagic Pocket Cinema Camera. As usual, the images coming from the BMPCC Digital Super 16 sensor is very filmic and organic. Although the colouring in post has been consciously minimal and neutral, the colours produced by this camera seem natural and quite saturated. I did notice aliasing coming from that small sized sensor though, in a couple or three instances. The moire is clearly visible against the walls of two buildings, a brick wall at 0:37 and in particular a wooden wall with vertical patterns at 1:10. The Zenit 16 OKS 3-10-1 10mm f/2.1 prime lens really pleases me in its optical quality. The chromatic abberation is moderate and it gives a neutral and sharp, almost Zeiss-like image of a professional quality, at least compared to my Zenit Meteor 5-1 17-69mm f/1.9 zoom lens. The big issue is of course the vignetting; it simply doesn’t cover the Super 16 sized sensor of the BMPCC. Hence the cropping of the image to 2.40:1, and even there I have to zoom in the image in post to remove vignetting entirely. However, I didn’t use any step-up ring this time to attach my Zenit Meteor 5-1 lens hood (a lesson learnt from my previous project, A Curious Day at Snooper’s Haven), which at least didn’t create an enormous vignetting this time around. Another test I did was to check out how well my cleaning of the lens and its ND-filters, using the Zeiss Lens Cleaning Kit, fared. Although I didn’t spot the large optical defect created by the spot on the back optics of the lens, which I did notice on my last test movie, unfortunately my NDx4 filter (belonging to the Soviet Kinor-16 kit) still showed optical artefacts in the form of small polygons of light, although less than before. The conclusion is that the cleaning itself worked out fine, but that the ND-filter is damaged beyond repair, as its tiny bubbles still distorts the light. This is only shewn when pointing the camera towards the sun, only seen in a couple or three instance, such as in 1:28 and 1:57. Some of that intermingles with actual lens flare and as the effect is quite similar (as seen in this lens flare at 1:57) it is hard to distinguish which is which.

BMPCC Shoulder Rig (front) prepared for the shoot

The Shoulder Rig: The main object with shooting this film was to test the functionality of my latest do-it-yourself shoulder rig. In assembling it I have used components from Zenit (Krasnogorsk-3 Pistol Grip and Telescopic Shoulder Brace), Manfrotto (088LBP Female 1/4″-20 to Male 3/8″ Adapter with Flange) and POOLi™ (D-BASE-2 and POOLiCAGE), as well as the following products from SmallRig: 1795 1/4″ Fixing Screw with D-Ring fitted to the 1087 Rod Clamp and screwed to the 1622 Rubber Non-slip Handle Grip. The BMPCC is always mounted to the POOLiCAGE half case, and when shooting I always attach the SmallRig 1638 Top Handle and cold shoe base, as an extra support. Assembling everything, I put a small piece of black felt tape over the chromed D-ring that is seen on top of the rod clamp, to prevent reflections to the front element of the lens. Although feeling somewhat loose, as I still have a hard time fitting the K-3 pistol grip to the POOLi™ D-BASE-2 base plate, even though I have attached rubber non-slid pads on top of the grip at the junction to the base plate, it moves a little revolving slightly. But during the shoot I wasn’t bothered by it at all, with no need to rescrew or anything like that (well perhaps maybe once or so); it’s assembly works good enough. The top handle really served me well during the shoot, as I could switch easily from ready-to-shoot position to a more relaxing position, carrying the rig from the top handle, relieving stress to my arm.

BMPCC Shoulder Rig (side) prepared for the shoot

When actually shooting, I always hold the rig with my right hand on the K-3 pistol grip and my left hand on the SmallRig forward grip, resembling shooting with an Assault Rifle. It feels a bit menacing to carry the rig in public; somewhat afraid that people (or the cops) will take me for a terrorist. But it does really provide for a much more steady hand compared to before, with pretty decent overall results if you ask me. No stabilisation at all made in post. Shaky, yeah. But not that much micro-shaking going on here compared to what I have been used to. The film is more or less edited in the order of the shots taken in cronological order (with the exceptions of the early shots of the children making barter which was the last shot that I did); the viewer will probably notice that the shakes become less as the film progresses, as I learnt how to handle the rig better after a while. I’m in particular pleased with how nice the results came out with the following and walking camera (beginning at 3:24); I used good sport shoes and did the stedicam heal-to-toe walk. I imagine that this type of footage will make it easy for me to stabilise in post. On this shoot I also realized that I could extract the shoulder brace somewhat longer than that I was aware of, just by pulling harder; nice surprise! The only thing I miss is a remote LANC-trigger attached to the pistol grip; it’s a bit ankward to turn on the camera and then go to the steady assault rifle positon, and back again, resulting in spoiled footage at the start and end of each take.

A screenshot of a RAW capture from the movie featuring a walking camera shot

Workstation and DaVinci: I simplified the process in post as much as I could. I just did a basic colour correction in Camera Raw retaining the in camera settings, according to the follwing: ISO 800, Color Temp 6647K and Tint 13.82. I lifted Gain +2 and lowered Lift -1.6 to make a basic streach of the image information for later adjustments in Grading. In Camera Raw I set the Saturation to 90% with no Color Boost; in the Colour Wheel tab I left the Saturation at 50%. In this way I made shure to present the image coming through the lens and from the sensor as faithfully as possible, to be better able to judge the character of the 16 OKS 3-10-1 lens. Then I proceeded to adjust the colour slightly with the Color Wheels. To refrain from any stylistic Grading or any such, I left the Offset wheel alone. I didn’t touch the Gamma wheel for the midtones much either. I simply adjusted the Gain (highlights) and Lift (shadows) to balance the individual colours (Green, Blue and Red), and to create a uniform look throughout the film. Doing the colour correction I learnt one thing about exposure: Low exposure gives you a lower contrast to the overall image when you lift the midtones up in post. ETTR (Exposure To The Right), i.e. slight overexposure, creates much more contrast in the image, making more use of the apparent dynamic range. Simply compare the image in 3:52 with that of 4:03, the former being overexposed in camera (and lowered in post) and the latter underexposed (and raised in post), giving a greyish and flat image in the latter case compared to the high contrast in the former.

Ramping the volume in DaVinci Resolve 12.5.5

Because of the heavy vignetting of the lens I had to zoom, pan and tilt the image in post even though I set the Output Blanking to 2.40:1, according to the following settings: Zoom 5.2 to 12% (averaging 8%), not that much actually compared to my previous project using the same lens. Pan -0.5 to +17 and Tilt -110.5 to -30 (generally -30). I did not stabilisation using the Tracking Window, as I wanted the film to foreshow the stability (or the lack thereof) of the sholder rig, which meant that no additional zooming was required. In the edit page I did a very basic and simple cut of the movie, more or less in a chronological order. No fancy stuff to report here. In sound mixing I generally lowered the volume a bit from the take-up sound, as it feels quite thin and lacking dynamics. I also wanted to have the music score in the foreground. There was one exception though, in the scene with the guitar player (at 3:01) where I found the music score mixing bad with the guitar licks, so I lowered the soundtrack for a few seconds right-clicking on the audio clip, pressing “Volume” in the bar at the left upper corner and enabling “Volume” in the checker box. Then I simply created four points (by pressing the <•> symbol) along the line (of the volume level) and adjusted the level of the volume by lowering with the mouse between the two middle points. This is very similar to the process of right-clicking on video clips and enabling Retime Speed. All in all, I don’t remember DaVinci Resolve 12.5.5 crashing or lagging, perhaps a few times. The post production went smooth and faster than it used to do because of the restrained manipulation of the image.


Leave a comment