Digital Film

When I started out in this adventurous business of independent guerilla style filmmaking back in 2006 there was a true dichotomy current between shooting movies with film or in video. With “film” we always meant moving images captured on photochemical celluloid emulsion filmstock, using such formats as Super 8, 16 mm, Super 16, 35 mm, and Super 35. With “video” we referred to electronic signals translating light captured on a sensor and then transferred to tape on camcorders, either analogue video (such as Betacam, VHS-C, Video8, S-VHS and Hi8) or digital video (such as DV, HDV and DVCPRO). This was shortly before the RED One had started to revolutionise the digital cinema technology and format, and well before the introduction of DSLR’s with video recording capacity. For me Video was vastly much inferior to Film because of the lack of proper colour bit depth and natural dynamic range compared to negative film emulsion that in the latter case has as much as 16 bits of colour latitude and 16 stops of dynamic contrast range (but approximately 12 bits and 12 stops as an average); video being highly compressed produced 1/3 of that at best. I even preferred (and still prefer) the Super8 film format compared to HDV.

The DSLR revolution made that contrast between Film and Video less pronounced, shooting 1080p in formats such as MPG4 with the H.264 compression codex. Still, there was Celluloid Emulsion Film and there was DSLR Video, visually differentiated, although Video actually started to look good with a 8-bit colour latitude and 10 stops of dynamic range, at least better than Super8, but still not better than 16 mm. Simultaneously, Silicon Images, and in particular RED Digital, turned the game around in introducing Digital Cinematography (Digital Cinema) in a more or less uncompressed RAW format. Although the Silicon Images SI-2K and SI-2K Mini predated RED One by a whole year, introducing a Super 16 sized CMOS sensor, both cameras were highly patented and still beyond the reach of guerilla filmmakers in 2007 because of the high price tag. In a similar manner that 35 mm film was above that same limit for half of a Century, confining many independent auteurs to shooting in 16 mm, the RED One Digital Cinema camera used a Super 35 sized CMOS sensor and cinema 35 mm lenses. The world still lacked the equivalent of a affordable digital 16 mm camera that shot in RAW.

That deficiency was remedied with the introduction of the Swedish Ikonoskop A-Cam dII in 2008, the first camera to have a analouge CCD 1080P sensor with the size equivalent of a Super 16 gate (actually approximating Super 16, at it was somewhat smaller than that) shooting in the Cinema Digital Negative RAW format (CinemaDNG), a flat and desaturated format containing much detail information in highlights and shadows to be extracted in post. The Digital Bolex D16  jumped on that bandwagon in 2012, one year prior to the Ikonoskop being discontinued, using a slightly larger CCD 2K sensor exactly matching the Super 16 size. Simultaneously, Blackmagic Design introduced its first camera, the Blackmagic Cinema Camera or BMCC. Like the Digital Bolex, the BMCC also utilised the Adobe open file CinemaDNG format using a slightly larger 2.5K digital sCMOS sensor than that of a Super 16. In 2013, Blackmagic Design introduced the Blackmagic Pocket Cinema Camera (BMPCC), shooting RAW CinemaDNG captured from a true Super 16 sized 1080P sCMOS sensor for the sensational price of $995 (compare that to the $9000 price tag of a Ikonoskop)!

Since the introduction of the legendary Ikonoskop A-Cam dII, and its important legacy preserved through the Digital Bolex D16, and now the BMCC and BMPCC, we no longer can talk of Film as being the sole prerogative of celluloid emulsion photochemistry. Today we can also talk about Digital Film which captures images in a digital negative format with 13 stops of dynamic range and uncompressed 12-bit colour depth, which is hardly distinguishable from celluloid emulsion. Considering that all exposed and developed celluloid film stock nowadays is scanned into a digital format and worked through a digital intemediate in postproduction computer editing, grading and rendition processes, the actual images projected on a contemporary digital 2K cinema are equivalent when compared between “Digital” and “Celluloid” Film. The fact remains that my small Blackmagic Pocket Cinema Camera is capable of producing a look and emotional tone to any audience watching a movie in a digital theatre that would match that same movie if it was shot on celluloid Super 16 instead; in all respects, it is truly a Super 16 Digital Film camera.

There is still one big difference though between the image produced from a Super 16 celluloid emulsion film camera and that of a Super 16 digital film camera, that cannot be bypassed. It has to do with grain structure of celluloid film emulsion which cannot be emulated in a digital film format. This is because a Super 16 sCMOS sensor, as that of the BMPCC, uses a pixel structure arranged in 1920 neat columns and 1080 equally neat rows, with 2.073.600 (2K or megapixels) of equally sized square pixels or photosites, whereas the grain structure of film emulsion is arranged haphazardly with each irregularily shaped grain being of different size. This latter phenomenon creates that distinctly filmic sharp look which at the same time is pleasently soft, which enhances reality and makes it look more beautiful, in particular human faces and skin tones as well as details of the face. Contrary to this, Digital Cinema Cameras using Super 35 sized sensors with resolutions of 4K and above only create a sharp image that looks uncompromising, surgical and cold, with each blemish painfully visible. The fact is that many agree that the 2K resolution of digital cinema cameras better emulate the filmic look of celluloid 35 mm film than cameras using 4K.


This quote was pasted from Kurt Lancaster’s holy bok Cinema Raw: Shooting and Color Grading with the Ikonoskop, Digital Bolex and Blackmagic Cinema Cameras (2014, Focal Press, New York), which has become something of a catechism in my filmmaking book collection that has inspired me tremedously, not the least from a spiritual level. What Lancaster and his colleague friend Michael Plescia is referring to boils down to the fact that each frame of a Celluloid Emulsion Film is unique, producing 24 different versions of that same image each second, whereas Digital Film reproduces that same image each frame (if we disregard taking motion in the image into account) where 4K or 6K resolution removes the classic filmic softness from the image which can be somewhat emulated in 2K RAW; the CinemaDNG format creates very sharp images in itself so to limit the native resolution of the camera to 1080p or 2K (or 2.5K even) softens that image somewhat, a fact which makes the Ikonoskop A-Cam dII, the Digital Bolex D16 and the Blackmagic Pocket Cinema Camera such a unique set of cameras in producing a true filmic look that can fool the best of eyes in believing they are actually watching classical Film.

Even the iconic Arri Alexa, introduced in 2010 and still being considered to be the uncontested holy grail and true reference of Digital Super 35 Film, in virtue of its unique colour science (which the 2.5K Blackmagic Cinema Camera and the 1080p Blackmagic Pocket Cinema Camera are considered to emulate rather faithfully), uses a 3.2K CMOS sensor and 2.8K uncompressed ARRIRAW workflow (similar in resolution to the BMCC, which is said to be similar to CinemaDNG RAW). Cameras shooting in 4K and above that show more detail and razor sharpness also tend to become more sensitive to moire (a video feature above all) compared to the softer filmic look of 1080p / 2K cameras (although all digital film cameras suffer from moire more or less). Another great benefit of shooting 2K (2.5K or 2.8K) over 4K (or 6K) is that RAW often requires compression in 4K, and even more so in 6K, to be able to handle and store all the wast amount of data that is being created whilst taking a shot, whereas RAW lacks compression or is losslessly compressed (i.e. zipped and unzipped into a mathematically perfect copy of the original) in 2K, 2.5K and 2.8K; preventing degradation and distortion to the image. There are of course also DSLR cameras, which are cheaper than dedicated cinema cameras, with 4K resolutions, but these compress it down to 8-bit (and in rare cases 10-bit) H.264 video which isn’t even considered as a serious alternative.

We must also consider the fact that larger photosites or greater sized pixels allow for the collection of more photons (i.e. light) into their “buckets”, resulting in the ability to record a greater dynamic range, and higher contrast ratio, as well as more detail. Smaller sized photosites or pixels are more limited in comparison to collect light, yielding less dynamic range, lower contrast, and less luminance detail. The ability of a photosite to collect light is determined by both the width and depth of its “bucket”; although a deep and narrow bucket could collect as much light as a wide and shallow one, the optimal solution is both deep and wide photosite “buckets”. Thus, lower pixel resolution indicate the potential of larger photosites, as a sensor with lower resolution will allow for each photosite to be larger compared to a similar sized sensor with a higher resolution. So, in theory at least, a 4K Super 16 size sensor wouldn’t be able to capture as much dynamic range as a 1080p or 2K Super 16 sensor, if both sensors shared equal depth to their photosite “buckets”. So, what we see here is a trade-off between dynamic range and resolution. Every cinematographer prefers dynamic range over resolution, as he prefers to control his exposure, to have a higher latitude for over- or underexposure over sharpness or detail in the image. So the optimal resolution of a Super 16 sized sensor is 1080p or 2K, and under 3K.

Thus, the Ikonoskop A-Cam dII, the Digital Bolex D16, and the Blackmagic Cinema Cameras (BMCC and BMPCC), as well as the Arri Alexa, embody the term Digital Film better than any other camera currently existing, including the RED One or Epic cameras. So, in my radical opinion, Digital Cinema is not necessarily the same as Digital Film (if RAW is shot on ≥ 4K and is compressed using a codex such as REDCODE), whereas Digital Film always is Digital Cinema (if RAW is shot on ≤ 3K and is either uncompressed or losslessly compressed using file formats such as CinemaDNG). As the Digital Bolex (with a introductory price of $3299) was discontinued in June 2016 the only Digital Film camera being produced today, which continues with the Digital Super 16 size format that shoots in RAW, is the cheapiest of the three, the Blackmagic Pocket Cinema Camera, and since 2015 also its kid brother, the Blackmagic Micro Cinema Camera (BMMCC) for the same sensational price of $995. This is truly a revolution for the independent guerilla filmmaker which cannot be overestimated! This is also a fact that I want to capture and exploit to the fullest potential on this blog.