Last week we went into detail why the Alexa 35 is ARRI’s first major leap in image quality since 2010. But the irony is that no one particularly complained about the image quality of Alexa, but there were a lot of other complaints about it. In this article, we’ll take a look at how ARRI responded to them with the release of the Alexa 35, how much the quality upgrade costs us, and try to assess how the new camera meets the needs of modern filmmaking.
Since the dawn of the digital age, ARRI has preached that pixel quality (and size) is more important than quantity. Indeed, the relationship between the number of pixels in a source image and the perceived resolution is subject to a severe diminishing returns effect. If you don’t believe me, then believe me Steve Yedlinand his legendary experiment. Pixels and “K” numbers have long been ammunition in the marketing wars of camera manufacturers, but that doesn’t mean high resolutions are useless. On the contrary, due to the low resolution, Alexa has traditionally lost out to competitors in projects with a high VFX component. The higher resolution, coupled with color processing optimizations, should make the Alexa 35 an attractive choice for VFX-heavy projects.
The Alexa 35 is also designed in part to meet the demands of streaming platforms. Infamously, Alexa with its 3.2K sensor has not been officially approved by Netflix. The Super35 crop mode on the Alexa LF also doesn’t capture enough pixels to meet the 4K mandate. In fact, ARRI didn’t have a Super35 camera to film for Netflix. Therefore, ARRI had to increase the resolution of Alexa 35 in order not to lose the clientele.
Streaming platforms require 4K as the minimum capture resolution. The Alexa 35 was originally conceived as a native 4K camera to meet the demands of streamers in all major sensor modes, both spherical and anamorphic (except HD modes, of course). And all this with the minimum possible increase in the number of pixels in order to preserve their size and light sensitivity to the maximum. As a result, we have a sensor with a resolution of 4.6K pixels with a size of 6.075 microns.
Post stabilization, framing and supersampling. ARRI encourages everyone who can afford it to shoot in 4.6K in open gate and cut 4K of any desired format from there with post-stabilization and a little cropping (cutting groats from the common fund, as RED shows in their demos, is a shame and savagery, in my humble opinion). But even without an open gate, the 4.6K 16:9 mode reads the full width of the sensor with a margin for bleed. As for supersampling, it is available in-camera when recording in ProRes, along with anamorphic stretching.
High resolutions are needed for VFX work. The increased resolution of Alexa 35 already gives sharper edges in sharp color transitions, which are so important for VFX. However, image optimization for keying and compositing is also available at the level of the new ADA-7 debayerization algorithm, as well as at the level of new “textures” of sensor settings for higher sharpness and microcontrast. VFXers should be happy.
The green screen has been replaced by virtual production, aka Mixed Reality, aka LED Volume. For ARRI, it was critical that the new camera meet the needs of the new shooting format. The Ethernet port on Alexa 35 is now capable of streaming real-time metadata directly into Unreal Engine via the ARRI plugin. Previously, this required an additional UMC-4 device and a bunch of different wires.
Is the Super35 format relevant in 2022?
For me, a much more interesting and more difficult issue is not resolution, but sensor size. Judging by the popularity of Alexa Mini LF, the demand for large format is huge. With the competing Sony Venice and RED Monstro / V Raptor, cinema seemed to have finally moved to the LF/VV format. Fortunately, it does not cancel the ability to shoot in Super35. Why isn’t ARRI’s new LF camera then? Won’t the Alexa 35 be seen as a step back, despite the impressive gains in image quality?
The optics argument doesn’t work because there is a crop option on LF sensors, which provides backwards compatibility with Super35 lenses. And in the case of 8K / 6K sensors, it also covers compliance with the 4K mandate of streaming platforms in crop mode. If we are talking about optics, then the LF format just wins, because it gives more choice – the entire S35 fleet is available, and the latest Full Frame cinema optics, like ARRI Signature, and countless photo lenses, both modern and vintage (Soviet hit Helios 44-2 regularly shoots movies) that can only reveal their character in full frame. And the crop mode also allows you to get additional range from full-frame lenses if necessary. In this sense, it is rather strange why the Alexa 35 comes with an LPL mount, which is designed for LF optics. Sets of native LPL lenses can be counted on the fingers of one hand.
Another argument in favor of LF is the quality and character of the image. You can’t argue against physics: other things being equal (sic!), a larger sensor collects more light. This means higher exposure and less depth of field. Yes, up to a certain point you can compare LF and Super35, but it breaks at wide angle. How to realistically compare 20mm/T1.5 on Super35? Personally, I don’t know 14mm/T1 lenses. Many cameramen want to shoot closer and wider, while being able to blur the background as well as they can. This “intimate” aesthetic is extremely difficult to create on the Super35. The point is not only in the viewing angle, but also in the nature of the curvature of perspective. LF makes close-ups acceptable at 35mm, while its counterpart at Super35 (24mm) will not be very flattering to the actor, and the farther into the wide angle, the more noticeable distortion.
There is also the question of generations. For the old guard that started out on film, Super35 is a native format to which LF is an extension, a new aesthetic suited to certain projects and situations. For the younger generation that started on the Canon 5D, Full Frame is the norm, it’s a native format, and Super35 is either a forced step back or vintage. Generations change slowly in the cameraman’s trade (Sir Roger Dickens is doing a great job at 73), but the future will inevitably be for those who have never seen the film. By insisting on Super35, ARRI may risk losing this generation to the competition.
Given the considerations above, I tend to answer “no”. The Super35 format is more relevant as a format of work and as a certain aesthetic of the frame, but not as a sensor size, given that LF does not require sacrificing Super35 capabilities. Of course, Super35 is not the end of the world (a lot of masterpieces were filmed in Super16 too). I am sure that the vast majority of operators will choose image quality if they have to choose between it and sensor size. But it’s a shame that you have to make such a choice at all when it comes to ARRI cameras. The trend towards increasing the format looks inevitable and LF is not the limit, IMAX and Alexa 65 still grow and grow.
Should I wait for the LF version of Alexa 35?
The sad news is that the option to “stitch” multiple sensors, as was the case with the ALEV-III in the Alexa 65, Alexa LF cameras, is missing in the new generation. According to an ARRI spokesperson, this is technically impossible with the new sensor architecture. I assume that the matter is in the stacked architecture, in which the ADC and memory are located on the back of the sensor board. Therefore, if there is a LF-version of ALEV-IV, then it will be a new sensor and a new camera, which means that they will have to wait a long time. I assume that if ARRI could technically make the new ALEV-IV 36x24mm right away, they would, but it turned out to be very powerful in terms of power and data speeds anyway. ARRI is currently positioning the Alexa (Mini) LF as a large format solution for the coming years. Especially considering that RAW from it can be processed through the new Reveal Color Science. But then again, the Super35 options on this camera don’t meet the 4K mandate, and it doesn’t have 17 stops of dynamic range either. And for those who feel cramped even in LF, there is also the Alexa 65, available to a select few exclusively through ARRI Rental.
4K Super35 camera in the world 8K LF
Given that competitors’ flagship cinema cameras are full-frame, the question is, is the Alexa 35 competitive? Based on the quality of the image alone, the components of which we analyzed in detail in the last article, definitely – “yes”. Personally, I can’t wait to see what the Alexa 35 is capable of in the hands of veteran operators, especially on an HDR display that can unleash its potential. In addition, ARRI addressed the shortcomings of previous cameras in terms of VFX projects and made the Alexa 35 as attractive as possible for work both on the virtual set and in post.
The main question is how important full-frame aesthetics are for modern cinematic language. Will filmmakers be willing to forego the high dynamic range and rich palette of the Alexa 35 for the more naturalistic and intimate sense of space that large format affords? Will this encourage more operators to try the Sony Venice II or the RED V-Raptor? What about projects where formats are mixed? Wouldn’t it be more convenient to use one camera with crop mode than two with different sensors? The answers to these questions will show only practice in the coming years. It will be interesting to watch the trends, but something tells me that those who shoot with Alexa will continue to shoot with Alexa in the future. Don’t underestimate the element of habit, reliability, and predictability in filmmaking.
PS: guys, throw a couple of tens of thousands of euros for donations, otherwise there is not enough to pre-order Alexa 35 🤣
So, we talked about the picture and about the sensor, but this is only half the battle. The camera can shoot as beautifully as you like, but if it is inconvenient to use, if it cannot be relied upon, if it slows down the process, then this is a bad tool. Therefore, next week we will discuss the physical design of the camera, the features of working with it, and most importantly, we will try to understand for whom it was made.
Thank you for reading the SHAGRAL Video Production Blog! My name is Grigory Shakhanov, and every Friday I post here my observations on industry trends and share the secrets of commercial video production. Subscribe to receive new posts!