C300: BMW M5 refuel mid-drift

[Run&Gun]

Right- look at my unedited post, it says Amira

Under good lighting the C300 II and Alexa/Amira/Mini etc. can look pretty similar. It's under real-world and mixed lighting where ARRI has the edge- looks better with less work. Since ARRI is a much smaller company than Canon, Sony, and Panasonic, they use large, power hungry components such as FPGAs instead of VLSI ASICs (expensive to set up, but cheaper at scale). In 2018 it's technically possible to have a cellphone-sized device to produce industry-leading performance and looks. However, they wont give that to us, for a variety of reasons one can reason out when looking at the big picture. Consider what you could do with an iPhone X wired to a Super35 or full frame sensor, today? As computational power increases, so will computational photography, and sensors and lenses will able to be much smaller, whereby shallow DOF can be simulated realistically from captured depth data. iPhone X has a low-res depth sensor which allows for real-time background subtraction without a green screen, as well as 'entertainment level' shallow DOF simulation. This will improve over time, and someday we'll be able to simulate Cooke, Zeiss, Leica, Canon, Nikon lenses, diffusion filters, etc., in the same way we can now simulate various amplifiers for guitars.

Indeed a heavy camera has 'built-in' inertial dampening, though wouldn't you prefer to have a light camera with (really good) in-body stabilization or a small, light gimbal? In 2018, if one wants the best quality, it's ARRI, Red, Sony, Canon C, Panasonic Varicam etc., all relatively big, heavy cameras. In the future we'll look back and have a good laugh at how big and clunky cameras and lenses used to be

I think a lot of people, in the race to having everything smaller, forget that if a human is going to use it/interact with it, it still needs to be of a certain size to allow that, especially in this context of talking about cameras and lenses. Heck, one only has to look back a few years to the tiny cell phones that were almost un-useable, because they were SO small.

And as far as (electronic)gimbals go, they have there uses, but good Steadicam looks better. Even good handheld, depending on the situation. Electronic gimbals still look too mechanical most of the time and are usually a step behind the action.

 

[Eric Coughlin]

AF is nice, but I'm typically using manual focus cine zooms on the Amira, so AF wouldn't do much good.

Seems like quite a lot of this is not sharp.

AF is great a keeping a car in the middle of frame sharp (drama less so of course)

The majority of it was shot with GoPros and drones. Primarily the telephoto shots, often in slow motion, were with the Amira. I see perhaps a couple Amira shots with focus issues. If the cars had enough contrast (the dark blue would make it more difficult) I could certainly see auto-focus being useful.

I did a shoot recently of an NHL game tracking hockey players with the Amira, and since they're moving all over the frame, left, to right, down the rink, up the rink, tracking different players, etc., I don't feel auto-focus would have been as useful for that scenario. In this shoot the car was going to different sides of the camera frame depending on what part of the circle it was in, which would also mess with auto-focus since auto relies on a fixed point to focus on (unless you're using the "tracking" feature, but that feature has always been spotty for me, particularly with fast moving subjects).

I can certainly see auto-focus having helped. C300's image quality is not as good, we wouldn't have been able to do 200 fps, and 120 fps on the C300 is rather poor. Also, frame rates past 60 on the C300 loose auto-focus anyway. It's all trade-offs. I look forward to a not C700 camera that is high end implementing good auto-focus, as well as high end lenses that have it (the 17-120 has been said to work with auto-focus).

 

[fhammond]

I think a lot of people, in the race to having everything smaller, forget that if a human is going to use it/interact with it, it still needs to be of a certain size to allow that, especially in this context of talking about cameras and lenses. Heck, one only has to look back a few years to the tiny cell phones that were almost un-useable, because they were SO small.

+1

One the features I like about the Amira is that it's got physical buttons for most functions I care about and enough user assignable buttons to customize the setup the way I'd like (I assume). Try changing something like frame rate on the Amira and compare the process to how long it takes to do the same thing on most other cameras. Small size is only beneficial up to a point and then it becomes a downside.

Also, I don't believe the original statement ("In 2018 it's technically possible to have a cellphone-sized device to produce industry-leading performance and looks") to be true, unless you think that "industry-leading performance" can come from a small sensor.

 

[jcs]

All fun to speculate about a future where we don't need expensive gear and even a cam op's talent may become less important, but for now that's not the case. I suppose worst case scenario we become more like painters, where the paintbrush may be cheap, and thus rely solely on our talent to justify our rates to make a decent living. Gear is an additional way to earn revenue by charging for the rentals, but plenty of DPs earn their living from labor rate alone. The cost of cameras will continue to go down, but I don't foresee the cost of high end support (lighting, grip, audio, stabilization, etc.) going down to dirt cheap consumer prices within our career span.

In-camera or lens IS has a different and less organic look than handheld with a heavy camera. IS tends to have a floaty bouncy feel to it, which typically isn't desirable and is good primarily for when whatever mode of stabilization in use is not sufficient (such as hand-holding a DSLR with no rig). I did a mockumentary style web-series with C300s with full shoulder rigs using L-series lenses (such as the 24-105 and 70-200 which both have IS), and if I recall rightly we left IS off most of the time to keep things looking more organic.

It's not speculating about cost: high-end bicycles are far more expensive as the weight goes down! So why not high end cameras that are super light, super functional, with the best ergonomics, best stabilization, best AF, best color, DR, intelligent menu systems and physical buttons, etc. Agree current lens IS has issues. However gimbal systems can literally be programmed to simulate whatever you want! Like the fluid motion of a steadicam? It can be simulated with a gimbal! Gimbals have a long way to go: 3-axis (yaw, roll, pitch) is a good start, now we need x,y,z linear for full 6-axis stabilization. The stabilized camera can still sit on the shoulder if desired, etc. And any mass / inertial (inertia tensor) distribution can be simulated (speaking from experience writing flight and driving simulators).

I rigged up the C300 II with a shoulder mount, 177Wh battery, etc., 24-70 F2.8L II lens and compared to footage I shot with a simple lightweight belt-rod system, and the light system was more stable. If the goal is smooth-ish shakey cam, sure heavy shoulder works great. I've seen the future, and it's light weight computer controlled systems . In the meantime, heavy camera gear is a great way to help stay in shape haha!

 

[jcs]

+1

One the features I like about the Amira is that it's got physical buttons for most functions I care about and enough user assignable buttons to customize the setup the way I'd like (I assume). Try changing something like frame rate on the Amira and compare the process to how long it takes to do the same thing on most other cameras. Small size is only beneficial up to a point and then it becomes a downside.

Also, I don't believe the original statement ("In 2018 it's technically possible to have a cellphone-sized device to produce industry-leading performance and looks") to be true, unless you think that "industry-leading performance" can come from a small sensor.

Agree with point on human factors when engineering products: seems like many manufacturers aren't aware of cognitive engineering. We're not talking about Zoolander cellphone-sized devices, instead cognitive-engineered products that work with humans- much smaller than current, but not too small. The point is that computational power has advanced so far that there's really no excuse for ginormous cameras anymore. A single SoC (system on a chip) can compute 4+ trillion ops per second, and with low power! https://www.anandtech.com/show/11771/intel-announces-movidius-myriad-x-vpu . Far more than enough to do the basic stuff current cameras are computing. These chips can do far more powerful things, such as AI - machine vision (already being used in drones).

Regarding cellphone-sized devices, note I stated wired to a Super35 or full frame sensor. If the sensor, electronics and power systems can be tiny, we can build a camera size based on human factors, not hardware size limitations. The only reason I said wired to Super35 or full frame sensor is to utilize current lenses. Once computing power gets on a bit further, tiny image sensors and lenses will blow away what was possible with current tech. How big is the human eye? How about an eagle or owl's eye? Biological cameras still smoke our current camera tech, and are much smaller. Someday tech will surpass even those systems. Open your mind. Think!

 

[Eric Coughlin]

It's not speculating about cost: high-end bicycles are far more expensive as the weight goes down! So why not high end cameras that are super light, super functional, with the best ergonomics, best stabilization, best AF, best color, DR, intelligent menu systems and physical buttons, etc. Agree current lens IS has issues. However gimbal systems can literally be programmed to simulate whatever you want! Like the fluid motion of a steadicam? It can be simulated with a gimbal! Gimbals have a long way to go: 3-axis (yaw, roll, pitch) is a good start, now we need x,y,z linear for full 6-axis stabilization. The stabilized camera can still sit on the shoulder if desired, etc. And any mass / inertial (inertia tensor) distribution can be simulated (speaking from experience writing flight and driving simulators).

I rigged up the C300 II with a shoulder mount, 177Wh battery, etc., 24-70 F2.8L II lens and compared to footage I shot with a simple lightweight belt-rod system, and the light system was more stable. If the goal is smooth-ish shakey cam, sure heavy shoulder works great. I've seen the future, and it's light weight computer controlled systems . In the meantime, heavy camera gear is a great way to help stay in shape haha!

Sort of like auto-focus, until gimbals can be wired to your brain (such as your human hands are) to actually deliver in real time the exact movement you want, I don't foresee them completely replacing more traditional means of stabilization which are directly controlled by your hands such as handheld, tripod, dolly, and Steadicam.

The high end already has access to computer controlled systems but more often chooses traditional means of stabilization because it is more befitting to the look they want, not just because it's simpler or cheaper to use. I have a Rhino arc slider and on a shoot recently I asked the producer if he'd prefer that, or to manually do a dolly move during the interviews, and he choose the latter so there could be more variety and organic feel to it.

I'm not sure heavy cameras are good for staying in shape. Too many stories of bad backs in the industry of cameramen. Really have to be careful about leaving a 30 lb camera on your shoulder all day on a regular basis. I'll probably buy an Easy Rig before long, though most of my shoots don't require long periods of shoulder mount. I've heard people say their camera job is their workout since they lift heavy equipment, but in my case I put on 20 lbs while living by that philosophy, and had to hit the gym to later get rid of that 20 lbs. Meaning, I don't feel our job is actually a good workout, even though we lift heavy things and walk (not run) a lot.

Yeah, ultra stability often isn't the desire, of which IS or other solutions can get things more stable. But if you really needed it that stable, put it on a tripod. The whole point of handheld as far as a look (not regarding convenience) is to not look completely stable. Sometimes something more in-between handheld and tripod, such as using IS, can be a desired look; it just depends. All I'm saying is that getting it more stable isn't necessarily a good thing.

 

[fhammond]

Open your mind. Think!

I am but since I'm not building cameras myself, that results in thinking about cameras are available - today - to purchase and to be supported by a company I trust. I'd agree that the Amira isn't cutting-edge but it clearly has a lot of benefits that aren't readily found (or found at all - usability being the main one) in other cameras.

 

[jcs]

I am but since I'm not building cameras myself, that results in thinking about cameras are available - today - to purchase and to be supported by a company I trust. I'd agree that the Amira isn't cutting-edge but it clearly has a lot of benefits that aren't readily found (or found at all - usability being the main one) in other cameras.

All Alexa's are great tools! Imagine an Alexa created with the latest in compute technology- would be amazing. The Alexa Mini is a step in the right direction: http://www.arri.com/camera/alexa_mini/camera_details/alexa-mini/subsection/technical_data/ . Wonder if they'll create an Alexa Mini LF (FF sensor)?

@Eric- was joking about the camera workout, and concur heavy cameras used frequently can result in long-term issues. Maintaining a healthy weight is easy and hard. Easy if one can 100% avoid processed foods, sugar, and alcohol (which typically is loaded with sugar) and exercise daily. It's hard when un-processed foods aren't readily available or one doesn't have the willpower to fight the processed food addiction and/or daily exercise doesn't happen.

Regarding stability, sure, it depends on the desired shot. In my 'heavy-cam' test case, I didn't get the look I wanted, and got the look I wanted with a rod and belt stabilizer, which has the added benefit of being comparatively lightweight.

 

[Parrot]

Wonder if they'll create an Alexa Mini LF (FF sensor)?

Arri will ship an Alexa Mini LF when an FPGA that's power efficient enough for the small form factor becomes available. Are there enough Geeks on this board to figure out which FPGA Arri uses in the LF and then look at release schedules for new FPGAs and then extrapolate when Arri might ship Alexa LF Mini? I'd like to know, but I'm not ready to geek that far out right now...

 

[jcs]

Arri will ship an Alexa Mini LF when an FPGA that's power efficient enough for the small form factor becomes available. Are there enough Geeks on this board to figure out which FPGA Arri uses in the LF and then look at release schedules for new FPGAs and then extrapolate when Arri might ship Alexa LF Mini? I'd like to know, but I'm not ready to geek that far out right now...

Here's Xylinx's most powerful + least power used FPGA: https://www.xilinx.com/products/silicon-devices/fpga/kintex-ultrascale-plus.html#productAdvantages . Also based on 16nm FinFET tech: https://www.anandtech.com/show/11771/intel-announces-movidius-myriad-x-vpu which is about 1 TerraOP per Watt (which is amazing). Didn't quickly find power usage for the Kintex, however I suspect it's a lot more power hungry. These chips can compute machine vision tasks, which takes a lot more computational power vs. a camera's color science, deBayering, image processing, and compression (especially if based on FPGAs developed 8 years ago, which is an eternity in tech). All of ARRI's (and probably every other camera maker) processing should run well under 1 TerraOP/s (haven't seen their code, obviously, so who knows what they are doing). However I have worked over 20 years in image processing, compression, and artificial intelligence / machine vision, and can't imagine a camera needing anywhere near 1 TerraOP/s. The Intel/Myriad-X likely costs around $10 (about the same price as the Myriad 2 it replaces). ARRI (or any camera maker) can charge whatever they want, the value is the combined hardware/software (firmware), and product design. There's absolutely no reason to have to suffer through ginormous cameras beyond 2018 due to hardware/power issues. We're still a few years away from decent computational cameras, where very small sensors and lenses can be used, and sensor size and lens behavior can be accurately simulated. So for the near future, most of the size/mass will be related to the sensor, lens mount, and lenses.

When are we going to see the CMOS sensor updated? https://fstoppers.com/gear/inventor...nces-breakthrough-low-light-photography-90305

Amazing things are coming...

[EDIT]
Now this makes sense, hopefully ARRI & Canon are going this route to stay competitive too: https://www.cinema5d.com/red-is-talking-with-foxconn-are-we-ready-for-affordable-8k-cinema-cameras/
Once component size is reduced, they can also make cameras that are not clunky jumbled and boxy, instead curvy, rounded, and sexy! I added sexy as some cameras have curves, but they are not sexy nor ergonomic
[/EDIT]