The next generation features of cameras

[Christoffer Glans]

I've been advocating for a long time that we should look past the current spec requests that keep driving the industry of cinema cameras. I have a hard time imagining DSMC3 with just a bump in Ks and DR, less noise and just a phase-detection AF. We are starting to enter the same kind of stagnation that phones have entered, in which people stop updating every year because there's no point to it. There's no real innovation and people don't need more when there's nothing more to get or need.

We need some change in the philosophy of cinema cameras. Since so much today is driven by post-manipulation and VFX enhancements, even for projects that don't have a focus on VFX specific shots, we need some innovation in how we view what our cameras can do. I would trade away a lot of past years innovation in cinema cameras for the ability to have in-camera stabilization metadata that actually exists within the RAW file settings in post. The same goes for 3D position camera tracking data that gets rid of tedious camera tracking in post.

The next company that actually puts things like these into their next lineup of high-end cinema cameras will win first prize as the top brand for cinema cameras. And I'll bet that the next Emmy for technical achievement would be up for grabs for any company that actually puts this into their cinema camera. I don't know why, but when my freakin iPhone can do all of these things for a fraction of the cost of a cinema camera, why is it so impossible to implement all of it into a cinema camera?

There are plenty who return counter-arguments, but all I hear is excuses that really don't add up. The tech is there, it's not rocket science worth billions of dollars and it's not really that hard to implement when a startup company like SteadXP solves the stabilization problem for a fraction of the cost many propose it would cost. Even with the cameras right now, we could have a stabilization slider in RedCine that utilizes the gyro metadata, but we don't, because of "whatever". I'm amazed that we don't see more requests for things like this. It's like people are just used to doing it the old ways, throwing thousands of dollars on unnecessary post fixes for stuff we could have as automatic features in-camera. And I'm not talking about doing it with boxes and wires and stuff strapped onto the camera. Get it into it, in-camera, as powerful metadata with the necessary tools for it in post.

The youtube channel Corridor Crew gave a pretty good rundown of the same types of features that I really want to see in future cameras going forward. These are features that I think people don't seem to understand the benefit of until they actually witness them, exclaiming: "why didn't we have this sooner?"

 

[Steve Sherrick]

I agree with you that this tech is coming and will be the next big jump. Usually, these kinds of innovations happen when the industry plateaus and sales of cameras slow or there is an immediate demand for something, which sometimes happens when big projects need a certain technology to exist to achieve their goals and so they invent it or push others to do so. I wouldn't be surprised if there has been a push for some of the features you mentioned, it just depends on how hard that push is and how much investment is needed to accomplish it. As we know, if it were easy, there would be 20, 30, 40 camera manufacturers making these high end cameras, but there's not and that's because of the investment needed to not only build the products but to build a following and break through, which is of course what RED was able to do.

I think on the surface, this seems like something RED would be primed to do since their cameras are used a lot on visual effects films and they have a lot of experience now when it comes to pushing boundaries and developing a following. I simply just don't know the economics of their company to know where they currently stand in terms of investment into R&D. Can't imagine that Jarred will stand still. I would expect they are probably already looking into these things to see how they can take advantage of current tech and what is needed to develop things that aren't quite there yet.

Perhaps in a few years, we'll have a digital cinema camera that will make the current crop of cameras look antiquated. It will be fun to watch the evolution/revolution.

 

[Hrvoje Simic]

That video with cinema camera experts' "this thing should have what I want because it's 2020 and it's expensive" train of thought is awesome.

I'm still waiting for a hoverboard.
And it's 2020. And still nothing. They shoulda coulda made it by now. I mean really. Lame.

 

[Phil Holland]

"why didn't we have this sooner?"

I like those guys as a former and moderately current VFX dude, so I'll keep some of my thoughts close to the chest as this industry has some interesting tides in the deployment of emerging technologies.

You are largely looking at very different ecosystems for larger cameras than smaller ones and moderately disposable ones.

When you are looking at a premium digital cinema camera, you are "ideally" looking for state of the art at the time of invention that will be a solid invention for some time to come. I the good news the pure emphasis is on image quality and general workflow. Bare in mind it takes a year or two and even longer in some cases to develop hardware. Important thing to remember.

With these smaller cameras often getting replaced on a yearly basis or even quicker they have the power to explore new chips, hardware, and even develop software surrounding them. Those are cameras many are buying on a yearly basis, which ideally you aren't doing for $15K and up premium cameras or we'd all go broke real fast. The market actually doesn't support that concept on premium cameras particularly well outside of super premium, like custom bodies RED makes for premiere filmmakers.

It's interesting that gyro data is mentioned because RED was way, way, way early to that party. But the tech could be improved greatly and most humans don't know at all how to tap into it. Interestingly that does empower independent filmmakers a great deal and is a great assist to VFX work and fixes. However, I've gone years now with out post stabilizing a shot on features through commercials. Literally not needed. But when you need it you need it.

4K SDI came out just after DSMC2 was actually finalized and launched and dammit I wish we had it around 2017, but DSMC2 and RED in general seems to be on a 5 year main body product cycle. However, eyeballing Komodo we are seeing RED flex a bit back in their innovative nature, which btw they have been greatly criticized for. We will have a 2lb camera with 6K, 4K out, internal compressed raw, PDAF autofocus, Global Shutter, and that's just what we know now.

That's insane. And thinking about it broader than that, I though about how crazy an actual Global Shutter PDAF sensor tech was the other day. I mean it's not something everybody is doing at all and certainly new for the cinema space in a camera on the fringe pricing levels of consumer/prosumer land.


TL;DR - I'm all for all of the new technologies ended up in our gear. Sometimes it takes time, moolah, and development for stuff to get refined even to the point that it's that usable in a system. Like you don't want to integrate something that is sluggish or may not become industry standard. However, some of this stuff is useful for sure. And yeah, audio needs to be looked at harder than ever 2020 and onward. Huge improvements can be made.

That said, I'm also really looking forward to some of the simple quality of life improvements we'll be getting. Better power consumption, more powerful cameras, speedier menus, faster startup times, etc.

It's awesome to have all the bells and whistles, but refinement truly can have a massive impact on every day usage.

Things are bloody fast paced in the general camera technology industry right now. 8K cell phone lands next month. 8K just a few years ago was a alien concept to many.

Profit breads innovation as well as technological advances over time, then it's just up to real smart minds doing some really hard work, some of which is vastly more difficult than many think. Add into that expanding on the ecosystem to support and develop for certain things and it gets nuts fast. A bit easier for multi-billion dollar companies for sure.


Heh. Hoverboards. That's so 2000-2010s.

Which means we're really on pace for dinosaurs soon. Living ones that is. Then again, we'll be in a "good idea or bad idea?" situation real, real quick. I mean, some people die lacing their shoes. We solved that problem long ago and really even before laces, but you know, fashion is fashionable.

 

[andrewhake]

"The tech is there" is an absolutely ridiculous statement to make. It is, and never will be, that simple.

Just because you know of something that is possible does not make it a viable option for actual use in another product. It's endlessly more complicated than that.

If only making products and tools was as simple as knowing that one things exists, and that another things exists, and just "thinking" then into existence..

 

[Hrvoje Simic]

Heh. Hoverboards. That's so 2000-2010s.

I like it how the arrow points to the device.
Otherwise I'd be confused where to look at for the said device.

I didn't say "desperate flying toolshed contraption" but a hoverboard.
Marty McFly had it so it should be made by now.
Because a) I want it and b) it's 2020 and c) DJI and Maglev

Ya just combine those two in some way so the board fits in my backpack, looks cool and goes anywhere. Boards exist and flying does. The tech is there. So just make it happen already. I mean really.
/rolls eyes/

While we're at it, I want smartphone control and autopilot with obstacle avoidance and all that new car/drone stuff, so I can input a route and chat and do selfies and take pics, and download some cool moves, and vlog in V-log while I hover, informing the planet about the achievements of my thought process.

 

[Zack Birlew]

I liked that video because it does touch on some pretty cool ideas and concepts that you do scratch your head about and wonder why more manufacturers haven't jumped on board some of these innovations. Really though, we saw the future with the Lytro digital cinema prototype but unless someone does something similar or gets that tech from Google's hands now that they own Lytro, we're stuck with standard imaging. Being able to break up a scene into composite pieces was brilliant to see and effectively eliminated the need for chromakey for only the most particular of effect shots and even then light field recording would help a great deal. I don't know if Google is still developing it or just sitting on it but if it could be licensed then that would be a new level of camera. Of course, barring that, having all the extra metadata and gyroscope mapping would be a useful addition to any cinema camera but as stated already, that stuff gets better and better all the time so it's better to have external devices or modular add-ons that can be replaced with newer versions as time goes on and they get better and better but without having to replace the camera. I would like to see RED advance further, how, I'm not sure, but I would caution that they've seen what going after pure resolution alone led to and how successful Gemini has been and how difficult shaking off Arri's color science still is. I'm sure they'll be able to get to 8K again but with Gemini's low light capabilities added in for revision and I'm sure they'll be able to jump to 12, 14, or even 16K in the not too distant future since they've reached 8K already.

I suppose a change of stance would be nice as the modular concept should never go away but crippling cameras with lower spec HDMI for instance or limiting media options shouldn't be the case anymore, even if it comes at the cost of features and yes, quality. Open the camera platform up, let people experiment, use the cameras however they may want to, even if it's just getting 4K/6K/8K Prores or even Prores RAW with their Atomos recorders and the like, at least to start and even as backup later down the line if that's the case. My RED One MX is so closed off and dependent on the availability of RED 1.8" SSDs and RED Compact Flash cards being available and operational that it's tough to want to use it more often than I do for fear of wearing down the drives let alone the SSD module itself. Yet I can just turn on my Sony A7S or Blackmagic Pocket 4K and record onto whatever's available basically, even my Atomos Shogun. Making the cameras easier to use is another matter as well, I know everybody is in love with LUTs, LOG modes, and using all the zebra functions they can but those just seem to counter the whole point of capturing images the way we see them or at least getting close enough to do that as has been done since the beginning of photography. If someone can't tell if they're clipping highlights too much or dropping down into the shadows more than they should with all the great monitoring tech we have now then alleviating that is not only up to a matter of constantly developing technique and gaining experience but is also a key point for innovation with digital cameras. If nothing else, we need to aim for HDR or full multi-exposure recording methods beyond basic RAW, or in this case, REDCODE, so that proper exposure is a thing of the past and this goes for any camera whether it's for stills or even a tiny Gopro. One other thing that bugged me about my RED One MX is why I couldn't change my mind about aspect ratio later on as I film in 4K 16:9 HD for most things but I have all this look around room according to the monitor with look around enabled, why not record the whole sensor image anyway in the background and at least give me that option later on in case we decide to go 2:35.1 or even just to have the extra room if the composition is off or if we needed more image area for an effect like a digital camera shake? That seems like a problem that shouldn't be there for such high end camera systems in the first place, especially since they're all very data intensive as it is. All in all, there's certainly a lot of things that could be done to make cameras better but ultimately it comes down to choice and being able to make the most of what your chosen camera can be and for manufacturers to push past that constant cycle of trickle down tech or stopgap limitations, it's time to innovate and move beyond what we've become used to, much like when RED burst the HD bubble and pushed the industry to 4K.

 

[Kwan Khan]

Thanks for sharing...

 

[Hrvoje Simic]

Making the cameras easier to use is another matter as well, I know everybody is in love with LUTs, LOG modes, and using all the zebra functions they can but those just seem to counter the whole point of capturing images the way we see them or at least getting close enough to do that as has been done since the beginning of photography. If someone can't tell if they're clipping highlights too much or dropping down into the shadows more than they should with all the great monitoring tech we have now then alleviating that is not only up to a matter of constantly developing technique and gaining experience but is also a key point for innovation with digital cameras.

No new innovation is necessary for proper exposure.
Basic camera skills and experience are.

No camera problem there.

If nothing else, we need to aim for HDR or full multi-exposure recording methods beyond basic RAW, or in this case, REDCODE, so that proper exposure is a thing of the past

You want a professional camera which removes the need for learning how to properly expose ?

Arri had parallel dual gain 10 years ago.

and this goes for any camera whether it's for stills or even a tiny Gopro. One other thing that bugged me about my RED One MX is why I couldn't change my mind about aspect ratio later on as I film in 4K 16:9 HD for most things but I have all this look around room according to the monitor with look around enabled, why not record the whole sensor image anyway in the background and at least give me that option later on in case we decide to go 2:35.1 or even just to have the extra room if the composition is off or if we needed more image area for an effect like a digital camera shake? That seems like a problem that shouldn't be there for such high end camera systems in the first place, especially since they're all very data intensive as it is.

There is no problem.

There are larger crops of the sensor, and OG options. You shoot wider then intended for delivery and crop.

All in all, there's certainly a lot of things that could be done to make cameras better but ultimately it comes down to choice and being able to make the most of what your chosen camera can be and for manufacturers to push past that constant cycle of trickle down tech or stopgap limitations, it's time to innovate and move beyond what we've become used to, much like when RED burst the HD bubble and pushed the industry to 4K.

Manufacturers innovate and push cameras already.
So no problem there neither.

 

[Hrvoje Simic]

The next company that actually puts things like these into their next lineup of high-end cinema cameras will win first prize as the top brand for cinema cameras.

https://www.arri.com/en/company/about-arri/awards


https://www.arri.com/en/company/pre...ri-wins-academy-award-for-alexa-camera-system

I don't know why, but when my freakin iPhone can do all of these things for a fraction of the cost of a cinema camera, why is it so impossible to implement all of it into a cinema camera?

Because there are factors like IP, R&D, licencing, time, facilities, OEMs, yields, testing, logistics, quality levels, certification, and a bunch of other stuff in the equation between imagination and tangible outcomes.

And makers of herd management devices have far greater resources and interest to give you what you want.

 

[Christoffer Glans]

I think on the surface, this seems like something RED would be primed to do since their cameras are used a lot on visual effects films and they have a lot of experience now when it comes to pushing boundaries and developing a following. I simply just don't know the economics of their company to know where they currently stand in terms of investment into R&D. Can't imagine that Jarred will stand still. I would expect they are probably already looking into these things to see how they can take advantage of current tech and what is needed to develop things that aren't quite there yet.

I can say that since Arri released their Mini LF, almost no one in Sweden rents a Red anymore. Sweden's industry has been pretty focused on Arri for the most part but with the Mini LF, it's a death blow to anything Red related. If this is true for more places around the globe then Red has to do something else than just Ks, DRs and so on. They need to focus on areas where Arri is lacking and will be lacking going forward. Red is often much further ahead than others, while Arri is behind but more reliable. The choice has to be one of those and if you are focusing on reliability you are competing directly with Arri on their turf, while competing in innovation you're on Red turf. If Red innovates with features like the ones mentioned, they would be the only ones having them for cinema cameras and will own the part of the global industry that needs such features. If anything, it would be a wise business move in order to compete against the Mini LF.

 

[David Rasberry]

As size, thermal and power envelopes drop with each new generation of chips and other electronic components it becomes easier to fit more powerful sophisticated tech into a given physical design package. Cost, vs development time, vs benefit trade offs for any particular set of design goals for a particular camera is an ever more complex engineering challenge.
Jarred has mentioned that as they turn on software features in Komodo, power consumption, currently somewhere in the 27 W range, continues to rise. What is the ceiling for that? 30W? 40W? How will that impact battery life? Thermal loads? Image noise? Long term reliability?
On the other hand there is much to be said for simplicity. I find shooting a camera whose outstanding feature is a total absence of features beyond the most basic necessities quite liberating.

 

[Bobby Farokhzad]

I am for the most part an owner operator and in some of the projects I film I am a one man band. Here is my two cents.
Don't need the built in camera stabilization, I use a gimbal with good to great results.
Don't need the autofocus but it would be nice to have.
3D tracking metadata is something many of us have been asking for, I would like to see it being implemented.
Wireless video is the one feature I would want the most in this list, I think it's a must have for any serious production.
Couldn't care less about wireless audio.
What I really really want to see in the future is a full RGB sensor with no debayering of any kind and with global shutter.
Just my humble opinion.

 

[Blair S. Paulsen]

Perhaps the most elegant solution would be a better selection of modules, targeted to particular use cases. Keeping the camera brain small is desirable so it can be rigged to suit myriad situations.

One of those module offerings could support one man band/docco style shooting situations implementing several of the capabilities noted in the video. A different one could be more DXL-ish with features aimed at fully crewed shoots. I can also see use cases for a Ranger style rig - imagine a DSMC3 brain docking into a Ranger body. RED could still market a "known" configuration to suit the rental house market, with the advantage of potentially being able to swap out brains down the road as the tech evolves.

IMHO, the modular approach of DSMC2 is a great concept that has not been well exploited to date. OMODs got some things right, others not so much. As noted by others, there's a lot steps on the development road. I know Mark has worked very hard to push forward through all the politics, technical issues, etc - but even with his knowledge and connections it still takes a lot of lead time. If DSMC3 retains the modular concept, I hope that RED is able to engage with 3rd party module makers far enough advance to support development of modules that offer desired capabilities.

FWIW, I also see the potential for a smart phone based approach. Very interested in seeing how the upcoming Komodos integrate with phones, could be a preview of coming attractions...

Cheers - #19

 

[Björn Benckert]

What I cant understand is. Why on earth did the steadyXP build just a little plastic shitty GoPro battery powered unit?

To me if they instead of the plastic thing would have made a Vlock 2 Vlock milled aluminium cake to sit in between camera and battery worked reliably and looked somewhat professional they could have sold those for + 2000USD a pop to steady cam ops etc.

And yes they try to lock in their features into their own app instead of pushing out raw data or something that could be applied in any post app.


And yes its a shame that red has such poor gyros in the DSMC2, or if its just the readout that is poor. I spent quite a bit of time and money investigating the RED SDK to extract gyro data, its there but its possibly the worst gyro data I ever seen. Its as bad and lagging as the leveling tool on the touch screen which I find very odd, every smartphone made has a really good gyro inside, one that can be calibrated and used as a quite accurate leveling tool with very little or close to no lag.

I also looked into manufacturing a gyro module like the steady XP stuff described above, even got permission from Autodesk to use their universal FBX format for ti. But as camera developments goes so quick these days I simply did not pull the trigger as its not really complex stuff and as soon as one camera manufacturer does this, the rest will need to follow and in such scenario being a third party provider of such system would not fly well.

This is how far we got with the DSMC2 gyro. Dont buy it as there is no real good use for it as the data coming out of the camera is simply very shitty. And the mac version does not kick on later iterations of OSX than I think, high sierra last Open GL OSX. on PC it should fly properly.

http://www.syndicate.se/axis/

But the program will work with komodo files and with a bit of luck Komodo has a better gyro than DSMC1 and 2.

 

[Tom Gleeson]

The next company that actually puts things like these into their next lineup of high-end cinema cameras will win first prize as the top brand for cinema cameras. And I'll bet that the next Emmy for technical achievement would be up for grabs for any company that actually puts this into their cinema camera.

Christopher,

I believe manufacturers have already started doing it. Check out the Sony FX9. It has an effective, configurable and practical auto focus system. Not perfect AF but I wish I had the same on my Monstro. The Sony also records gyro information into its video files that can be read by its own Catalyst Software to allow the stabilisation of footage. The Canon C500mk2 goes one step better and stabilises images actively in the body. (3D tracking in camera would be a game changer but accessing that IP may not be straightforward.) The Sony FX9 will also natively transmit video to a phone App that allows others to see the camera image. It's not high resolution and it has a small delay but the camera will even upload proxies to a server around the world as you continue to shoot. Sony has chosen to pioneer these "2020" features in the "medium level" FX9 but if Sony integrates this tech into their high end Venice its a game changer. I seriously hope Red is not ignoring the FX9.

At the ACS we did an in depth look at the FX9 and you can find it here on https://vimeo.com/381979485 or https://www.youtube.com/watch?v=XkoG4JRekNo&list=PLeFHbaPND1RJGtaVqMlGjwzH2DiaTtIrr&index=2&t=0s. Unfortunately the Software stabilisation was released a few days after we published so we did not test it.

 

[Björn Benckert]

Christopher,

I believe manufacturers have already started doing it. Check out the Sony FX9. It has an effective, configurable and practical auto focus system. Not perfect AF but I wish I had the same on my Monstro. The Sony also records gyro information into its video files that can be read by its own Catalyst Software to allow the stabilisation of footage. The Canon C500mk2 goes one step better and stabilises images actively in the body. (3D tracking in camera would be a game changer but accessing that IP may not be straightforward.) The Sony FX9 will also natively transmit video to a phone App that allows others to see the camera image. It's not high resolution and it has a small delay but the camera will even upload proxies to a server around the world as you continue to shoot. Sony has chosen to pioneer these "2020" features in the "medium level" FX9 but if Sony integrates this tech into their high end Venice its a game changer. I seriously hope Red is not ignoring the FX9.

At the ACS we did an in depth look at the FX9 and you can find it here on https://vimeo.com/381979485 or https://www.youtube.com/watch?v=XkoG4JRekNo&list=PLeFHbaPND1RJGtaVqMlGjwzH2DiaTtIrr&index=2&t=0s. Unfortunately the Software stabilisation was released a few days after we published so we did not test it.

Camera gyro stabilization needs to know the lens, Red had this for a longer than all the others trough their early EF an nikon mounts. Aslo they been on the forefront on resolution. With a decent gyro and very little bit of coding red would have killed the competition for kinds of camera stability work. It hurst, it´s like watching someone jumping up way above the basket holding the ball firmly with both hands... and then for some odd reason not slam dunking it.

 

[Nick Morrison]

Christopher,

I believe manufacturers have already started doing it. Check out the Sony FX9. It has an effective, configurable and practical auto focus system. Not perfect AF but I wish I had the same on my Monstro. The Sony also records gyro information into its video files that can be read by its own Catalyst Software to allow the stabilisation of footage. The Canon C500mk2 goes one step better and stabilises images actively in the body. (3D tracking in camera would be a game changer but accessing that IP may not be straightforward.) The Sony FX9 will also natively transmit video to a phone App that allows others to see the camera image. It's not high resolution and it has a small delay but the camera will even upload proxies to a server around the world as you continue to shoot. Sony has chosen to pioneer these "2020" features in the "medium level" FX9 but if Sony integrates this tech into their high end Venice its a game changer. I seriously hope Red is not ignoring the FX9.

At the ACS we did an in depth look at the FX9 and you can find it here on https://vimeo.com/381979485 or https://www.youtube.com/watch?v=XkoG4JRekNo&list=PLeFHbaPND1RJGtaVqMlGjwzH2DiaTtIrr&index=2&t=0s. Unfortunately the Software stabilisation was released a few days after we published so we did not test it.

I have a feeling RED understands where the industry is going.

You can see them testing ideas in the Komodo.

A lot of that will get folded into DSMC3.

I have no fear.

Also I'm looking forward to an 8K Gemini sensor, or something similar that's IMAX size.

Full Frame is starting to feel so 2019.

 

[Mike P.]

I have a feeling RED understands where the industry is going.

You can see them testing ideas in the Komodo.

A lot of that will get folded into DSMC3.

...

Full Frame is starting to feel so 2019.

Yeah no doubt, but Komodo isn't out yet (much less DSMC3), whereas most other manufacturers have a lot of these features implemented already (especially the FX9) and have had some for quite a while now (like AF). If RED understands where the industry is going, it'll be because everyone else is already there (except Arri, but they still have the best image with 10year old sensor tech, so they get a pass).

Camera gyro stabilization needs to know the lens, Red had this for a longer than all the others trough their early EF an nikon mounts. Aslo they been on the forefront on resolution. With a decent gyro and very little bit of coding red would have killed the competition for kinds of camera stability work. It hurst, it´s like watching someone jumping up way above the basket holding the ball firmly with both hands... and then for some odd reason not slam dunking it.

On the topic of gyro data, what's to stop a phone from being used like a "SteadyXP unit"? That'd be pretty ideal for Komodo since it's looking like phones will be the primary display for a lot of people (so it could be built into the RED display app). Similarly, it wouldn't be surprising if Komodo's "global shutter" is actually in-camera/actively applied gyro stabilization (like the C500m2's) combined with a really-fast rolling shutter that mitigates 99.99% of skew.

As other's have mentioned, I'm inclined to believe the reasons for these things not being implemented sooner/faster is likely IP, licensing, and legal reasons, more than anything else/engineering.

 

[Björn Benckert]

Yeah no doubt, but Komodo isn't out yet (much less DSMC3), whereas most other manufacturers have a lot of these features implemented already (especially the FX9) and have had some for quite a while now (like AF). If RED understands where the industry is going, it'll be because everyone else is already there (except Arri, but they still have the best image with 10year old sensor tech, so they get a pass).



On the topic of gyro data, what's to stop a phone from being used like a "SteadyXP unit"? That'd be pretty ideal for Komodo since it's looking like phones will be the primary display for a lot of people (so it could be built into the RED display app). Similarly, it wouldn't be surprising if Komodo's "global shutter" is actually in-camera/actively applied gyro stabilization (like the C500m2's) combined with a really-fast rolling shutter that mitigates 99.99% of skew.

As other's have mentioned, I'm inclined to believe the reasons for these things not being implemented sooner/faster is likely IP, licensing, and legal reasons, more than anything else/engineering.

Yes I guess it would be quite easy to make an iPhone app making he phone do the same job as a steady XP. But there is a few buts... The phone would need to be mounted in a known way, like flat ontop of camera facing a certain direction. You would also need to feed synk from the phone, so you would need to wire the phone headphone jack / cable into the camera sound in and feed sound tc/sync signal. And Im not sure possibly the phone could run such app, recording the gyro data from the phone while also getting a feed from camera for picture etc. But to me thats a bit messy, not so much better than the steady XP, yes you dont have to deal with the GoPro battery, and a phone is possibly more reliable. But it would kind of, to me, be like replacing one bad solution with another and I think the steady XP is a really really good gyro, phones might not be having the same spec. I think they sample very precise rotations at 160Hz or such. Dont know the frequency of Iphones possibly its the same but could be less. The DSMC2 got something like 10Hz...

Now I would be very surprised if komodo does not have an internal gyro. Very likely Komodo has the same kind of horizontal indicator etc as the DSMC2 has and if it does it also very likely write the gyro data into the r3d´s the same way as DSMC2 does. And at this date it would be strange if the gyro on Komodo, or / and the readout from it is not more accurate and higher frequency than the DSMC2. If so, if its actually fast and accurate then I will give my little Syndicate AXIS app just a little some more attaention and make it so it creates one fusion camera stab setup for each clip on a mag right there in the r3d folders as a one button click operation.

We pretty much have the above done allready, but with the current quality of the DSMC2 gyro output, I just halted the AXIS project as the noise introduced from poor gyro data is bigger than the shakes you normally would want to remove. But we have built it on the last SDK so it will in its current state be able to detect and read Komodo files and if there is a gyro and the gyro data is good. Its basically a gun ready to fire.

At around 55s into this rough demo you see DSMC2 data pulled using AXIS play inside Flame, if this was good gyro data coming in then: roll, pan, tilt would simply be locked / not moving. Now there is a heavy gitter and also a drift due to the inaccuracy of the DSMC2 gyro.