KOMODO....

[Neil Abeynayake]

I think at $2495 Red could sell a LOT of Komodos.

Probably more Komodos than Red would want to sell even maybe.

Could become more of a liability almost at that volume level.

Baller move to license that canon RF mount if that's what happened.

100% agree with you Dan.

 

[Neil Abeynayake]

The $7k price being thrown around is what I mentioned for a complete entry level package, Komodo and phone with some lenses, media, batteries, essential accessories. That's about what I have in my Bolex kit. About the most I can justify budget wise for a serious hobby camera.
I hope well underl $5k for the Komodo alone, plus the cost of the phone keeping it under $6k.

There is more to true value than initial price. BM sold thousands of the original 1080p pocket cams at $995 or less, compared to $3995 for my 2K DCI D16 512K version. Five years in my D16 is still functioning flawlessly and used ones if you can find one sell for close to original purchase price. There were less than a 1000 total made. I'd venture to say that a third or more of the pocket cams that old are already in a landfill. It was not a very robust camera.

Where did you get those numbers from? If it is pure speculation then, you are full of it. I spoke with BM marketing folks and they'd dispute that number by tenfold compared to their world-wide sales.

 

[Phil Holland]

Strangely I'm privy to a lot of market numbers of successful and unsuccessful cameras. Reliability, market response, support, etc.... Lots of reasons things work and don't work in the short or long term.

The hard truth is there are different market tiers when it comes to price point which is also generally represented in features as well. Though all of these cameras are technology and technology advances, some of the low tier/low priced models are doing things a few years ago would have cost you a fair bit of cash.

Here on REDuser you've been hearing me say "the industry is very dynamic" a whole bunch. Some of this is based on my localized experience of working with companies, individuals, and projects as they need technical solutions to creative problems. I've found myself consulting for new agencies, video game live stream events, camera testing/certification for various entities, and the normal daily influx of "what camera should I buy" questions from the full spectrum of filmmakers ranging from huge to beginners.

The lower pricing tier cameras sell like hot cakes and along the lines of some of the concessions that are typically made in these arenas surround reliability and quality control. This is also related to how much and how hard you use the cameras. If you're paying a premium for a digital cinema camera these days, typically you are getting something fairly robust and extremely well supported if things go pear shaped somehow.

The volume of these smaller cameras out there exceeds the high end stuff by a very large amount because the market is there for that. I think the more interesting tier stuff that's occurred in digital cinema cameras is also related to the DSLR/Mirrorless revolution as well. I've brought up the frightening absence of the $8000 DSLR from the market before, but the market of motion picture production has exploded since those days and much of the effort has been about developing the expanding world of various types of cameras with many focused on still and motion. On the digital cinema side of things, since RED has been early to the game of high resolution motion acquisition, naturally they translated well into the still world with various magazines. The DSLR/Mirrroless crowd has been going the opposite route where naturally they are good at stills, but most of the those cameras are considered entry level when it comes to motion work, despite having impressive capabilities.

Every manufacturer knows the lowest price point will sale the most in volume. I would dare say the new BMD Pocket 6K shipping so fast is likely partially a response to the 4K supply/demand/manufacturing issues that were alluded to in the stream, but also echoed by the horde of people I see online who pre-ordered and haven't received their cams yet. Tough world honestly on that front as I feel for them and BMD in this case.

Komodo is certainly the lowest price RED camera to date and it will be compared ad nauseam to anything near it's price point and likely to things well above it whatever the final feature set ends up being. The general world of click bait driven content for things like "iPhone versus RED" is YouTube gold in so many ways and easy for creators to run through. Me, I'll certainly test it out to see what the hell I can do with it and where it's ceiling and floor is, mostly what's inspiring for me is the form factor and I'm hoping the image quality is there to fit into the world I produce work for, which I'm feeling fairly good about considering all things. I don't think it's going to be chewing up against DSMC2 cameras in terms of features or image quality too much despite predictable fears of such things.

And I've already mentioned here, there's stuff going on in Komodo that I 100% feel will benefit the world of DSMC3, which I'm equally if not more excited about as that's my A Cam world and the bread and butter tool I use for making pretty pictures.

 

[Phil Holland]

Oh, just to respond the question I put forth earlier.

I'm hoping this is a 6K DCI aspect ratio sensor, though it would be really nice mainly for anamorphic peeps if it was something like 3:2 as well or the Gemini aspect ratio, but we just don't know the sensor size at all really.

So taking that into consideration, my hopes are:

6K DCI (FF in the RED world, 6144x3240) landing somewhere between 24-60fps. Maybe 48 or 50fps might make a lot of sense, unknown as to what the hardware can do. Whatever is going on at max resolution needs to be "at speed" for major markets, which would be at least 30fps.


I don't think this sensor will have the horsepower of Monstro or Helium, which top out at 80fps at 6K FF.

My next major concern is sensor size as that will play a role in some of the applications I can see for Komodo, but I'll pocket those thoughts for now until well know more. Glass-wise, whatever format size is in there, we have options. But I'm thinking about things like 360 video and similar use cases where you might be using a few Komodo cubes in an array of some sort, ideally genlocked *prays for that feature*. In terms of being viable as a B-cam, action cam, crash cam, POV cam and if this sucker is at least Super 35mm, I think personally that's a safe bet. I don't think RED wants any part in the tiny sensor game (like MFT) for now.

 

[Mathias Erichsen]

I’m starting to think that Komodo has a S35 sensor and that it’ll have an option for a Speedbooster in the mounts RED sell. Canon have allegedly been working on their own version of a speedbooster/focal reduser for the RF mount, so maybe this is it?
This way you’ll have S35 if you prefer, or a FF equivalent if that’s what you want
If this is it, then I hope the new PL mount with speedbooster will be able to take most, if not all, PL lenses, and not like the current PL speedbooster that only fit a handful of lenses...

I’d really like to see 50fps @ 6K

 

[Ugo Del]

I think at $2495 Red could sell a LOT of Komodos.

Probably more Komodos than Red would want to sell even maybe.

Could become more of a liability almost at that volume level.

Baller move to license that canon RF mount if that's what happened.

We just have to wait for the usual hundreds of «*less than 50 hours*» used Komodo for 2k$ on the 2nd hand market next year. No rush. I don’t think there is any other camera brand so heavily linked with G.A.S.

 

[Hrvoje Simic]

That's what I mean to say. Pocket 4K at ISO 1000 is not bad, but I find the Ursa Mini Pro has better overexposure latitude, especially at 800 or 1600. Anything above ISO 1000 on the pocket when using the first gain stage is a bit noisy.

Attempt #458

Digital ISO equivalent dial has no effect to lattitude nor DR whatsoever, only analog gain implementation affects this. The rest is an image presentation software fantasia. Image differs in capture just because you expose differently.

 

[Misha Engel]

Some weekend reading:
https://www.premiumbeat.com/blog/exclusive-interview-reds-sensors-resolution/

That's an interview with Graeme from a bit ago in which he moderately deep dives, but mostly hints about topics surrounding ideal resolutions for Bayer Pattern Sensors.

For resolution purists, traditionally you ideally want to have somewhere between 71% to 80% over provisioning of resolution towards your delivery resolution. i.e. 4096 / .80 = 5120.

Outside of the Bayer Pattern there's things like the OLPF/Anti-Aliasing Filter that will impede on potential captured resolution very much on purpose to ideally avoid aliasing and moire via potential high frequency detail, which is a real issue in production, post production, and VFX.

To highlight an often overlooked part of IPP2, there is a new demosiac algorithm at play and you'll notice a slight improvement in rendering captured detail over the previous gen. One of the benefits of REDCODE RAW really is it's actually raw data, though done through a variable wavelet compression scheme. So those particular advancements in color science and image processing can be discovered, unlocked, and put into place. In the RED world even based on the general resolution scheme the cameras deploy you can observe they are much closer to an ideal 80% oversampling towards the finishing resolution. More so now after IPP2 new jazz.


That said, there's nothing wrong with shooting 4K for 4K, many people do it. I personally shoot 5K and 5K+ for a 4K finish and that's my preferred workflow with 8K for 4K being rather ideal on many levels in my pea sized brain. And I'm not even getting into the various scaling algorithms you can tap into when developing your image.

Nice story, already read it a while ago. I still haven't seen any prove that 5120 bayer pixels is enough to get a perfect 4k-DCI (4096 RGB-output) (prove as in a scientific paper, checked and approved by other scientist), what I have seen is that Venice at 6k and Helium at 7k gives a perfect debayer to 4k-DCI/UHD (RGB). Up to now we always use a Helium as A-cam when a customer demands perfect debayered 4k-DCI/UHD output.

ARRI uses 3200x1800 with the AMIRA to produce a UHD RGB output, I don't know what to make of that.

Sony uses 3 8K(UHD) cmos sensors in it's most high-end 8K UHD camera, https://pro.sony/en_FI/products/4k-and-hd-camera-systems/uhc-8300 I think they have good reasons to do this.

Mr. marketing himself Michael Cioni claimed that you need around 11k to deliver in true 8k (for once I agreed with him).

Convident camera manufatures just send their 4k+ camera's to Slashcam.de to have them tested in a German scientific way for free (ARRI isn't very confident about their 4k performance).

 

[Jacek Zakowicz]

Nice story, already read it a while ago. I still haven't seen any prove that 5120 bayer pixels is enough to get a perfect 4k-DCI (4096 RGB-output) (prove as in a scientific paper, checked and approved by other scientist), what I have seen is that Venice at 6k and Helium at 7k gives a perfect debayer to 4k-DCI/UHD (RGB). Up to now we always use a Helium as A-cam when a customer demands perfect debayered 4k-DCI/UHD output.

ARRI uses 3200x1800 with the AMIRA to produce a UHD RGB output, I don't know what to make of that.

Sony uses 3 8K(UHD) cmos sensors in it's most high-end 8K UHD camera, https://pro.sony/en_FI/products/4k-and-hd-camera-systems/uhc-8300 I think they have good reasons to do this.

Mr. marketing himself Michael Cioni claimed that you need around 11k to deliver in true 8k (for once I agreed with him).

Convident camera manufatures just send their 4k+ camera's to Slashcam.de to have them tested in a German scientific way for free (ARRI isn't very confident about their 4k performance).

There is a very easy way to verify 4k, I've been doing it for years:
Point the camera at a 4k target, record, review footage to see if you can make out the 4k target. If not see which lower resolution target you can make out.
DONE!!!!

 

[Savannah Miller]

Attempt #458

Digital ISO equivalent dial has no effect to lattitude nor DR whatsoever, only analog gain implementation affects this. The rest is an image presentation software fantasia. Image differs in capture just because you expose differently.

Yes, but exposing differently is also how you rate the camera. If you rate the cam as a 1000 ISO cam, then you have made a pretty clear choice as to where you want your dynamic range to sit. And in 10-bit Prores I wouldn't feel as confident adjusting ISO/exposure, even though you could linearize the images and evenly gain them.

 

[Misha Engel]

There is a very easy way to verify 4k, I've been doing it for years:
Point the camera at a 4k target, record, review footage to see if you can make out the 4k target. If not see which lower resolution target you can make out.
DONE!!!!

Use 1 red, 1 green, 1 blue and 1 white background 4k chart, when all give a 4k resolution sharpness you can't go wrong.

 

[Hrvoje Simic]

Yes, but exposing differently is also how you rate the camera. If you rate the cam as a 1000 ISO cam, then you have made a pretty clear choice as to where you want your dynamic range to sit.

If you "rate" a camera through digital gain ISO equivalent rating, you are doing nothing with the camera's capture properties, you are just using a different presentation of what is captured, imagining that the camera is more sensitive and exposing lower.

You are not gaining any frikking thing in the image with higher digital "rating", it is just a software boost of the same frikking image, which, if compensated with lower exposure leads to reduced bits and image density. "Rating higher" you don't just "relocate DR" literally as if it's movable thing, you capture less light .

You are not just "trading shadow range for highlight range" that way, as understood if taken literally and oversimplified, you are trading shadow reach AND image density, tonal richness, the "depth" of the whole image.

 

[Phil Holland]

I like the good folks as slashcam, they follow me a fair bit. I'll post some charts up when I can Misha, likely at the end of the week. Once again, there's the googled answer and the results answer and there is a reason I prefer to be at 5K or 5K+ for a 4K finish with all the current cameras here. This is one of those situations where more is better if you are looking to improve a result, but more importantly having access to very high resolving lenses for such a test is important as well. Stay tuned.

 

[Misha Engel]

I like the good folks as slashcam, they follow me a fair bit. I'll post some charts up when I can Misha, likely at the end of the week. Once again, there's the googled answer and the results answer and there is a reason I prefer to be at 5K or 5K+ for a 4K finish with all the current cameras here. This is one of those situations where more is better if you are looking to improve a result, but more importantly having access to very high resolving lenses for such a test is important as well. Stay tuned.

I hope you can convince me to go the 5k route for 4k-DCI delivery, rolling shutter isn't the best on Helium 7k and you need an OTUS 85mm quality lens to get the max out of those tiny pixels.

For Venice or Gemini we would need to change our S35 lenses to FF lenses.

Dragon-X would be the best camera for our work(flow) when it's output can be de-bayered to a perfect 4k-DCI output.

 

[Christoffer Glans]

if it was Full Frame it would compete with the Alexa Mini LF for under 10K...

problem with Dragon 6K was that it was larger than super35 and a lot of cinema lenses are super35 - which isn't the case anymore - e.g. if you had a set of Zeiss Ultra Primes you would need to window the sensor down to 5.5K or 5K...

This might make the Medium Format Camera, that Jarred post seem relevant and if someone developed speedbooster to use Medium format lenses on a full frame sensor...

It all depends on compression rates, framerates etc. The Alexa Mini LF is a cinema grade camera, if the Komodo is closer to things like the Canon R, but with a high compression R3D Raw infrastructure instead, that would be a one of a kind camera without really disrupting Alexa Mini LF.
But even if it did, do you think Red would care if they disrupted a competing camera company's flagship low-profile camera?

Problem with Dragon 6K was that it was larger than super35 and a lot of cinema lenses are super35 - which isn't the case anymore - e.g. if you had a set of Zeiss Ultra Primes you would need to window the sensor down to 5.5K or 5K...

There are plenty of professional grade lenses that cover Dragon and even if some don't, most used the extra resolution as a look-around and safe-net since most normal lenses still covered everything. Most grade a series of lenses cover-area by how the lowest focal point looks on it. If a 24mm vignettes slightly, there's a good change the 35mm or 50mm don't.
Even so, with the RF, it's easier to make lenses for larger sensors and it's a vastly superior mount which will open up a lot of interesting choices. If the sensor is Dragon-sized, it won't be a problem, much less today since everyone goes full format.

A medium format camera would probably be a $300K brain only camera, just imagine the yield on that sensor production.

 

[Nick Morrison]

It all depends on compression rates, framerates etc. The Alexa Mini LF is a cinema grade camera, if the Komodo is closer to things like the Canon R, but with a high compression R3D Raw infrastructure instead, that would be a one of a kind camera without really disrupting Alexa Mini LF.
But even if it did, do you think Red would care if they disrupted a competing camera company's flagship low-profile camera?

Problem with Dragon 6K was that it was larger than super35 and a lot of cinema lenses are super35 - which isn't the case anymore - e.g. if you had a set of Zeiss Ultra Primes you would need to window the sensor down to 5.5K or 5K...

There are plenty of professional grade lenses that cover Dragon and even if some don't, most used the extra resolution as a look-around and safe-net since most normal lenses still covered everything. Most grade a series of lenses cover-area by how the lowest focal point looks on it. If a 24mm vignettes slightly, there's a good change the 35mm or 50mm don't.
Even so, with the RF, it's easier to make lenses for larger sensors and it's a vastly superior mount which will open up a lot of interesting choices. If the sensor is Dragon-sized, it won't be a problem, much less today since everyone goes full format.

A medium format camera would probably be a $300K brain only camera, just imagine the yield on that sensor production.

I think the 6K Dragon sensor size was only a problem when it came out - at the time the industry was entirely focused on S35.

But since then...much has changed.

Full Frame sensors are all the rage now, and even Arri embraced S35+ years ago with Open Gate.

If the Komodo has a 6K 30mm wide sensor, that is only a GOOD thing.

If it's a 6K 25mm wide sensor (Super 35), that's fine too.

Anything smaller will feel less disruptive, especially for the cost (10K shooting package seems safe).

 

[Gregory Karydis]

Re the sensor aspect ratio & FPS debate:
Anything over S35 will be great for me as I would love to use use all of my EF glass and hopefully in FF 3:2.
If we could have the same dimensions as the 6K Dragon, I'd be over the moon! :emote_happyhappy:
Considering the smaller brain size and the electronics you could fit in there, it would be fair to expect something between 30 and 50 fps at most!
Probably higher FPS at crop sensor resolutions.
Anything over that and we risk cannibalising the DSMC2 line.

But I'm thinking about things like 360 video and similar use cases where you might be using a few Komodo cubes in an array of some sort, ideally genlocked *prays for that feature*. In terms of being viable as a B-cam, action cam, crash cam, POV cam and if this sucker is at least Super 35mm, I think personally that's a safe bet. I don't think RED wants any part in the tiny sensor game (like MFT) for now.

I can already think of a few implementations where a Komodo paired to a Hydrogen device can have some sort of very reliable and wireless type of genlock... :wink:

Absolutely agree on the last bit about MFT sensors.

 

[Andrew R Lewis]

What if Red isn’t worried about cannibalizing DSMC2? What if this camera does disrupt NEW DSMC2 sales but Red isn’t worried because DSMC3 is right around the corner? Maybe Komodo is more than we expect.

 

[Gregory Karydis]

What if Red isn’t worried about cannibalizing DSMC2? What if this camera does disrupt NEW DSMC2 sales but Red isn’t worried because DSMC3 is right around the corner? Maybe Komodo is more than we expect.

I understand what you are saying, but:
1. as a Helium owner, how would that make you feel?
2. I kinda find it hard to fit all of the high-horsepower electronics into such a small (and hopefully energy efficient) brain...

 

[David Rasberry]

If you "rate" a camera through digital gain ISO equivalent rating, you are doing nothing with the camera's capture properties, you are just using a different presentation of what is captured, imagining that the camera is more sensitive and exposing lower.

You are not gaining any frikking thing in the image with higher digital "rating", it is just a software boost of the same frikking image, which, if compensated with lower exposure leads to reduced bits and image density. "Rating higher" you don't just "relocate DR" literally as if it's movable thing, you capture less light .

You are not just "trading shadow range for highlight range" that way, as understood if taken literally and oversimplified, you are trading shadow reach AND image density, tonal richness, the "depth" of the whole image.

Have to agree with Mr. Simic on this one. The DR of any sensor is a fixed range between the noise floor voltage of a pixel and the full well voltage at clipping. That is hard physics. On say a 13 stop sensor if you expose for equal shadow and highlight range you will have 6 stops above and 6 stops below middle gray. The exposure setting, aperture + shutter speed + light, that produces that middle gray exposure represents the native ISO for the camera and the maximum total DR that can be extracted. In practice you have about +/- 2 stops around that center value to preserve a high quality image as far as light hitting the sensor. Moving exposure outside of that range generally results in progressively more flat lifeless images regardless of what gain structure and processing are applied afterwards.