KOMODO....

[Ryan Genova]

I looked for tests but didn't find anything I thought was definitive....

DxOMark has Helium at 15.2 stops (some of their testing methodologies can be a bit weird, but the DR method seems reasonable). There's really nothing I can find for Gemini either, but maybe it's a bit closer to the claimed 16.5 than Helium? Regardless, if Komodo is even somewhat close 15.2 I don't think anyone could be disappointed, right?

 

[David Rasberry]

DxOMark has Helium at 15.2 stops (some of their testing methodologies can be a bit weird, but the DR method seems reasonable). There's really nothing I can find for Gemini either, but maybe it's a bit closer to the claimed 16.5 than Helium? Regardless, if Komodo is even somewhat close 15.2 I don't think anyone could be disappointed, right?

Exactly! What’s not to love about a 15-16 stop global shutter camera with Red DNA at a relatively modest price?

 

[Jarred Land]

I didn't mention it because it might sound antagonistic and I value the leader of a camera company interacting with the users in this way, but I actually had a question about that. Jarred said Komodo will have similar DR to Helium, but also the Gemini has more DR than Komodo, which by default means that Gemini has more DR than Helium. But as you say, Dragon, Helium and Gemini are rated the same by RED everywhere I've seen (see here).

So I'm confused as to the DR differences between the sensors (Dragon; Helium; Gemini; Monstro and now Komodo). I looked for tests but didn't find anything I thought was definitive....

There are a few different ways to test dynamic range, some more subjective than others. Our engineers use science, I look at actual charts, And marketing uses a combination of both. It is confusing as some of our sensors are better at the top and some are better in the lows and its not as simple as just sliding the middle to offset. Plus there is always some non-linearity in the shadows so step charts can be a bit deceiving, and that non-linearity gets improved over time.

For example Helium actually has improved a bit over the last couple years from a hardware standpoint by the engineers and Graeme's work continues to squeeze more out of the sensors from a software standpoint, but I reference Helium #0001 so it is a fair worst case not taking improvements into consideration.

Resolution doesn't technically count in dynamic range , but it actually does if you are oversampling, which is why you can see Monstro still beat out Gemini in the lows even though Monstro is biased to perform better in the highs.

And then you add the low pass filter, and those variations even same to same on different sensors also effect dynamic range...

i'm gonna stop there because I realize im probably starting to confuse things even more now

 

[Bob Gundu]

Not confusing at all. Bottom line, every iteration of sensor or body has been an improvement in image quality.

 

[Eric Lange]

Exactly! What’s not to love about a 15-16 stop global shutter camera with Red DNA at a relatively modest price?

Yeah really / "Ditto".

For a long time globally shuttered sensors have existed but traditionally they have had small sensor sizes and lousy dynamic range and puke awful color science... etc. etc.

So a reasonably larger sized sensor with very decent DR (+ RED's coding scheme) means you can dig very deep on the Art and Science for pretty much everything that a rolling shutter has been very obstructive to (up to now).

In other words beautiful imagery that smashes new technical frontiers ? (many practical and technical uses yet "Grown up" image quality(ies) ~ Finally ! .

+ price ??????

[That may not be interest to everyone but will be interesting to see how various groups / folks decide to run with KOMODO.].

 

[Eric Lange]

There are a few different ways to test dynamic range, some more subjective than others. Our engineers use science, I look at actual charts, And marketing uses a combination of both. It is confusing as some of our sensors are better at the top and some are better in the lows and its not as simple as just sliding the middle to offset. Plus there is always some non-linearity in the shadows so step charts can be a bit deceiving, and that non-linearity gets improved over time.

For example Helium actually has improved a bit over the last couple years from a hardware standpoint by the engineers and Graeme's work continues to squeeze more out of the sensors from a software standpoint, but I reference Helium #0001 so it is a fair worst case not taking improvements into consideration.

Resolution doesn't technically count in dynamic range , but it actually does if you are oversampling, which is why you can see Monstro still beat out Gemini in the lows even though Monstro is biased to perform better in the highs.

And then you add the low pass filter, and those variations even same to same on different sensors also effect dynamic range...

i'm gonna stop there because I realize im probably starting to confuse things even more now

[Emphasis added].

" Our engineers use science"

Lol (I'm glad someone said that).

And I'm glad i'm not the only one that feels a bit underwhelmed / "Meh" about so called Dxo "scores" as a system or method of "evaluation".

I'm old school and look very closely at system MTF s / Combined MTFs but at the same time I also understand that Graeme has a major handle on much deeper less obvious aspects of how higher resolution sensors can be made to "interpret" and process "color " vs resolution in a whole host / smorgasbord of different ways. - [ There's a lot of cool sh*t you can do in the frequency domain that a Dxo whatever test is not gonna elucidate.].

Nice to know that Graeme is still on the "Case" in a hands on way … :coolgleamA:

 

[Ash Worth]

Science is great however if you claim 17+ stops and the image doesn't look that different than the 14+ stop cameras then there's misleading perception in the marketplace. A 3+ stop difference is big, there needs to be proof that it exists because I've not come across too much documentation that says it does.

 

[Lee Saxon]

And I'm glad i'm not the only one that feels a bit underwhelmed / "Meh" about so called Dxo "scores" as a system or method of "evaluation".

I think you are reading too much into that. If you look at the rest of that sentence, it seems he was just talking about the various ways *Red* looks at things, not criticizing any other organizations. Feel kinda bad for Jarred, he has to be quite careful what he posts and y'all are often still trying to get him in trouble extrapolating things.

 

[Eric Lange]

I think you are reading too much into that. If you look at the rest of that sentence, it seems he was just talking about the various ways *Red* looks at things, not criticizing any other organizations. Feel kinda bad for Jarred, he has to be quite careful what he posts and y'all are often still trying to get him in trouble extrapolating things.

I'm not dissing your sentiment here … that's your view through the prism.

Prefacing something with 'We use science" speaks for itself...

I'm not sure "Who" you think by extension that in an abstract sense that various extrapolations on a forum get Jarred in trouble with whom ?

He's a grown up and knows his business. What he writes is what he chooses to write.

What I write is not going to "Get Jarred in trouble " correctly construed or mis-confabulated ?

But nice that you care that Jarred does not get into trouble with DXO group etc.

For me MTFs i.e. lens MTFs (sagittal and radial) and sensor MTFs + MTFs of signal processing + electronics + final output ---> Combined system MTFs

Beyond that for deeper and more technical applications + the 'ART" each project / application has special needs and requirements that a DXO evaluation is not really gonna help much ?


Some people like those kinds of high level mashed together single number analyses … I don't.

So I dare express my opinion that I don't find the "DXO" thing useful. That's not to say DXO is "Bad" it's not a good nor appropriate fit for everything.

Jarred is making the point that perhaps due to the inherent nature of resolution versus color space in the Various flavors of R3D that perhaps DXO tests are not always that meaningful nor appropriate.

 

[Jarred Land]

heh heh.. the only ones that I can get in trouble is with you all , so don't worry about me.

And I am not criticizing DXO or anyone else.... not sure how this conversation went down that road.. But measuring Dynamic Range as I said is very much subjective and all depends on how you measure.

If you have ever shot a latitude chart, opened the files and then started counting the boxes you almost always say an "ish" for the very last 1/2 , 1/4 or 1/8th of a stop. the value of the "Ish" is the subjective part.

And... As much as you would assume that Engineering measures conservatively and then Marketing turns that into something higher.... It is actually usually the other way around. There is a difference between technical dynamic range and visible dynamic range. As much as some may think our numbers are inflated, marketing actually ( usually ) ends up falling on the visible side. To me the only important number is visible.. the amount of stops you can actually stuff in your image, subjective or not. I think Phil did a article a few years ago actually testing visible DR between the sensors.

Anyways... The comparison is much more telling than the numbers. It takes the subjectivity out of it.. If you shoot a chart shot with Monstro and Gemini and Helium and Komodo , all side by side, the difference between them, at least to me, is a much better reference than just counting the boxes.. When you compare the difference between you generally are equalizing your own subjectivity . your own definition of "ish" is measured across all.

 

[Andy Jarosz]

Hi, just making sure Komodo will still have the oil extraction feature as well as come bundled with a coupon to redeem three wishes. If possible, I'd also like to request it be able to turn into an inflatable life raft. Thank you

 

[yann LHENORET]

Hmmmm. Modified for Komodo perhaps?

https://www.instagram.com/p/B6323EiBwif/

Very nice!!

 

[Phil Holland]

I think Phil did a article a few years ago actually testing visible DR between the sensors.

Yes sir and I suspect I'll be writing about it more coming up.

And the latest round of "big boy" camera tests surrounded mainly finishing for UHD 4K with each camera system from each manufacturer.
When it comes to measuring what I call Total Captured Dynamic Range I do it several ways to cross reference my results to come up with a solid rather controlled answer. The individual patch test however takes forever, but is worth doing.

Getting away from the numerical landmark, much of the conversation these days has been extended to "what is a usable stop".
But really what is going on is for a given image pipeline and production workflow where is a studio comfortable with working with each camera system.

This is why we test everything at all available ISO Ratings to discover how far we can push it to finish out to most recently a UHD 4K finish.

The fun thing here is we also are not just showing a still of a Xyla, Patches, or Measured Stops; we also show it in motion, which I've started to upload in my public tests more often. That's where it gets hairy because on a still image you see one measurement and in motion a different one. More or less much of my numerical testing is to make sure manufacturers aren't claiming one thing and we are seeing something else entirely, which actually still happens. I'll just take Monstro for instance, in motion and even in still you can "see" 19 stops. But that isn't what people will use and certainly not use safely. And those last stops are noisy. 17+ is a relatively safe call and I totally get why + is included especially taking into account different OLPFs and even shooting "clear".

People get lost in the numbers pretty often because it's hard to really equate what that means and how that information is used. Especially way up high and way down low.


The only real thing happening at the moment is a studio wants to know what ISOs a system shouldn't be used with because too much noise is a dead end unless you are aiming at a specific aesthetic that needs a crazy amount of noise. We also test for light source issues and a variety of other chaos, but I keep it simple.

This is one of RED's strengths because oversampling helps a bit on this front and famously there is no in camera noise reduction in the RED world mainly to keep as much captured color and detail as possible.


Just to remind people, motion picture film is 13 stops of Total Captured Dynamic Range. If you really know what you are doing you can squeeze 15 stops out of it. It's really how film sees light and color and translates that into an image via it's own unique properties and development methods. Much of that in truth is related to how film is exposed layer by layer. Digital has surpassed this on technical merit, that's not even really a question at that point. For me however it should be the baseline of what we should want and where we can go from here.


I'll end the evening on a mind poisoning thought. It's all about how you use those stops.


Just my opinion, but if Komodo lands where Jarred is stating, everybody should be pretty happy especially at this price point.

 

[Robert Parkin]

Is it too late to add pogo pins on the side for record start/stop functionality through a handle? Seems such a shame to miss this on a camera that can be set up so minimally. Catching one off moments is easier when you don't have to move your hands to press the record button. Just a little teeny tiny wish!

 

[Christopher A. Bell]

I remember back in the 'Old days" making the switch from Film to Digital that the dynamic range was less than what is available now (almost substantially so) and yet we still were able to generate world class imagery … More mindful of lighting in some instances.

I've brought up similar thoughts on other forums many times before. I started out back in the 90's on SD analog Betacam. What we have available now is insane compared to just a few decades ago, BUT it's ironic that now we have cameras with so much more DR and the ability to shoot in LOG and RAW and the ability to literally 'fix it in post', but I see so much more over exposed, blown out shots with horrible color than I ever did when we "had" to get it right in-camera. We were more careful and mindful. Now because so many think that it can all be done in post, no care is taken during acquisition.

 

[Jarred Land]

Is it too late to add pogo pins on the side for record start/stop functionality through a handle? Seems such a shame to miss this on a camera that can be set up so minimally. Catching one off moments is easier when you don't have to move your hands to press the record button. Just a little teeny tiny wish!

no room on the side for pogos but we have those pins on the top... and there will be a new outrigger handle for Komodo that uses them.

 

[Jarred Land]

The fun thing here is we also are not just showing a still of a Xyla, Patches, or Measured Stops; we also show it in motion, which I've started to upload in my public tests more often. That's where it gets hairy because on a still image you see one measurement and in motion a different one. More or less much of my numerical testing is to make sure manufacturers aren't claiming one thing and we are seeing something else entirely, which actually still happens. I'll just take Monstro for instance, in motion and even in still you can "see" 19 stops. But that isn't what people will use and certainly not use safely. And those last stops are noisy. 17+ is a relatively safe call and I totally get why + is included especially taking into account different OLPFs and even shooting "clear".

People get lost in the numbers pretty often because it's hard to really equate what that means and how that information is used. Especially way up high and way down low.

Yes I really like the way you test Phil.. and it's a nice bias to real world, which should be what matters the most.

Internally, If you ask Myself, then ask Graeme, and then ask our Sensor Engineers how many stops any given sensor of ours has, you will usually get 3 different answers. Just like if you ask 3 different D.P.s how many stops any given sensor of ours has ( or any camera for that matter ) , you will likely get 3 different answers.

 

[Alex Stolpakov]

no room on the side for pogos but we have those pins on the top... and there will be a new outrigger handle for Komodo that uses them.

Do you have any pictures in the works?

 

[Lee Saxon]

not sure how this conversation went down that road...

Yeah, and I shouldn't have responded. Just wild sometimes how much speculation and extrapolation the least comment (or a little phone snapshot) from you can start.

Question:

Since that top port isn't monitor-out signal like DSMC, and we know a lot of indies are going to want this as an A cam despite its intended usage, I'm curious about the choice to go SDI-only since an HDMI port would've allowed far more affordable monitors.

Also, since you've mentioned that Komodo was not at all conceived as a "Hydrogen camera module", can you say more about the choice of the CSI interface? Was the camera originally not going to have its own screen? Can you talk yet about any of the use cases you're imagining for what y'all will do with the CSI link?

 

[Hrvoje Simic]

Internally, If you ask Myself, then ask Graeme, and then ask our Sensor Engineers how many stops any given sensor of ours has, you will usually get 3 different answers. Just like if you ask 3 different D.P.s how many stops any given sensor of ours has ( or any camera for that matter ) , you will likely get 3 different answers.

Yes, the perception varies between total signal range in vacuum to cinematographic context of light, contrast and density.
With trips in between to pink unicorn land.