komodo gemini

[Ashley Stracke]

think komodo form factor would ever have a gemini LL style sensor in the future? worried ill buy a gemini and then regret it due to some new cam..

 

[M Harvey]

Looks like nobody has specific intelligence pointing to that in the immediate future. (Or they do, and they're not sharing!)

The product refresh cycle has become so fast now, with the availability of exciting new technologies ever imminent, that it's natural to have the concern that your new camera will instantly become obsolete. I get where you're coming from. For me personally, I find it helpful to remind myself that the camera I own now won't become less good simply because a new camera just arrived.

My favorite thing I shot in the last year was on a DSMC1 Epic Dragon, which came out in 2012. And I regularly use a Sony FS7 (born 2016) for all kinds of things. Check out the cameras used in David Fincher's films ten years ago-- and these movies all still look great.

 

[Andrew Reese]

I own both Komodo and Gemini and a friend of mine also owns Gemini, Komodo and a Dragon X. We both find Gemini IQ superior to Komodo so don't get too caught up in "newer" is better mindset. I personally still love my Komodo and it's a great B-cam next to DSCM2 cameras and a great A-cam on its own but the image just feels much "thicker" and more filmic out of Gemini so much so that Komodo hasn't been getting used as much as we thought. Monstro/Gemini are still king IMO in the realm of maximum IQ and dynamic range and even if RED came out with a FF 8k komodo sensor or a Komodo with dual ISO the Gemini would still be my first choice because of its larger pixels and just how beautiful the image is from that sensor. Don't get me wrong, the form factor of Komodo is killer for run and gun and hard to shoot spaces and it definitely serves a great purpose in my arsenal but DSMC2 cameras are still a cut above in terms of maximizing IQ and rigging them up with tons of accessories on more demanding productions. I think now would be a great time to buy Gemini at a discount since it has been out for a while and with all the attention Komodo is getting and trust me you won't be disappointed. Also, keep in mind DSMC2 cameras are in the upgrade pipeline to DSMC3 where I believe Komodo is not...

 

[Nicholaus James]

RED Gemini FF 6K DSMC3, just saying........

 

[Andrew Reese]

What interest me the most for the future of RED cameras would be a 6-8k FF global shutter sensor in a Komodo form factor but with better HFR capabilities and larger photo-sites. FF obviously seems to be the next step for Komodo and the next generation Monstro but one major change I think needs to happen going forward.... In order to get my attention larger sensors will not just need to provide increased resolution but rather they will need to have a pixel pitch more comparable to Alexa (+/- 8.0 μm).

Gemini currently has the largest pixels out of any RED sensor at 6.0 μm which is even larger than Monstro pixels at 5.0 μm. Full frame sensors and increased resolution is great but not if the pixels are very tiny like on Komodo, Helium or even worse UMP12K. Take a look at this list I posted over on DVXUser a while back below for some interesting comparisons....

I personally find that increased resolution is not as much a benefit (outside of green screen work or frame grabbing for prints) as is the sensitivity and quality of the photo sites and I still firmly believe this is why many prefer Alexa image over competitors higher resolution sensors that may seem to "outperform on paper" but just can't get the same quality in terms of the "look" as these lower resolution sensors that focus on "quality over quantity". We too easily get caught up in resolution today instead of prioritizing Image Quality (i.e. dynamic range and characteristics of highlight retention, shadow information and color science during under and overexposure).

"Resolution" is just a format and does not mean it is "better or worse" and is really only a benefit if the production prioritizes it for the job. Same goes for "Sharpness" as many still say Gemini is "soft compared to Helium" which I have always found funny and can tell you is absolutely not the case when using sharp lenses and hitting your focus marks. Alexa at 3k can produce very sharp images and so can Gemini at 5k and Helium at 8k and side by side in a 1080 or 2k timeline you would be hard pressed to judge a difference in perceived sharpness between any of those cameras if not re-framing.

I hope image quality is prioritized more in the future over marketing terms like "full frame" and "12k resolution". Arri seems to understand this and it is evident in their soon to be released 4k S35 line of cameras. To this day, I still find something special about the footage I capture with my Gemini and there is a unique balance between perceived sharpness / micro-contrast and the smoothness of how highlights are treated within the image and the cleanliness of lifted shadows and I think this all boils down to the science behind the photo sites and not the size of the sensor or the formats it can record.


Sony A7siii - $5000
+/- 12 DR (Usable Stops)
4k @ 10-bit / 16-bit RAW
Sensor Size: 35.60mm x 23.80mm
Photosite Pitch: 8.40 μm

ALEXA Mini LF +/- $70-80,000
+/- 14 DR (Usable Stops)
4.5k @ 16 - bit RAW to 12-bit linear
Sensor Size: 36.70 x 25.54 mm
Photosite Pitch: 8.25 μm

ALEXA Mini +/- $40-50,000
+/- 14 DR (Usable Stops)
3424 x 2202 @ 16 - bit RAW to 12-bit linear
Sensor Size: 23.8 x 17.8 mm
Photosite Pitch: 8.1 μm

Canon C300iii $11,000
+/- 13.5 DR (Usable Stops)
6k @ 12-bit RAW
Sensor Size: 26.2 x 13.8 mm
Photosite Pitch: 6.40 μm

RED Gemini +/- $30,000
+/- 13.5 DR (Usable Stops)
5k @ 16-bit RAW
Sensor Size: 30.72 x 18mm
Photosite Pitch: 6.0 μm

Sony Venice +/- $50,000
+/- 14 DR (Usable Stops)
6k @ 16-bit RAW
Sensor Size: 36.2 x 24.1 mm
Photosite Pitch: 6.0 μm

Red Monstro +/- $70,000
+/- 13.5 DR (Usable Stops)
8k @ 16-bit RAW
Sensor Size: 40.96 x 21.60 mm
Photosite Pitch: 5.0 μm

RED Komodo +/- $10,000
+/- 12.5 DR (Usable Stops)
6k @ 16-bit RAW
Sensor Size: 27.03 mm x 14.26 mm
Photosite Pitch: 4.40 μm

Canon R5 $5000
+/- 12 DR (Usable Stops)
8k @ 10-bit / 12-bit RAW
Sensor Size: 36 x 24 mm
Photosite Pitch: 4.38 μm

BMP6K $2,000
+/- 12 DR (Usable Stops)
6k @ 12-bit RAW
Sensor Size: 23.1 x 12.99 mm
Photosite Pitch: 3.76 μm

RED Helium +/- $40,000
+/- 13 DR (Usable Stops)
8k @ 16-bit RAW
Sensor Size: 29.90 mm × 15.77mm
Photosite Pitch: 3.65 μm

BMUMP12K $10,000
+/- 13 DR (Usable Stops)
12k @ 12-bit RAW
Sensor Size: 27.03mm x 14.25mm
Photosite Pitch: 2.20 μm

 

[Jason beaumont]

if you buy a gemini, you can use it for atleast 5 more years no problem. videos from 5 years ago still look great. the fact better cameras have came out doesnt really matter

 

[paulcurtis]

What interest me the most for the future of RED cameras would be a 6-8k FF global shutter sensor in a Komodo form factor but with better HFR capabilities and larger photo-sites.

Sure, we don't know what DSMC3 will bring and Red does have a higher tiered range.

I disagree with the photosite size, on paper it makes sense but there's more to a photo site than it's size (construction, micro lenses, backside illumination, where electronics go and so on). We don't know how the global shutter in Komodo is made and what the compromises are. We do know that originally it was said that GS was going to be an option because of reduced range but actually it turned out better.

I also have no idea where your Usable Stops come from, because my direct experience doesn't match all those (of the cameras i've used)

cheers
Paul

 

[James Sielaff]

Sure, we don't know what DSMC3 will bring and Red does have a higher tiered range.

I disagree with the photosite size, on paper it makes sense but there's more to a photo site than it's size (construction, micro lenses, backside illumination, where electronics go and so on). We don't know how the global shutter in Komodo is made and what the compromises are. We do know that originally it was said that GS was going to be an option because of reduced range but actually it turned out better.

I also have no idea where your Usable Stops come from, because my direct experience doesn't match all those (of the cameras i've used)

cheers
Paul

The larger pixel argument is a "all things being equal" argument, and of course, things are never all equal.

 

[Andrew Reese]

Outside of bit-depth, sensor size, and internal processing there is absolutely a physical relationship between the surface area of photo sites and light sensitivity. A good analogy is droplets of water in a "bucket". A wider diameter bucker will simply capture more rain drops (photons) and have a greater "full well capacity".

The best low-light camera available is the Canon ME20F-SH which literally can see in the dark. It has a pixel pitch of 19µm. Coincidence? I think not... The Sony A7sii (an industry standard low-light camera for 5+ years) has a pixel pitch of 8.4 µm but does have some internal noise reduction under the hood.

Alexa Mini and Mini LF are over 8.0µm. Now they aren't the best low-light cameras but alternatively can produce amazing range in the highlights (probably best in the industry) with very smooth roll-off to clip.

Gemini is 6.0µm, Komodo is 4.4µm and Helium is 3.65µm. Helium is not the greatest RED camera in low-light conditions or under-exposure but the argument is that the image cleans up when down-sampling from 8k to a 4k. What if you wanted a perfectly clean 8K image in lower light situations with Helium, how would you change that sensor to make it more sensitive?

From my initial experience Komodo is little better than Helium in lower-light scenes but Komodo still does not come close to Gemini in LL mode. I have found Gemini to be much better than Helium, Dragon and Komodo in terms of cleaner low-light images but Gemini has the largest photo sites, slightly larger sensor and dual-iso calibration to its advantage.

My original point was, there is definitely something to be said about the surface area of photo sites and overall sensitivity. How the photo-sites are "tuned" and what goes on in terms of processing under the hood also comes into play. This is where I think Arri leans more towards favoring highlight retention than low-light performance in how they tune their sensors. Whereas the Canon ME20F-SH is obviously tuned to be most sensitive in low-light scenes at the cost of highlight retention and loss of overall DR.

With smaller surface area photo-sites there is physically less photon capture and without a pixel pitch of 19µm the Canon ME20F-SH simple would not perform the way it does in low-light conditions no matter what processing and noise reduction is activated under the hood.

The same fundamental principle is encountered with sensor size and lens coverage (i.e. S16 < M4/3 < S35 < Full Frame) with full-frame allowing for better sensitivity in low-light conditions simply by letting in more light. This is why Monstro is somewhat on par with Gemini in low-light even though it has a slightly lower pixel pitch (5.0µm), no dual-iso, but instead a larger VV sensor.

I agree it is the combination of how all specifications are "playing together" but what is going on behind the scenes in terms of internal processing will never make up for a sensor that physically is not large enough or does not have large enough photo-sites to capture more photons.