I am a professional colorist & would like to try improving your raw footage

[Pritchboxer]

The project is great but basically as a professional, I've seen nothing impressive from this camera until recently, with the raw update it looks almost usable in a professional video. The bird footage looks very weak, artifact-ed, noisey and just bad coloring but I'd like to get my hands on some raw data to see if thats down to hardware or poor post processing which I suspect.

If I can I will share my workflow. If someone could be so kind to either film under studio lighting or failing that, get me some static shots of a subject on a cloudy day, so the shadows (dark areas) are fairly well exposed and no clipped highlights. Just a quick few seconds worth of frames ie 100 or so. If I could get the raw image data and some dng's even better and I will try a variety of debayering methods.

I dont need anything fancy, just some movement in the shot such as a falling coin and well exposed.

Thanks

[Nikon1]

Nice to have you here.
I would realy like to see someone push this Camera and its Footage to its maximum Potential.
Still dont have a chonos myself ... Mainly because of money and because i still havent got my hands on one to test it.
I also feel like there is a need for somebody to come up with some realy impressive Footage filmed on the Chonos, because i always feel like this camera could (or at least should) be able to do somewhat better, than what i have seen untill today.
I am more of an camera guy myself and also into compositing and animation, but sometimes i also do grading and color work and i also found most of the footage i´ve seen rather weak in terms of color/ contrast and dynamic range.
Would love to see if it is just a thing about grading or debayering.
Most people here seem to use the 12,5-75mm f/1.2 c-Mount "Kit-Lens". Maybe it is just the look of that lens that looks somewhat wrong. I have a similar lens like that one, just mine is older, and that is a realy soft lens with poor contrast and an realy weird look to it.
I dont know if the newer Lens here has the same Look, but i never use mine because it is that bad....
As an kit lens it is quite ok, especialy for the price and size and for the fact, that you could go to f/1.2 if you need to, but it gets very soft wide open and isnt even super sharp when at f/5.6 or less...
Whish you best luck in finding some Raw Footage and would recommend to try get something taken with a lens of known performance, because that could be a huge factor for image Quality.

[BiduleOhm]

The main problem is in the debayering method but the last firmware update (0.3.0) has greatly improved it (but 99.9 % of the videos on the internet are videos from before the update was available).

The raw data is very good since the beginning: http://forum.krontech.ca/index.php?topic=132.msg1115#msg1115

[Nikon1]

As an professinoal i think you allready know that, but i post those images anyways for everyone else reading along.
Here is an quick side by side comparison of three lenses. All of them are realy sharp.
.
The Leica on the Left is the oldest of them and an very early one, completely uncoated.
The Schneider in the midle is somewhat newer but still quite old, Glass has coatings.
The Nikon is a Modern lens design, realy sharp and Contrasty, coated.
.
Images Flat, Graded and an 1:1 Crop for sharpness. All taken on the Nikon1 J5, 1" Sensor Size. Grading is just done to my personal taste for every shot individualy, to see what i could make out of every shot.
That is just to show how much difference in terms of the Look a lens could make, even if sharpness would be basicaly the same. Those test shots where Taken in realy heavy backlight, a quite demanding scene for older and uncoated lenses.
For some things like Cinema or Music videos those Older Lenses could produce some Desireable results, but for science ... i would rather use other, newer lenses...
.
So if virtually everyone uses the chronos with that Kit Lens, the look of that lens could be mistaken easily with the look of the Camera itself just by looking on the footage on the Web...
Still waiting for someone with proper Lenses and some good knowledge about grading and Lighting to make an testshoot with the chronos.

[NiNeff]

I did as you requested and filmed a coin drop.
Lighting: direct sunlight, about 5 m distance to camera, fstop about 9
https://youtu.be/J8meUUESwJk
The raw will take quite a while to upload, i will edit as soon as it's up.

EDIT:
here's the original mp4: https://nineff.de/nextcloud/index.php/s/tJzYHCr4gZSRatZ
and here's the raw: https://nineff.de/nextcloud/index.php/s/tmpRbK9TefSceM6

[Pritchboxer]

Thanks for your responses guys. Thank you so much NiNeff, I will take a look at the raw data but I'm afraid it's too blurry to begin with. Maybe its the lense you're using? I'd suggest stopping down the aperture as far as you can go without having to push the iso above the lowest setting to improve sharpness. I do worry the jagged edges and noise we see is to do with the 2x2 binning method I believe this sensor uses which is likely gonna create jagged edges. If someone could do a test with a known sharp lense such as the nikkor, stopped down a bit to increase sharpness, it would be perfect.

[Nikon1]

As far as i know, the sensor does not use bining but is just 1280x1024px, but in a bayer-Pattern.
If you look up "Bayer-Pattern" (or whatever the exact name of that Article is) on Wikipedia, you will find, that even good algorythms will only give you about 70 to 80% of effective Resolution of the original bayer-Pattern, even with opimal algorithm for your Scene...
So, if you take that 1280x1024 resolution (which is allready quite Low, especially by todays standards with the ongoing megapixel war of big manufacturers) and multiply it by 70%, you will get around 1 Megapixel effective Resolution (in terms of Sharpness and Ability to resolve Details, renderd by the lens on the Sensor).
IF you use the full Sensor Resolutin, that is...
If you are going for 16:9 and use 720p Resolution, it is even Worse effective Resolution.
So, i dont think it is especially blurry or something, it is just "Low Resolution". But Given The Framerate it is still quite Respectable...
So the "kit Lens" is actually almost ok for that camera and most people, since the Pixels are that big and because you wont even see the unsharp lens, unless you are Pixel-Peeping like me and you maybe also. But Color, glare, flares and contrast are somewhat different, since it will make the final image look wrong in some Way.
Just for your info: i have very slow internet connection and because of that i watch almost everything in 240p or 360p HD Resolution. I still could tell immedeately what shot in a video was filmed on the chronos and which one was filmed on some other Camera...
.
On my 20 Megapixel 1" sensor i would never want to use a lens with low Resolution, because it would be way to blurry for me.
.
That Nikon1 32mm Lens (and all my other Nikon1 Lenses...) CAN be mounted on the Chronos!
I allready built an adapter for that, but i never got my hands on a chronos to do a comparison against my Nikon J5 in the Same Light with the same Lens and same Settings.
I would not wonder at all, if the Nikkor Lens would be WAY better than the Kit Lens, since it costs arround 800€ or even more when new, and it is built for extreme High-Resolution 1"-Sensors, so it should also perform extremely well on the Chronos also.

[Pritchboxer]

The sensor is advertised as doing 2x2 binning... maybe it's just an option but would make sense considering the jagged edges. There's certainly more bad things going on than just low resolution. Also I think maybe what you're referring to is the way an image is debayered depending on the bitrate. A 14 bit raw video should retain all the details providing there are enough photosites. 8 bit video is where some of the pixels are guessed effectively, based on key pixels.

[Nikon1]

you could do 2x2 Bining on the chronos, thats right, but that is just for increasing the FOV for very high Framerates.
Think about it:
If you have a realy small resolution it is just croping the Sensor to a realy small size, so you would need realy crazy wide Lenses like around 1 or 2mm focal length in some cases if you just use 1:1 pixels because of that small sensor size to beginn with and the aditional crop you would apply.
2x2 bining is in that case just to make life a bit easyer, that you could get along with fewer, less exotic lenses.
If you would use 2x2 bining on the full sensor it would give you 640x512px total resolution from my understanding. That would allow for faster Framerates while still being able to get the "look" of the 2/3" sensor (like DoF-Effects and so on), just lower Resolution.
Native Sensor resolution should be 1280x1024px

[Pritchboxer]

I'm not saying I want to do it, I'm saying on the sensor's tech page they say it does 2x2 binning. Maybe this is how it achieves such high frame rates but at the expense of reduced true resolution like you say.

Hopefully thats not correct and it's simply bad post processing. I'd love to get a hold of the recent raw bird footage. That looked almost ok... I think if I can get my hands on it potentially I can make it clean 720p. If it is down to the sensor, I'm not sure why they havent upgraded to the newer 1080p sensor available from the same manufacturer. The markup between the cost of the sensors and selling price is huge already.

[Nikon1]

https://en.wikipedia.org/wiki/Demosaicing
.
So, there is what i was saying:
Quote from that Wiki-Article above:
.
"The reconstructed image is typically accurate in uniform-colored areas, but has a loss of resolution (detail and sharpness) and has edge artifacts (for example, the edges of letters have visible color fringes and some roughness). "
.
You will always loose Resolution with debayering, since You never know what exact color your pixel realy is, you kind of always have to estimate and interpolate, and that will introduce blurryness.
I dont know where that 70 to 80% effective Resolution fact came from, but i read it somewhere.... it is somewhere around that.
That is kind of the reason why almost any modern digital camera has a sensor of higher resolution than what it will output for video-Footage. There Will always (/mostly) be in-Camera downscaling to, for example 1080pHD from a 14 Megapixel sensor.
That way you realy know what color and brightness your final Pixel should realy be.
But for Highspeed-Cameras it is mostly Native sensor Resolution = Footage Resolution because of computing power limits and Data Speed Limits on Memory. And then, additionally to the quite low resolution of those Sensors, you have to deal with the Demosaicing...
.
And about that 1080p thing you said:
Problem is, you need way more processing power for that kind of resolution at this framerate. And that could get expensive quickly.
Also, in todays Cameras the whole camera is built around the sensor, so you could not just upgrade the sensor that easily... you would have to change a lot about your camera if you wanted to do so.
If you want a better sensor in that kind of camera, the only way to do it, is basicaly built a completely new camera around your new sensor and use, what you have learned from the previous one... and maybe reuse most of the software and maybe also the I/O-Part of the Camera aswell as the buttons and stuff.
I realy hope the Kron Guys are working on an follow up model with higher resolution sensor. I allready asked David about that but never got an answer about that.

[Nikon1]

test?
did you miss to add an atachment?

[NiNeff]

I uploaded another set including raw files in the footage thread.
Also there's another coin video uploading right now

[NiNeff]

More Coins:
https://nineff.de/nextcloud/index.php/s/SfWB3kt3fe4aMgn
Raw:
https://nineff.de/nextcloud/index.php/s/pMWDi2eRqScxp3z

[patrickrebstock]

i could try and shoot some for you tomorrow,
i have shot alot but not any raw yet