Copper Bones - RED Monstro - Low Light Test

[AndreasOberg]

Thanks Peter. Dynamic Range stuff coming today. Yesterday ran about 6 hours longer than it should have.




ISO is fairly subjective to taste when it comes to just how you utilize it for general filming. I know of two shows being shot at ISO 2000 right now no Helium. Low light is low light, but if you have the freedom and power over your exposure you can sort of work with it how you like. ISO 250, 800, etc.... Just whatever produces the best image for you. The only real thing is clipping and how you feel about it and image noise and how you feel about it. Thus far, even though my journey with Monstro is relatively new I've delivered ISO 250-3200 material that I enjoy.




Interesting you say that about the web compression! There's actually a rather specific IRE range that YouTube and Vimeo encodes rather dislike in terms of compression. It's in the lower shadows, which is why lifted shadow material encodes slightly better for their respective codecs. It's not a huge deal, but you can certainly see the difference on the ProRes or higher bitrate encode. Thank you for the compliment, I do spend a bit of time on my web encodes though. It's a moving target to upload "the best" material and there's of course file size limitations that truly prevent you from going crazy.

I like Monstro for wildlife, you are wider however on all your lenses, so be aware of that if shooting 8K FF using the full VV sensor. That said many who shoot wildlife will explore 5K-6K to get that S35 range. It's pretty common to use longer still lenses for wildlife for sure, pulling focus is of course harder, but the optics are so very good. Honestly some of Canon's best work optically has been in the 200, 300, 400, 500, and 600mm primes. I have a few long lens solutions like RED's 300mm T2.9 and more recently I enjoy shooting with Sigma's 120-300 for outdoors stuff, great freedom of zoom there. You are correct that the 50-1000 does not cover 8K FF, but down at 5K-6K you'll be having a good day. Personally I have 12-600mm solutions for VistaVision at the moment and in those extreme situations a cheater/doubler/tele-extender are a good idea. Expanders actually work well here too. I did something neat with the Tokina 1.6X expander late last year and the Zeiss 70-200mm T2.9.


Doing the typey typey, then lunch, then uploading some Monstro Dynamic Range Tests and Measurements. I'll drop it here in the Weapon sub-forum.

Fantastic job Phil. You are an inspiration!
I agree I see that youtube is destroying the dark parts, but I'm using my psychic mind to see what it would look like without the compression. Still the quality is fantastic, especially in the brighter parts.
I think that in our case if we had the dream monstro sensor we would probably shot much in 5k. That would give us the 96 fps that we usually crave. As a bonus the crop factor would make our zoom even more brutal. How do you think a 5k monstro crop would compare to an Epic MX 5.1k? I would think monstro would still look better and have less noise.

Do think there is a change that Red will ever put in a burst slow motion mode that lets us shot in higher resolution for a few seconds. Even 3 seconds in 200fps in 4k would be a big step forward.

Btw, the Canon IS makes wonders. I can even film handheld by just leaning it towards something and get good results. Yes I need to stabilise, but the movement is is soft so its easier to get steady in post.

/Andreas

 

[AndreasOberg]

Show the folks at work and they were all very impressed. Such a pleasant image.
/Andreas

 

[Phil Holland]

Show the folks at work and they were all very impressed. Such a pleasant image.
/Andreas

Excellent to hear!

Side note this particular look from philmColor was created for a Director who as of yet only truly works with film, so the pressure is on Got something fun coming hopefully before NAB to show off on that subject matter.

Fantastic job Phil. You are an inspiration!
I agree I see that youtube is destroying the dark parts, but I'm using my psychic mind to see what it would look like without the compression. Still the quality is fantastic, especially in the brighter parts.
I think that in our case if we had the dream monstro sensor we would probably shot much in 5k. That would give us the 96 fps that we usually crave. As a bonus the crop factor would make our zoom even more brutal. How do you think a 5k monstro crop would compare to an Epic MX 5.1k? I would think monstro would still look better and have less noise.

Do think there is a change that Red will ever put in a burst slow motion mode that lets us shot in higher resolution for a few seconds. Even 3 seconds in 200fps in 4k would be a big step forward.

Btw, the Canon IS makes wonders. I can even film handheld by just leaning it towards something and get good results. Yes I need to stabilise, but the movement is is soft so its easier to get steady in post.

Web compression is a strange world, though a good one in terms of empowering various levels of content creators. The nasty bit being the fact that we must encode and they must re-encode. If I was "king of the internet" I would work closely with various companies that build popular encoders and include passthrough metadata for "to spec" encodes for YouTube and Vimeo so it would just pass through. This would be ideal for professionals producing content and while it's available to those very deeply entrenched in the gullies, to us meager professionals we are left to experiement to find the best results which most of the time entails uploading a very large and high quality file to get a decent serverside re-encode, which of course limits how much and how long our content is compared to encoding it to the way it should be seen from the get go.

I can say safely that Monstro at 5K will be better than MX in a variety of ways. Beyond the noise, you have more color information and less evil IR contamination to battle against.

I can see RED introducing a burst concept with a new body perhaps. At the moment we are tied to the sensor's max readout speeds. Jarred knows my biggest wishes on where we go from here hardware-wise. One of those things is 8K at 120-240fps. Which would provide the ecosytem for a few things like 5K at 192-385 fps (4K 240-480fps), which would be a major new world for motion picture technology if that can be done in a format size like VistaVision. That would require new hardware, new media, and a new sensor design. I can see a few ways to get there tech-wise. Whether they focus on burst or susttained frame rates, I'm certain when we go beyond DSMC2 brains in a while we'll have some major powerful hardware at our disposal. Interestingly much of the ecosystem to pull this off is there MINUS the sensor tech that I know of at the moment. To that note, it's been interesting to see the tick/tock nature being slightly out of whack with new sensor tech being explored. Which I think I think is a good thing, especially in the face of the competition coming to the 8K arena. The biggest difficulties of doing anything like this is data rates, heat, and QC with the focus of avoiding a negative effect on color and dynamic range. RED could make a camera now with about 10-ish stops of DR that likely could do all of this if I've done my math right, but nobody working today wants 10 stops if they can avoid it.

Optics stabilization was a good call long ago. It's been interesting to see Sony's efforts on the stabilized sensor. Outside of cellphones and perhaps camcorders I'm not a massive fan of digital stabilization. Of course gimbals and similar tech are useful when going further into smooth. I wonder often how much of this can manifest into a professional camera system, but I suspect between the rapid improvement of more optimized stabilized setups we'll have access to something spectacular soon. Outside of the Shotover F1 and K1 I use for helicopter work, there's recently been a couple compact systems I can use on good old ground that have attracted my interests.

 

[AndreasOberg]

Excellent to hear!

I can say safely that Monstro at 5K will be better than MX in a variety of ways. Beyond the noise, you have more color information and less evil IR contamination to battle against.

I can see RED introducing a burst concept with a new body perhaps. At the moment we are tied to the sensor's max readout speeds. Jarred knows my biggest wishes on where we go from here hardware-wise. One of those things is 8K at 120-240fps. Which would provide the ecosytem for a few things like 5K at 192-385 fps (4K 240-480fps), which would be a major new world for motion picture technology if that can be done in a format size like VistaVision. That would require new hardware, new media, and a new sensor design. I can see a few ways to get there tech-wise. Whether they focus on burst or susttained frame rates, I'm certain when we go beyond DSMC2 brains in a while we'll have some major powerful hardware at our disposal. Interestingly much of the ecosystem to pull this off is there MINUS the sensor tech that I know of at the moment. To that note, it's been interesting to see the tick/tock nature being slightly out of whack with new sensor tech being explored. Which I think I think is a good thing, especially in the face of the competition coming to the 8K arena. The biggest difficulties of doing anything like this is data rates, heat, and QC with the focus of avoiding a negative effect on color and dynamic range. RED could make a camera now with about 10-ish stops of DR that likely could do all of this if I've done my math right, but nobody working today wants 10 stops if they can avoid it.

That is interesting to know. For us slow motion is so important so its a bit hard to know how useful 60fps at 8k would be. 60fps is still miles better than 30fps in my book.

I did a bit of research and since the canon sensors have 36mm wide sensor and monstro has 41mm this means that you could definitely use the 600mm F4 up to 7.2k resolution and stay withing 36mm image circle. However since Monstro is not as high as the Canon and the problems usually are visible in the corners I would not be surprised if you could go outside the 36mm a bit. Also the 600 has such a good image quality with almost no defects at the corners which indicates that the image circle is quite big. How do you actually measure the image circle? With a bright light and the exact correct distance (same as sensor would be) to a paper I would think?

/Andreas

 

[Les Hillis]

Hi Phil, nice low-light test, thanks for sharing that.

Curious if you've considered designing a standardized exposure test setup (or group of setups) for motion picture shooting.

With calibrated spot, reflective and colour meters, wouldn't it be possible to select widely available (or makeable) objects, materials and surfaces that could be arranged (and lit and made to move) in a pre-determined way to enable reproducible testing of various cameras and image-forming elements?

 

[AndreasOberg]

I did some more research about lenses. From what I understand it will be tricky to use many long lenses with such a large sensor. Canons 50-1000mm for example only has an image circle of 32mm since its intended for S35 so that will be impossible to use. I love the low light capabilities of Monstro but sounds like Helium may be a better choice to be able to use long cinema zooms.

/Andreas

 

[Karim D. Ghantous]

Phil, Nice work and thank you again as always for giving back to this community

That's Positive Phil!

And here's a shot straight from camera:


And with the philmColor baseStock color grade:

Don't hate me but I much prefer the ungraded version. The graded version has too much life taken away from it.

It's been interesting to see Sony's efforts on the stabilized sensor. Outside of cellphones and perhaps camcorders I'm not a massive fan of digital stabilization.

From what I hear, Olympus's implementation of IBIS is so good, you can take handheld photos with surprisingly long shutter speeds. We're talking more than half a second, apparently. I forgot exactly what I read, but it was pretty impressive.

 

[Patrick Kendall]

Is this Chester Copperpot before he sets out on his adventure of finding the pirate ship, pre-Goonies?