Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

KOMODO....

Status
Not open for further replies.
David Rasberry said:
Two new Canon RF lenses just announced
Click to expand...

Yep, that makes up the rather popular holy trinity of Canon zoom focal lengths. Coming up next is the RF 70-200mm f/2.8L, which is apparently significantly shorter than the EF counterparts.

I personally wouldn't have minded them going all f/2 for a trio to accompany the RF 28-70mm f/2 L. But that would have been a bulkier lens for sure.

But as they say, since these cameras are getting better and better in low light, perhaps not a huge deal.

I will say that I liked the blip about the newer 24-70 breathing less. I did a really deep dive into their new focus movement tech last night. Very interesting stuff.
 
Phil Holland said:
(snip) I will say that I liked the blip about the newer 24-70 breathing less. I did a really deep dive into their new focus movement tech last night. Very interesting stuff.
Click to expand...

I know they went hard core on a trolley system in the CN 30-300mm to control breathing. Curious if Canon has been able to use similar techniques on other lenses in their portfolio. FWIW, if I were Canon, I'd make all my newer lenses motion capture friendly. RED may have jump started the wave with their DSMC program, but at this point...

Cheers - #19
 
Bastien Tribalat said:
Seems BRAW is limited to 12 bit anyway. Plus, it seems there's de-mosaic in-camera process that impose that limit. It's beyond my technical skills and it doesn't change the fact that BMD Log is not as flat and therefore has not as much lattitude than others. Even Canon RAW Light, which is also 12 bit (or 10 bit even beyond 30p on C200) has a more usable log gamma curve.

But hey at the end of the day none of this matters because the DP is supposed to do his work properly and light a scene correctly. (and I like the images coming from URSA Mini Pro G2 and Pocket 4K so far but I'm just saying that it's not as "good" what RED, Sony, Canon, Arri and Panasonic offer and it shoudn't be : those cameras are hell cheap and come with DaVinci Resolve. It's an AWESOME package for low budget projects and filmmakers, students and beginners. BUT it's not as "good" as an Alexa Mini LF 10 times less money)
Click to expand...

Blackmagic log curves are less flat BUT I believe they store information outside of the 0-1 range of their log curves, giving you the full DR while still having a less flat log curve. I think they chose that because people complained their log curves where too flat.

Especially the Blackmagic Raw files are more than a good replacement for 10-bit prores because of the better color subsampling, so I don't see any reason other than for maybe network TV that you wouldn't shoot BRAW.
 
Bastien Tribalat said:
Seems BRAW is limited to 12 bit anyway. Plus, it seems there's de-mosaic in-camera process that impose that limit. It's beyond my technical skills and it doesn't change the fact that BMD Log is not as flat and therefore has not as much lattitude than others. Even Canon RAW Light, which is also 12 bit (or 10 bit even beyond 30p on C200) has a more usable log gamma curve.

But hey at the end of the day none of this matters because the DP is supposed to do his work properly and light a scene correctly. (and I like the images coming from URSA Mini Pro G2 and Pocket 4K so far but I'm just saying that it's not as "good" what RED, Sony, Canon, Arri and Panasonic offer and it shoudn't be : those cameras are hell cheap and come with DaVinci Resolve. It's an AWESOME package for low budget projects and filmmakers, students and beginners. BUT it's not as "good" as an Alexa Mini LF 10 times less money)
Click to expand...

It's 16bits logged to 12 bits, same as all other camera RAW outputs (except X-OCN which is 16 bits linear according to Sony).
You can download some UMP clips here https://www.blackmagicdesign.com/products/blackmagicursaminipro/blackmagicraw. The lattitude of the UMP G1 is tested by Cinema5D and comes out on par with the AMIRA and 0.5 stops behind the ALEXA on a Full-HD/2K rgb-output(full sensor readout).
The canon C300 II and the FS7 are about 1.5 stops behind the ALEXA (full sensor readout) on a Full-HD/2K rgb-output.

ARRI ALEXA's, RED 5k+ and Sony Venice have more lattitude than the UMP, Canon's and Panasonics don't at the moment.
 
Phil Holland said:
Yep, that makes up the rather popular holy trinity of Canon zoom focal lengths. Coming up next is the RF 70-200mm f/2.8L, which is apparently significantly shorter than the EF counterparts.

I personally wouldn't have minded them going all f/2 for a trio to accompany the RF 28-70mm f/2 L. But that would have been a bulkier lens for sure.

But as they say, since these cameras are getting better and better in low light, perhaps not a huge deal.

I will say that I liked the blip about the newer 24-70 breathing less. I did a really deep dive into their new focus movement tech last night. Very interesting stuff.
Click to expand...

It will be interesting to see how Komodo plays with the electronic lens controls on these. AF? OIS? Manual control?
 
Misha Engel said:
The lattitude of the UMP G1 is tested by Cinema5D and comes out on par with the AMIRA and 0.5 stops behind the ALEXA on a Full-HD/2K rgb-output(full sensor readout).
Click to expand...

Complete nonsense.

Resolution factor on Alexa doesn't change anything in that context.


Bastien Tribalat said:
Even Canon RAW Light, which is also 12 bit (or 10 bit even beyond 30p on C200) has a more usable log gamma curve.

But hey at the end of the day none of this matters because the DP is supposed to do his work properly and light a scene correctly.
Click to expand...

It matters the next day.
 
Cole Sprague said:
+1 for Monochrome!
Click to expand...
RED has offered monochrome cameras for years now, and they deserve credit for that. "Change it in post" misses the whole point, which I don't need to explain to anyone here. :-)

Christoffer Glans said:
I would also add that far too many aren't good enough cinematographers to be able to utilize everything on a Red or Arri and far too many have a Red's or Arri's with far more capabilities than they make use of in their work.
Click to expand...
I agree with that, though the RED is the real powerhouse - it's like the digital equivalent of a film negative, when you think about it.

It makes me wonder why beginners go into debt to buy these high end cameras. There is no point to this extravagance - and believe me I love my cameras and lenses, just like most people. But there is no shame in sensibility, especially when you are starting out.
 
I’m not a big fan of really flat log profiles. It is a form of compression. If you are shooting a video codec like Prores you need them sometimes, but you give up a lot of subtle rich color and image texture in deep shadows that you can grade into display range from a linear raw image. That is just what happens when typically 4 stops of linear values get compressed into less than one at the bottom of a log curve. You can restore contrast , but you can’t recover the lost data.
 
Given that most of us make money off of our REDs, it's not debt, it's an investment. Why delay that profit by dropping smaller $5k/$10K/$15K debt bumps on un-rentable cameras along the way? If you're gonna get wet, might as well go swimming.
 
Hrvoje Simic said:
Complete nonsense.

Resolution factor on Alexa doesn't change anything in that context.
Click to expand...


What happens by downscaling and how is the dynamic range increased?

Well, the signal value of four neighboring pixels has a high degree of correlation (luminance will be roughly the same), whereas noise for those pixels in general should show no correlation as it is random.

Now, by downscaling (or averaging) four pixels of an 3840×2160 image to 1 pixel of an 1920×1080 image the signal value should stay very much the same, but the noise value will reduce because noise is random (uncorrelated), hence averaging (or in mathematical terms the root mean square value of noise which is measured in the recorded image) scales with the inverse square root of the number of samples that we average.

Hence 4 pixels into 1 gives a factor of two (sqrt(4)=2) higher signal to noise (SNR) ratio. Higher signal to noise ratio for a given luminance = higher dynamic range.

How is the LUT business these days?
 
Misha Engel said:
What happens by downscaling and how is the dynamic range increased?

Well, the signal value of four neighboring pixels has a high degree of correlation (luminance will be roughly the same), whereas noise for those pixels in general should show no correlation as it is random.

Now, by downscaling (or averaging) four pixels of an 3840×2160 image to 1 pixel of an 1920×1080 image the signal value should stay very much the same, but the noise value will reduce because noise is random (uncorrelated), hence averaging (or in mathematical terms the root mean square value of noise which is measured in the recorded image) scales with the inverse square root of the number of samples that we average.

Hence 4 pixels into 1 gives a factor of two (sqrt(4)=2) higher signal to noise (SNR) ratio. Higher signal to noise ratio for a given luminance = higher dynamic range.

How is the LUT business these days?
Click to expand...

This is also why the S1H gets 4:2:0 at 6K and 4:2:2 at 4K subsampling. Same goes for Red, having around 4:2:2 at 6K (Dragon) and 4:4:4 at 4K output. Now, my math might be off, but in general principle.
 
Misha Engel said:
What happens by downscaling and how is the dynamic range increased?

Well, the signal value of four neighboring pixels has a high degree of correlation (luminance will be roughly the same), whereas noise for those pixels in general should show no correlation as it is random.

Now, by downscaling (or averaging) four pixels of an 3840×2160 image to 1 pixel of an 1920×1080 image the signal value should stay very much the same, but the noise value will reduce because noise is random (uncorrelated), hence averaging (or in mathematical terms the root mean square value of noise which is measured in the recorded image) scales with the inverse square root of the number of samples that we average.

Hence 4 pixels into 1 gives a factor of two (sqrt(4)=2) higher signal to noise (SNR) ratio. Higher signal to noise ratio for a given luminance = higher dynamic range.
Click to expand...


No, that's not dynamic range, no it doesn't get increased whatsoever by scaling down and it is still nonsense.
 
Christoffer Glans said:
This is also why the S1H gets 4:2:0 at 6K and 4:2:2 at 4K subsampling. Same goes for Red, having around 4:2:2 at 6K (Dragon) and 4:4:4 at 4K output. Now, my math might be off, but in general principle.
Click to expand...

The myth that Bayer raw spatial color filter distribution produces chroma resolution equivalent to video codec subsampling dies hard, but it is definitely not true for any decent modern deBayer algorithm and uncompressed Bayer raw data. Raw compression schemes no doubt have some effect on that though.
 
Gabriel Mann said:
Given that most of us make money off of our REDs, it's not debt, it's an investment. Why delay that profit by dropping smaller $5k/$10K/$15K debt bumps on un-rentable cameras along the way? If you're gonna get wet, might as well go swimming.
Click to expand...

But you're not making money by possessing your camera; you're making money with the work you do, which could conceivably be just as good using any number of cameras. I think cameras are closer to expensive tools than investments. Investments should have the goal of increasing in value for profit, which digital cinema cameras as a rule don't do. Considering how quickly they become obsolete and then no longer hold value, they're closer to a computer (tool) than, say, real estate (investment).

Am I taking you too literally?
 
Fahnon Bennett said:
But you're not making money by possessing your camera; you're making money with the work you do, which could conceivably be just as good using any number of cameras. I think cameras are closer to expensive tools than investments. Investments should have the goal of increasing in value for profit, which digital cinema cameras as a rule don't do. Considering how quickly they become obsolete and then no longer hold value, they're closer to a computer (tool) than, say, real estate (investment).

Am I taking you too literally?
Click to expand...

If a particular camera choice increases the profit from your work either via higher perceived quality and market value of the end product or more efficient production workflow then yes it qualifies as an investment.
 
David Rasberry said:
I’m not a big fan of really flat log profiles. It is a form of compression. If you are shooting a video codec like Prores you need them sometimes, but you give up a lot of subtle rich color and image texture in deep shadows that you can grade into display range from a linear raw image. That is just what happens when typically 4 stops of linear values get compressed into less than one at the bottom of a log curve. You can restore contrast , but you can’t recover the lost data.
Click to expand...

You are confusing two things. One can define a log profile such as Log3G10 which has less than a 1-bit error compared to whatever linear value you want to encode. One can define a container format that only holds so many bits per pixel. If the log container only has 8 bits per channel vs. 16 bits per channel of linear data, yes there will be some compression. But it's also true that at the bottom end of the linear light curve, noise floor may well be above precision error, likely for the first 3-4 stops. If you want to add noise as a source of richness, starting with a log profile won't stop you from doing so.

I think it leaves a wrong impression to tar all log files with the same brush. A log profile that's invertible from linear and back within the parameters of the noise threshold is perfectly acceptable, and that in that case, any deviations between them can be approximated by adding noise.
 
Status
Not open for further replies.
Back
Top