It's official: Canon USA offers 5D2 firmware update

[Dylan Couper]

Originally Posted by Steve Witt

Thanks for the help Dylan. I have some manual focus Nikon lenses for 35mm film cameras that I am using with a Brevis adapter.....can I use them on this 5D MarkII camera??

With the help of a Nikon - EF adapter, yes. Do a quick search in this section and you'll come up with detailed threads on what is available. Many of us are using cheap $8 adapters off Ebay and are quite happy.

[Steve Witt]

Wow...It seems this could out do the Red Scarlet.

[Nigel Barker]

Originally Posted by Dylan Couper

Not to derail the thread, but I purchased an FD adapter as well (probably the same one) and was unimpressed by the results. Besides the softer image, the amount of light loss was very significant. 2 stops roughly. I have detailed notes & numbers, but never posted them since everyone was buying Nikons anyway.

Having done a bit more testing I would agree that the image is a bit softer Perhaps I didn't do such detailed testing as I honestly didn't notice 2 stops difference in exposure & I did stop down to F22 but it was a very bright & sunny day. The one FD lens I do have (35-105mm F3.5) has a reputation for being particularly sharp so maybe this offsets the softening of the adaptor to some extent.

The problem with FD->EF adapters (for those reading apart from Dylan) is this. When Canon moved from FD to EF they changed the focal point of the lenses & without adding a lens in the adaptor they cannot focus to infinity when mounted on an EOS. Given the price of the adaptors you can bet that the adaptor lens doesn't match the quality of the Canon lens so in general these adaptors have a poor reputation. There is also a increase in focal length of approx 1.25X.

Lenses of most every other type of mount can be fitted to the 5DII using a simple adaptor with no glass in it & thus no possibility of diminishing the optical quality. Given the potential for problems & disappointment with FD lenses on the 5DII & the fact there are so many other lenses that will work without problems I wouldn't recommend that anyone buy an FD lens & adapter.

[Jon Fairhurst]

Originally Posted by Steve Witt

Wow...It seems this could out do the Red Scarlet.

Yes and no. Comparing the 5D2 to the Scarlet 2/3" (the Scarlet S35 and FF35 are in a very different price range), here's a quick summary...

5D MkII
* Fantastic low light capabilities
* Full frame DOF - more than Super 35 film
* Support for many, many lenses; you can rent glass in any large city.
* Affordable and widely available

Scarlet 2/3" (expected, anyway)
* Blazing fast capture for slow motion (180 fps burst, I believe)
* 24/25p
* Superior audio support
* No rolling shutter artifact to speak of
* Better ergonomics for video
* Electronic zoom
* Continuous autofocus
* RAW video for smooth grading
* True 1080/2k resolution without aliasing (the 5D has moire artifacts)
* Longer shooting times

If you don't need the 35mm DOF, Scarlet has the potential to be superior in many ways. However, if you want 35mm and you want it now, the 5D MkII is king.

[Dan Brockett]

Originally Posted by Jon Fairhurst

Scarlet 2/3" (expected, anyway)
* No rolling shutter artifact to speak of

Jon:

Have you shot with the RED One? I have. Scarlet will have some rolling shutter artifacts because it will have a CMOS imager with a rolling shutter. It is just a matter of how apparent the rolling shutter artifacts will be.

Dan

[John Vincent]

The new firmware - something I've been quietly waiting for -is here. But, despite my reading of all the threads, I'm a bit confused -what does it all mean?

My understanding is that this new firmware essentially makes the video grabbing aspect easier - no more changing f stops and shutter rates - but alters nothing else.

If I bought a new, up-graded Mark II, what exactly can I do as a film maker? How close can I get to 24p? Does the firmware change the compression rates or ability to transfer the video data to editing systems? Does it (can it) address rolling shutter?

I understand the pros/cons of shooting a feature w/ a DSLR (ie - no eyepiece, strange form factor), but I'm not sure exactly how the new firmware would impact trying to shoot a feature.

Any ideas/thoughts would be greatly appreciated...

john

[Julian Frost]

The thing is, we don't know. All the press release says is that we can control the aperture, ISO and shutter speed, within the stated limits. It doesn't say if we can change any/all of those settings during filming, or if they can only be set prior to pressing the record button. The press release says nothing about 24fps, so I would assume, at this point, that the camera will be stuck at 30fps, as it is now. The same for compression rates... no change has been stated.

Julian

[Jon Fairhurst]

Originally Posted by Dan Brockett

Scarlet will have some rolling shutter artifacts because it will have a CMOS imager with a rolling shutter. It is just a matter of how apparent the rolling shutter artifacts will be.

I expect that the 2/3" Scarlet rolling shutter will be much less than the Red One, simply because of that 180 fps burst speed.

Clearly, to achieve that frame rate, Scarlet 2/3" must be capable of scanning from top to bottom in no more than 1/180th of a second (5.55ms.) By contrast, the 5D MkII scans in 25 ms (based on research by Mark Hahn, and confirmed by me.)

Anyway, in theory, Scarlet's rolling shutter should be about five times faster than the 5D2's - and better than the Red One's as well.

Of course, this is all theory. We will see when Scarlet actually hits the streets.

[Jon Fairhurst]

Originally Posted by John Vincent

My understanding is that this new firmware essentially makes the video grabbing aspect easier - no more changing f stops and shutter rates - but alters nothing else.

That's the working assumption. Rather than the camera setting its own aperture, shutter and ISO - and forgetting all the settings between shots - we should be able to set each of these variables and have them stay fixed from shot to shot.

In the end it will speed up production (no more messing with the camera to get the settings we want), it will reduce the chance of error due to simplicity, and it will help us ensure good matching from shot to shot.

It might also help us get a specific look - especially with Canon lenses - that wasn't available before. The automatic software would often jump between levels, and effectively made some settings impossible to achieve.

So... in theory, we will be able to 1) dial in exactly what we want, and 2) not have to adjust anything from one take to the next.

We will know more next week...

[Dan Brockett]

Originally Posted by Jon Fairhurst

I expect that the 2/3" Scarlet rolling shutter will be much less than the Red One, simply because of that 180 fps burst speed.

Clearly, to achieve that frame rate, Scarlet 2/3" must be capable of scanning from top to bottom in no more than 1/180th of a second (5.55ms.) By contrast, the 5D MkII scans in 25 ms (based on research by Mark Hahn, and confirmed by me.)

Anyway, in theory, Scarlet's rolling shutter should be about five times faster than the 5D2's - and better than the Red One's as well.

Of course, this is all theory. We will see when Scarlet actually hits the streets.

Your logic seems pretty sound, I am impressed. Of course, anything and everything we are discussing is really nothing but conjecture since RED is constantly changing and tinkering with the specs.

Dan

[Peer Landa]

Originally Posted by Dan Brockett

Your logic seems pretty sound, I am impressed. Of course, anything and everything we are discussing is really nothing but conjecture since RED is constantly changing and tinkering with the specs.

I agree, hence, I believe the Scarlet will look a bit different now after the 5D manual control dust has settled. (I also agree that Jon's logic, as always, is impeccable ;^)

-- peer

[Charles W. Hull]

Originally Posted by Jon Fairhurst

Clearly, to achieve that frame rate, Scarlet 2/3" must be capable of scanning from top to bottom in no more than 1/180th of a second (5.55ms.) By contrast, the 5D MkII scans in 25 ms (based on research by Mark Hahn, and confirmed by me.)

I'm sure you're right - but how then can the manual 5DII have shutter speeds up to 1/4000 sec?

[Evan Donn]

Originally Posted by Charles W. Hull

I'm sure you're right - but how then can the manual 5DII have shutter speeds up to 1/4000 sec?

Shutter speed and read out speed aren't connected. The degree of rolling shutter is a result of how long it takes to read the lines from the top to the bottom (read-reset speed), shutter speed only affects how long each line is exposed for. So each line can be exposed for 1/30th-1/4000 of a second before it is read, but it will always take 25ms to read out all of the lines on the sensor before resetting to the top.

[Jon Fairhurst]

Originally Posted by Evan Donn

Shutter speed and read out speed aren't connected. The degree of rolling shutter is a result of how long it takes to read the lines from the top to the bottom (read-reset speed), shutter speed only affects how long each line is exposed for. So each line can be exposed for 1/30th-1/4000 of a second before it is read, but it will always take 25ms to read out all of the lines on the sensor before resetting to the top.

Well said, Evan,

Some people are also confused as to how the camera can shoot close to a 1/30 exposure with a 30 fps frame rate and also have rolling shutter.

The answer is that the lines can expose independently as short or as long as they'd like - and they can expose simultaneously. The problem is that they are read out sequentially, and each read takes some amount of time, so the exposures are all offset in time.

In fact the first line can start exposing for the next frame while the last line is still exposing for the current frame, and has yet to be read. It's a bit of a brain bender, that!

[Charles W. Hull]

Originally Posted by Jon Fairhurst

The answer is that the lines can expose independently as short or as long as they'd like - and they can expose simultaneously. The problem is that they are read out sequentially, and each read takes some amount of time, so the exposures are all offset in time.

Evan, Jon, thanks for the explanation.