RED ONE to have slow shutter/intervalometer capability?

[ColinSmith]

I'd agree with pretty much all that Jim.

As far as I can see doing the frame accumulation is post gives more options for control and post processing, with the only plus for doing it in-camera being easier exposure control in low light additive blends.
My stuff has all been done in post, and I wouldn't change that even if I had a camera that'd do it as it shot.

My one niggle... A 24fps sequence blended in groups of 3 frames is still different than a 8fps sequence though ;-) Just depending on the subject it might not LOOK any different ;-)

 

[Keith Brust]

Tom, I shot the sequence on the Saguaro cacti in the Deserts program.

Keith

 

[Keith Nealy]

Realtime access to the creative process

Realtime access to the creative process

JIm and Colin,

Thank you for making the "in post" version a little easier to understand.

Jim said:

but if done "in camera" then you are locked into the look the camera gave you with no option to change it.

The advantage of doing it in camera as in the XDCAM-HD system is with immediate playback. You may be shooting in such a dark situation that you can't even see well to compose and can only judget the take when it is reviewed. Then you can evaluate exposure and change accordingly and review again until you get the look you want. No amount of post wil make up for a badly composed shot or one that has really been exposed badly - so each system has it's advantages.

Jim also agrees by saying:

Of course there are times when you just want lots of footage processed quickly and be done with it, and see the effect real-time on your evf, and hardware based in camera processing is good for that, BUT you aren't excluded from the ultimate final result of an accumulation effect with the existing operation of the RED as we currently understand it.

Frame accumulation is like "nightvision" but without the noise and realtime playback.

Jonathan Ames has done some wonderful dramatic work with it by shooting scenes of an entire room lit by a single candle.

In-camera is also easier , faster and cheaper which also have their places in the production process of some films.

What excites me about computer controlled digital recording is the new frontiers it might open up for us. What's beyond frame accumulation?
And, I'm not looking for new gimmicks... but new tools.

Non-linear editing gave us creative feedback and freedom we never had in linear editing and filmmaking showed improvements because of it.

I have cut as many films on a moviola as I have on an Avid and I assure you the NLE films are better because the tool was better for the creative process.

I am not a big fan of special effects in place of story but as a creative director 35 years I am in favor or any tools that give me realtime access to my creative process.

The potential we have with RED to "Go where no man has gone... before" is enormous and we are fortunate to be a part of making this transition into the democratization of digital cinema a reality. And we are doubly fortunate that we have leaders that are astute enough to listen... which, in a market driven environment, can only improve success.

aloha,

Keith

 

[Ralph Oshiro]

Tom, I shot the sequence on the Saguaro cacti in the Deserts program.

Keith

Wow! I saw "Deserts" on a 50" 720P plasma. That was GREAT stuff! Really beautiful work, Keith!

 

[ahusain]

Ralph, I would be very shocked if any camera could expose for stars in under 10 seconds... even with the ISO cranked through the roof. Maybe with a full moon, but still, I can't imagine any camera being able to do it at 1fps.

The question a lot of people want to have answered is whether the RED will do long-exposure timelapses, like 45-second night exposures? Right now people can do it with DLSR's, but the post production chore is a real pain compared to the type of recording an HVX does, for example. With an HVX, you shoot a timelapse, and you can play it back right there on camera, which is pretty amazing. With DLSR, you shoot the frames and then you have to stream them together in an NLE later. And for me, because I shoot RAW DLSR timelapses, the post is even more complicated. If the RED could do long-exposure timelapses the way the HVX does short exposure timelapses, that would be amazing.

But as others have pointed out, with DLSRs, they are so cheap you can have several running at once. And if you are going to burn out your sensor or shutter, better to burn out a $700 DLSR body than a $17,000 RED.

hmm. sony vegas makes this kind of thing trivial: just go to the directory in which all your DSLR pics live, sort them by time using the windows directory bar at the top, then select-all and drag and drop on a vegas timeline. you're done!

you can select the default time you want any still pictures to occupy, so you can change the framerate to whatever you want (you have to do this before you drag and drop though).

took me a few seconds to create video from a huge sequence of jpegs.

 

[Greg Syverson]

I still remember a few years ago asking a major camera manufacture about frame accumulation on the bigger more expensive cameras. They replied that is was only found in less quality cameras. Now it appears that is does have a use as an effective tool. 4k HD frame accumulation. I like it.

 

[tj williams]

Colin thanks for the lesson in "frame accumulation"
Can his effect achieved from standard 24P footage in Avid or Final Cut?
Will such a post process increase the cameras ability to look into dark areas?

 

[Jim Arthurs]

My one niggle... A 24fps sequence blended in groups of 3 frames is still different than a 8fps sequence though ;-) Just depending on the subject it might not LOOK any different ;-)

Yes, you're absolutely right! My mistake!

I'd argue that the overall total temporal information presented over the run of the sequence is the same, but it is certainly displayed differently... 8fps will give you a distinct "stepping" feel, while the accumulated 24fps will be more of a smooth flow with the lowest and highest steps dropping off and appearing on at a faster 24fps basis...

As a final note, it will be interesting to see what the final RED hardware allows, which is the bottom line in this discussion. My gut feeling is that the desired hardware and displayed "accumulation" will be beyond the scope of the first run of RED... but who knows? Some things might be firmware upgrades, and some things will be board level swaps at a time in the future perhaps.

It sure will interesting to see what is allowed for slow shutter... at least the RED Team is aware of our various interests and desires along with the rationales...

 

[Jim Arthurs]

Colin thanks for the lesson in "frame accumulation"
Can his effect achieved from standard 24P footage in Avid or Final Cut?
Will such a post process increase the cameras ability to look into dark areas?

TJ, the answer is a qualified "Yes"... as long as you shoot 360 degree shutter, so that the blur continues from one frame to the other without a gap in the motion.

For example, you could literally shoot 2 stops under-exposed and do an additive blend over three frames to get your exposure back to normal... as to the actual steps in FCP or Avid... I'm uncertain as I'm not as familiar with either...

 

[ColinSmith]

To do it in post you'd ideally shoot with a 360 degree shutter, so when you combine frames there are no "gaps" in the motion blur. Apart from that, combining frames is easy enough though, and almost totally flexible in post.

I do it in After effects, but it should be workable in most multi-layer apps.
Normal setup is like this;

Layer 1 is straight video
above that, Layer 2 is offset 1 frame, 50% transparency
above those, Layer 3 is offset 2 frames, 33% transparency
above those, Layer 4 is offset 4 frames, 25% transparency
................

and so on, for as many frames as you want to blend.
The settings above make an equal combination of all the frames, truly like a single long exposure per frame, but you could make the transparency change in different steps too, maybe to make the long motion trails fade off more with time for example.

Now, in the above example each frame should be exposed normally.
To shoot in low light you may want to "add" frames instead of just blending.
I've not tried this one yet, but it would be the same setup, except put the layers in an "add" blending mode. I would expect that for most shots you would still ramp off the transparency to be able to combine enough frames without it all blowing out, but that would depend on the subject and exposure.

In the "add" case you would not have to get a full normal exposure in camera, as you will be boosting the exposure by adding the frames......if you want to add 24 frames, the camera exposure can be 4 1/2 stops under exposed for example...

I'd need to play with this a bit and see how best to work with the exposure settings to combine frames without killing it with noise. The big plus with this post work though is that you will almost always get it to work one way or another ;-)

 

[ColinSmith]

My gut feeling is that the desired hardware and displayed "accumulation" will be beyond the scope of the first run of RED... but who knows? Some things might be firmware upgrades, and some things will be board level swaps at a time in the future perhaps.

Yes, thinking about it, a frame accumulation would need hardware to buffer 24 to 48 4k raw frames, and processor power add or blend them and write out the new frames at 24 fps.... which would be a fairly heavy task in addition to everything else that is going on....

 

[Tom Lowe]

hmm. sony vegas makes this kind of thing trivial: just go to the directory in which all your DSLR pics live, sort them by time using the windows directory bar at the top, then select-all and drag and drop on a vegas timeline. you're done!

you can select the default time you want any still pictures to occupy, so you can change the framerate to whatever you want (you have to do this before you drag and drop though).

took me a few seconds to create video from a huge sequence of jpegs.

Sure, any NLE can do this. But I was talking about RAW photos. Also, keep in mind that you have to resize the images as well. Doing post on RAW timelapses is fairly labor intensive, if you ask me. The HVX is much, much easier. But then again, at least it's not chemical film!

 

[Adrian T.]

For example, you could literally shoot 2 stops under-exposed and do an additive blend over three frames to get your exposure back to normal...

But you're also increasing the noise level. When adding up frames you're also adding up the noise.

 

[ColinSmith]

Yes, but then again, it is the exact equivalent of what the Sony 350 does in camera.....
I'm interested to try it out, and see what different combinations of frame blending and adding and post noise reduction come out like.
Blending smooths out noise, so possibly adding and blending equal numbers of frames comes out not so bad.....

 

[Jim Arthurs]

But you're also increasing the noise level. When adding up frames you're also adding up the noise.

Think of it this way... the noise pattern is different from frame to frame... so when you blend 2 or 3 or more frames together, any one final pixel is a dithering of multiple noise positions, while the subject matter, if it isn't in fast motion, stays more or less the same from frame to frame...

People use this technique to produce grainless "freeze frames", and many of the common stop motion capture softwares do this multiple frame blending to make inexpensive web cams and dvcams look far better than they normally do.

One trick I do when I want to sample noise or grain is to find a black shot with no camera motion, sample it over enough frames to make it grainless, then do a difference between a single original frame and the clean mix frame to pull the actual grain pattern out, then do it again and again on sequential frames, creating a grain/noise only sequence, authentic to that particular camera, useful for overlaying onto CGI elements to make them match the live action better...

 

[Stuart English]

Frame Accumulation

Frame Accumulation

So to continue this exporation. When the frame accumulation is done, is it a rolling process (a) In camera (b) in post

i.e.
1st frame recorded = Frames A + B+ C + D + E
2nd frame recorded = Frames B+ C + D + E +F
3rd frame recorded = Frames C + D + E +F +G

Or
1st frame recorded = Frames A + B+ C + D + E
2nd frame recorded = Frames F + G + H + I + J
3rd frame recorded = Frames K + L + M + N + O

Understood that you could choose to record one or more frames at a time with the output of these calculations ....

 

[Anders Holck]

What I would like to have in the camera is the ability to do Real long exposure with essentially an open shutter which Stuart described as
1st frame recorded = Frames A + B+ C + D + E
2nd frame recorded = Frames F + G + H + I + J
3rd frame recorded = Frames K + L + M + N + O

This saves a lot of space on the reddrive when doing something like a 24 hours time lapse, when essentially you are going to merge the frames in post anyways.

The frame accumulation that colin describes is pretty easy to do in post as has been discussed, it's essentially a additive video delay effect.
After effects has a plugin called Time>Echo, which allows you to do something like this.

 

[ColinSmith]

The first way is how I have post processed my shot.
It gives a long exposure, 5/24s in your example, while keeping a 24 fps recording rate.

The second example gives 5/24s exposure too, but now the recording rate is correspondingly down to 24/5 fps..... a traditional timelapse in other words.

The first method is what we (I) have mostly been talking about, although constructing the sequences in post you would have the option to build up your composite frames from any combination of shot frames (1/24s 360 degree shutter) that you chose.....

And OK, having just read Anders post, yes, the second option has advantages for regular long duration timelapse...

 

[JohnF]

Why are we talking about post based frame accumulation when slow shutter would do the job just as well, if not better (and simpler), and save much in storage and time in post?

And as someone has already mentioned if frame accumulation is not processed correctly it can increase noise not decrease it...

And OK, having just read Anders post, yes, the second option has advantages for regular long duration timelapse...

It has many other advantages for creative filming applications full stop.

Slow shutter please!!!

JohnF

 

[ColinSmith]

I'd see them as 2 different option to choose from.
Slow shutter cannot do everything that frame accumulation can do, and frame accumulation may have it's own disadvantages.

In Stuart's examples you can see that the "slow shutter" processing is just a variation of the "frame accumulation" processing in any case.... the only differences are the frame offsets and the fps write speed.

It may be that frame accumulation in hardware has too much of a price to pay in terms of adding capacity to the camera, but if not..... nice to have the choice..... otherwise..... just keep doing it in post, although that could apply to slow shutter as well ;-)