Intervalometer Ramping Module (adv_int.mo)

[dmilligan]

here's an example, to ramp Av from f/22 to f/4 in 20 frames do this:

• set the Av to f/22 in canon
• go into ML menu
• A > New Keyframe...
• Select "Aperture" menu item and turn it ON (this means this parameter will be included in the keyframe)
• Select "Create Keyframe"
• go back to canon and change the Av to f/4
• go into ML menu
• A > New Keyframe...
• Change "Keyframe Time" to 20
• "Aperture" should still be "ON"
• Select "Create Keyframe"
• Go back up one level and select "List Keyframes", you should see two, T=1 and T=20

You can set keyframes based on time of day as well, just turn "Use Global Time" on, then in the "New Keyframe" menu under keyframe time, you sill see a time (hh:mm:ss) rather than frame # (and on the right side is the current camera time)

To change a keyframe you already created, you can set the keyframe time to the time of the keyframe you wish to edit (if you do this you will see a warning that you will be overwriting an existing keyframe).

[saltador]

weird...i did a short test and upload all  20 images time-lapse to lightroom and i see av ramping but all images are just white and weird ping stripes..can't see nothing...

follow all the steps as you wrote.

only thing i had is advance intervalometer ON , in A menu..could that be the problem..doubt it..will try
without.

thanks.

[saltador]

also..i install the las built..sept 28..for 5dii...and drop the advent.mo in the module folder...is this right?

thanks

[dmilligan]

If you can see the A menu and can create keyframes then you have it installed correctly.

If you turn it off then it won't do the ramp when you run the intervalometer, but do a test without it to verify it's not the module causing the white frame thing (could it perhaps be your exposure settings? If your in M mode ramping the Av like this without compensating with shutter or ISO is going to make your image brighter and brighter possibly pure white unless your scene is changing i.e. a sunset) so you should probably run it in Av mode or with AutoETTR turned on.

[saltador]

yes i see the A menu..


I'm doing a test in M mode..auto ETTR..always on....no deflickering...av ramp... intervalometerevery 6 sec..

weird..

doing another right now..with a higher intervalometer..now at 12sec..lowest shutter at 1/6..ETTR always on..no deflickering..
will let you know..if i get same messy frames again..

thanks

[saltador]

ok...reporting again..

did a few test...looks like it has to do with this module cause the image preview in the importing window in LR looks fine
but once imported into light room the become a white and colored mess...

I tried it with deflickerin on and off, with different intervals and always the same...

anybody else having issues..

arrrrrrr...

i rather be doing timelapses...hehehe

thanks.

[saltador]

my mind keeps insisting...

so i deleted ML from card..droped unified version..add it built 29-09..add it ad_int.mo..did a test and same things..frames look white as it updates previews..

so, i checked the settings and it looks like is doing the av ramp but the iso and shutter are not changing at all..from looking at the exposure setting (frame is all weird colored, not visible at all).

good nite. 

[dmilligan]

You say the previews look fine in Lr, but then it doesn't open right, to me this sounds like a problem with Lr. Do the images look okay in camera? Do they look okay if you open them in a different program? What happens if you just take some pictures with the same settings you have and open them in Lr. What if you put the camera in JPEG mode instead of RAW? Do you have dual_iso turned on? if so you'll have to preprocess them with cr2hdr first or they wont open correctly in other programs.

[saltador]

You say the previews look fine in Lr, but then it doesn't open right, to me this sounds like a problem with Lr. Do the images look okay in camera? Do they look okay if you open them in a different program? What happens if you just take some pictures with the same settings you have and open them in Lr. What if you put the camera in JPEG mode instead of RAW? Do you have dual_iso turned on? if so you'll have to preprocess them with cr2hdr first or they wont open correctly in other programs.

yes...you save me here..thank youuu...i was tired testing variations to find the error..and you help me solve this.

In case it happens to others.
I had press the link to dual iso in the auto ETTR menu but i did not load dual iso in the module menu.

I think this is what solve my issue.
now it loads fine in LR
back to more timelapses test.

thanks

I promise to read more ...hehehe...thanks

[audunbrekke]

Hi. This seems like a greate module with many options for manual setting of keyframes. Would it be possible to make a function like the intervallometer with sunset / sunrise that was in earlier ML releases ? The sunset function was like this: the ML would adjust the Tv and iso as needed from exposure metering, but only in one direction for brighter exposures. It was also possible to adjust what the maximum average change could be, to prevent flickering. I know everybody is talking about AETTR, but the intervallometer with sunset function was a fast and easy function.

[dmilligan]

This module was not designed to do anything "automatically", IMO AETTR is much faster and easier than the previous sunset/sunrise function. You could however create a manual sunset or sunrise keyframe sequence with the module and save it. The module allows you to save your keyframes, I'm currently working on expanding this a little, right now the filename it saves to is hardcoded, so you can only save and restore one keyframe sequence.

Have you tried the post deflicker option in ML? I've found it works really well as long as you give it enough time to work between frames (5-10secs usually). I think alex is working on making this a tool you run in post, which would be even better.

[dmilligan]

Update: added looping

[Joachim Buambeki]

IMO after effects needs a way to keyframe ACR settings, that would be freakin awesome (also very useful for RAW video)

... I might be able to write a script for AE that makes ramping ACR settings easier. I know that the script API for AE has some pretty good integration with the native AE GUI, but I'm not that familiar with the AE scripting..

Did you find time to look into this? Beeing able to do everything within After Effects would be revolutionary!

[dmilligan]

I decided to go with Br instead (hence the script I created), it's much faster. AE takes several seconds to render a RAW frame, even on low quality (this is so annoying, why adobe why?!), Br has some kind of fast preview mode that takes fractions of a second. It's just much easier and faster to do in Br.

[jezhunziker]

Wow amazing module!! I've been testing this on 5Dii some great results, a couple of things to report and would be interested to hear thoughts on these... Most of my testing has been with Focus changes...

1./ I have found that focus keyframes require more than 4 second interval otherwise the intervalometer slows down to compensate for change, (could this be improved by adjusting any of the ML focus parameters?)

2./ I find that the most accurate way to determine how many focus steps are required is to frame up scene then choose "Focus Rack End Point" (from ML menu), from here I adjust focus in steps to the desired focal end point and record this number to input into the Adv Int keyframe. Is there a quicker method? ie is it possible to change focus on camera manually and store as a keyframe?

3./ Using 24-105 L lens at 80mm has more steps than at 24mm and at higher zoom (ie 105mm) it can include more than 1000 steps, is it possible to include a higher number in next build ie -3000-3000, (I know it's possible to change the step size to 2 but the focus change isn't as smooth)?

4./ As focus keyframes need camera to be in live view. Is there any way to save battery or turn off Liveview whilst running a time-lapse?

Thanks so much for all amazing work, I'll post some tests shortly,

 

[dmilligan]

1. Yes, I've found the same to be true. I used the code in the core that does the focus change. I tried taking some stuff out to speed it up, but it stopped working altogether. I assume that whoever wrote that code knew what they were doing and probably tried to make at fast as they could (they certainly know more than me). I think the only possible improvement would be reverse engineering a lot further Canon's focus control mechanisim, which is way beyond my current abilities.

2. No, there's not really a better way to do this, I wish there was, but again it would probably require more reverse engineering. ML has no way to really figure out how many "steps" the current focus position is at when focused manually. Basically all it can do is step the focus and count the number steps you take. I could potentially at least pull in the focus rack end point setting automatically, but that requires a change to the core (to allow modules to read this value). I've already submitted a pull request for what I would need to do this (and reading/writing many other ML settings), but its still pending.

Perhaps you can spend a little time creating a chart for each lens of steps vs. focus distance. You could even put marks on your lens like the distance markings, with step numbers, if you dont mind defacing your lens. I've thought about doing this to my nifty fifty (also add distance markings too since it doesn't have those either).

3. Thanks for the feedback, I had no idea what the maximum for any possible lens was so I just guessed (none of my lenses step more than several hundred). I'll update that to 5,000 just to be safe.

4. LV has to be running for focus to work but you can try this: http://wiki.magiclantern.fm/userguide#powersave_in_liveview

[a1ex]

I guess it's a good idea to remove the focus ramping from the core, and just leave a basic intervalometer there, right?

Also I'm thinking to merge the module this week, but I want to solve the menu issue first. I thought a bit at how to implement menu with submenus (basically I'll rename submenu_mode to something like submenu_level and keep the existing data structures).

[jezhunziker]

1. Yes, I've found the same to be true. I used the code in the core that does the focus change. I tried taking some stuff out to speed it up, but it stopped working altogether. I assume that whoever wrote that code knew what they were doing and probably tried to make at fast as they could (they certainly know more than me). I think the only possible improvement would be reverse engineering a lot further Canon's focus control mechanisim, which is way beyond my current abilities.

2. No, there's not really a better way to do this, I wish there was, but again it would probably require more reverse engineering. ML has no way to really figure out how many "steps" the current focus position is at when focused manually. Basically all it can do is step the focus and count the number steps you take. I could potentially at least pull in the focus rack end point setting automatically, but that requires a change to the core (to allow modules to read this value). I've already submitted a pull request for what I would need to do this (and reading/writing many other ML settings), but its still pending.

Perhaps you can spend a little time creating a chart for each lens of steps vs. focus distance. You could even put marks on your lens like the distance markings, with step numbers, if you dont mind defacing your lens. I've thought about doing this to my nifty fifty (also add distance markings too since it doesn't have those either).

3. Thanks for the feedback, I had no idea what the maximum for any possible lens was so I just guessed (none of my lenses step more than several hundred). I'll update that to 5,000 just to be safe.

4. LV has to be running for focus to work but you can try this: http://wiki.magiclantern.fm/userguide#powersave_in_liveview

Thanks for reply, adding extra focus steps will be really useful. The Rack focus end point isn't a bad workaround as we all know a good timelapse is all in the setup...

It's strange how much the focal distance effects the number of steps, I wonder if it's constant from one lens to another, I only have one variable focus lens but I could see if it matches number of steps at 50mm against my nifty 50.. I'll try and post a chart for 24-105mm

Thanks for the advise with LiveView - I'll try and find a way to shut off it off once TL is started either that or will need to purchase a few extra batteries.. 

[dmilligan]

I guess it's a good idea to remove the focus ramping from the core, and just leave a basic intervalometer there, right?

Yeah

Also I'm thinking to merge the module this week, but I want to solve the menu issue first. I thought a bit at how to implement menu with submenus (basically I'll rename submenu_mode to something like submenu_level and keep the existing data structures).

That'd be great! I'll make the change to the module as soon as you get that done.

[Francis]

Thanks for reply, adding extra focus steps will be really useful. The Rack focus end point isn't a bad workaround as we all know a good timelapse is all in the setup...

It's strange how much the focal distance effects the number of steps, I wonder if it's constant from one lens to another, I only have one variable focus lens but I could see if it matches number of steps at 50mm against my nifty 50.. I'll try and post a chart for 24-105mm

Thanks for the advise with LiveView - I'll try and find a way to shut off it off once TL is started either that or will need to purchase a few extra batteries.. 

It definitely is not constant from lens to lens. So many variables such as the motor in the lens as well as the range of the focus ring. It's not even consistent within a lens. My 100mm macro feels almost exponential in the proportion to each step (or physical turn of the ring). By this I mean as you go from the closest focal distance to infinity, the sensitivity of the ring substantially increases. This would make sense for a macro lens where you want a lot of control when focusing on objects at a macro range.

[J24]

I use a motion control system so I'd love to have an option to trigger the ramping and shooting with an external (wired) intervalometer.

Another good option would be if it would be possible to do the reverse: sending the shoot signal from the camera to the external controller? My controller can take the signal as input but can the camera send it? In this scenario ML and adv_int would control both the camera and the moco system.

Either functionality would be cool, since it would make the syncing of the shooting with the camera movement MUCH easier. I think Francis might be trying to achieve something similar here:

I want to be able to trigger an intervalometer sequence using an IR or wired remote.

[dmilligan]

I use a motion control system so I'd love to have an option to trigger the ramping and shooting with an external (wired) intervalometer.

Should be pretty easy to do, I'll look into it.

My controller can take the signal as input but can the camera send it?

What kind of signal can it take? ML can send various types. Most flexible is probably PTP over USB. It's also possible send a signal on the mic in line by turning the audio amp on/off. Something like this should probably go in it's own module, really easy to write one for it though, there's a CBR for when the intervalometer takes a pic.

[itsskin]

Hi! I think it's really cool to add feature like automatic interval ramping.

"ETTR exporure time" + user set X seconds. This will help to avoid overlaping shoot interval with actual exposure time and give AETTR algorithm enough time to read the preview to determine correct exposure.

Lets say we are doing day into the night timelapse. We want 15 sec intervals during the daytime. But later at night exposure time will be 30 seconds. The feature I'm asking for will do this automatically, so we do not have to predict times and set key frames for manual interval ramping.

Thx!

[dmilligan]

Doing it automatically could lead to some situations where you end up with a lot of jitter b/c ettr might be going back and forth between different shutter speeds. It would be kind of like the exposure flicker you get with ettr, except in the time domain (which is not possible to fix like exposure flicker). If you want to ramp the interval period, it is better to carefully plan it and do it manually so you get a nice smooth ramp in speed. Any good timelapse requires meticulous planning anyway.

[itsskin]

You are absolutely correct. I do a lot of timelapses and they force me to stay next to the camera for several hours. Your module + AETTR is a way to put camera on "Auto" and do other stuff. I'm getting really great results of this combo + your deflicker script. And do not mind some jitter in time domain - can fix it in post for what I do. If it's not hard to implement, I think a lot of people will say thank you.