Help design a time lapse rig

[Jrnation109]

Hi,
This is my first post on CR. I have been a viewer on this site for some time and have read tons of posts on the forums. I am in my first year of being a pro photographer and recently put my education on hold to pursue something I am extremely passionate about (please don't try to lecture me as making $30,000 a year vs $100,000 doesn't matter to me as long as I love what I do).

Anyways, I have been in school for 4 years for Computer Engineering. I am currently working on a motorized time lapse rail system and I would love to have input from the community as to what to provide. My main goal in this whole process is to make it as cheap as possible, while still making a profit obviously, so that way anyone can buy it. I'm sick of seeing these rigs that offer little options but somehow get people to pay $600+ for them; it's not right and we shouldn't be forced to spend that kind of money on something so basic.

Right now I have the rail system almost done and am working on the programming aspect of everything. What features would you like it to include?

Here is what I have down so far:

• Set exposure time/shutter speed
• Set length of time lapse (30mins, 1hour, etc)

I plan to use this information to determine the time needed between each shot. The rig will stop while the shutter is open and then once it closes it will move a few cm or inches down the rail based on the timing you selected.

I would love to have your input as to what additional options I could provide in the menu of my circuit board that will be the "brains" of the rig.

Thank you for your help and I can't wait to make a great product that I hope you all will enjoy.

 

[Aglet]

Might just be me but those great videos made with time lapse and camera movement are starting to all look alike.

Add the ability to do non-linear camera speed on that rail in the process and that'll be a nice feature.
Think like an engineer.

add:
- pan-tilt ability too, also non-linear
- software that can help set up the shoot by automatically calculating the necessary movements, like pan-tilt and motion for precise tracking of sky or very close subjects. Could base it partly on *** & compass direction data from tablet/cell phone. iOS/Android app to do all that and DL the data to hardware controller.

I'll add more if I think of it.

 

[Jrnation109]

Might just be me but those great videos made with time lapse and camera movement are starting to all look alike.

Add the ability to do non-linear camera speed on that rail in the process and that'll be a nice feature.
Think like an engineer.

Are you talking about having a curved or circular rail? If so, I'm way ahead of you.

 

[Jim Saunders]

I think he meant non-linear steps rather than direction; The idea of moving the camera around while pointing at something in the image sounds intriguing.

Jim

 

[Hillsilly]

If it is cheap and portable, count me in.

Perhaps look at a system option?

1 - The "Starter Kit". This incudes a 2m length of track. You put your camera on one end (onto a dolly...if that's the right word?) and have the camera move along the 2m track over a 30 minute to 12 hour duration.

2 - Extra track which clicks into the starter track to extend the length. Also includes curved pieces (assuming you can work out how to do that. I wonder if a sushi train disc-style track would work for curves?). Maybe this could be the garden railway of the future?

3 - Steping motor attachment module which provides programmable stop-start features, accelerated steps etc.

4 - Motorised dolly, which provides programmable camera rotation and tilt features.

5 - Command module, which provides a usb and/or blutooth / wifi connection so that all features can be programmed via a laptop / phone.

6 - Remote controlled / programmable "car" for the motorised dolly (not necessarily for time-lapse, but might be fun for general videography).

People can then initially opt for the cheap starter kit and then they can steadily buy more components from you.

 

[Jon Gilchrist]

You could do most of this with LEGO Mindstorms.

 

[paul13walnut5]

Ability to program in smooth ramping start stops (bezier curve)

Fast movement preview mode.

Position store (manually move head to start, store, manually move head to end, store, repeat movement with timelapse)

Iphone / android control.

Easy GUI.

Future devs:

Programmable motors to work with geared lenses, ability to set zoom start stop, focus position, aperture position...

How cool to be able to do a timelapse trumbone shot, or assist dust dawn sequences by smoothramping of aperture?

Too greedy?

 

[Aglet]

Might just be me but those great videos made with time lapse and camera movement are starting to all look alike.

Add the ability to do non-linear camera speed on that rail in the process and that'll be a nice feature.
Think like an engineer.

Are you talking about having a curved or circular rail? If so, I'm way ahead of you.

I've thought about curved track but those become "hardware options." (how many different radii can you build and stock?)
Perfect for some applications, for others, I'd rather the pan-tilt control and as someone else mentioned, build onto that with zoom and focus pull options too. May as well motorize everything that can be motorized.
And yes, Lego Mindstorm/National Instruments may be a good place to start for such an application. NI developer edition is bloody expensive but you can then spin off standalone apps that can be sold, maybe by now also compiled for iOS and Android.

Would be nice to accelerate the dolly, along a straight stretch then slow down around a 90 degree curve and come to a decelerated stop. All the while the camera pans and maintains focus on a nearby subject, like a growing seedling? Then zoom out while maintaining focus, smoothly adjusting aperture for rapidly increasing depth of field to put the thing into environmental context over a much shorter period of time, actually accelerating back to real-time.

And I'm not even a director - but I can imagine being able to do such a shot would excite a lot of people in the film industry. Great way to add some zip to a documentary!

 

[Jrnation109]

If it is cheap and portable, count me in.

Perhaps look at a system option?

1 - The "Starter Kit". This incudes a 2m length of track. You put your camera on one end (onto a dolly...if that's the right word?) and have the camera move along the 2m track over a 30 minute to 12 hour duration.

2 - Extra track which clicks into the starter track to extend the length. Also includes curved pieces (assuming you can work out how to do that. I wonder if a sushi train disc-style track would work for curves?). Maybe this could be the garden railway of the future?

3 - Steping motor attachment module which provides programmable stop-start features, accelerated steps etc.

4 - Motorised dolly, which provides programmable camera rotation and tilt features.

5 - Command module, which provides a usb and/or blutooth / wifi connection so that all features can be programmed via a laptop / phone.

6 - Remote controlled / programmable "car" for the motorised dolly (not necessarily for time-lapse, but might be fun for general videography).

People can then initially opt for the cheap starter kit and then they can steadily buy more components from you.

1 - I have various lengths of track and different packages that will fit all budgets.

2 - I am in touch with a supplier of a curved system right now that can also go straight. This would give the photographer the ability to make any style of track they want; just depends on the pricing.

3 - This is going to be included. I like the accelerated start and stop thing you mentioned; I'll look into it.

4 - Dolly is motorized however I have not looked into the tilt and pan option. I would love to bring that on board but that would require a whole new motor system thus driving up the price. Maybe I could make an advanced package that offers both linear motion and tilt/pan so people who don't want it have the option to get the cheaper version.

5 - It will have usb and be open sourced so if you know how to program you could go in at any time and make your own changes to the code. I am not going to have an iphone app or wifi simply because theres no point. Yes, it is cool but all it does is drive up cost. There will be an lcd screen on the circuit board with buttons so you can just use that to make changes.

6 - Already have this in the works. This will be apart of a higher package that includes a rail system and a basic dolly that is not on tracks but can be controlled that same way as the rail system (steps, non stop motion, etc)


Offering various packages and features for different price points is a must for me. I don't want to force people to spend $500 when they only want a rail and a dolly but I want people to have an option for a system that packs in a ton of cool features if they want it. Thank you for taking the time to throw a few suggestions at me. I'm glad our ideas were almost the same; lets me know I'm on the right track.

You could do most of this with LEGO Mindstorms.

Financially it would be impossible for me to sell the system. Going the Arduino with stepper motor and rail system route makes it much more possible to have a product at a low cost.

quote author=paul13walnut5 link=topic=16650.msg307043#msg307043 date=1377676644]
Ability to program in smooth ramping start stops (bezier curve)

Fast movement preview mode.

Position store (manually move head to start, store, manually move head to end, store, repeat movement with timelapse)

Iphone / android control.

Easy GUI.

Future devs:

Programmable motors to work with geared lenses, ability to set zoom start stop, focus position, aperture position...

How cool to be able to do a timelapse trumbone shot, or assist dust dawn sequences by smoothramping of aperture?

Too greedy?
[/quote]

This will give people the option to have continuous motion so you can use it for video if you would like.

I'm not exactly sure what you mean by this, "Position store (manually move head to start, store, manually move head to end, store, repeat movement with timelapse)", could you go into more detail for me?

I won't be offering iOS or Andriod control at this time. I do not see a need for it considering every thing you would be able to do on the app you can do on the Arduino. These apps are more or less "fads" and I can't justify driving up my costs considerably just to seem cool; great idea though if I didn't have the Arduino as the brains of the system.

At this time I only plan on having the system be able to control the shutter speed. I don't know how to go about changing any of the settings on the DSLR itself through the remote port; I don't even know if this is possible considering the 1D line has a closed firmware. It would be cool to have the option to have an attachment that could keep the lens in focus while doing video. However, I honestly am not sure how to go about that at this time.

I wouldn't say you are being "greedy" I'd just say what you want isn't possible with an AIO solution. All you would have to do is run your pc/mac to your DSLR and change the settings as needed throughout the duration of your TL.

Might just be me but those great videos made with time lapse and camera movement are starting to all look alike.

Add the ability to do non-linear camera speed on that rail in the process and that'll be a nice feature.
Think like an engineer.

Are you talking about having a curved or circular rail? If so, I'm way ahead of you.

I've thought about curved track but those become "hardware options." (how many different radii can you build and stock?)
Perfect for some applications, for others, I'd rather the pan-tilt control and as someone else mentioned, build onto that with zoom and focus pull options too. May as well motorize everything that can be motorized.
And yes, Lego Mindstorm/National Instruments may be a good place to start for such an application. NI developer edition is bloody expensive but you can then spin off standalone apps that can be sold, maybe by now also compiled for iOS and Android.

Would be nice to accelerate the dolly, along a straight stretch then slow down around a 90 degree curve and come to a decelerated stop. All the while the camera pans and maintains focus on a nearby subject, like a growing seedling? Then zoom out while maintaining focus, smoothly adjusting aperture for rapidly increasing depth of field to put the thing into environmental context over a much shorter period of time, actually accelerating back to real-time.

And I'm not even a director - but I can imagine being able to do such a shot would excite a lot of people in the film industry. Great way to add some zip to a documentary!

I will look into the zoom/focus ability as well as the tilt/pan option. These would have to be offered at a higher priced package obviously considering the amount of work I would have to put in to build each one of these.

The Lego thing is way to costly like I stated earlier. For $30 bucks I can get all the motors and circuit boards I need through various suppliers.

As far as the rail system I am in talks with a company who offers straight rails and curved ones. You can take them and attach them to each other and create whatever shape of rail you would like. You can have it go straight for 2 feet then have a 90 degree turn, a 4 foot curve, and so on. I'd love to use this one but if the price is to high I will need to go to a different one. My current choice for a rail system can get me 6 feet for $20 so that will play into my deciding factor. Then you have to take into account how to program it. I can make it so it will slow down during a curve but what if someone moved the position of the curve? Now you have it going fast through the curve and slowing down during a straight portion of the track; programmatically it would suck using a system people can manipulate. I would have to remove the ability to slow down during a turn because the program wouldn't know where the turn was; if that makes any sense.

Like I said earlier, being able to adjust aperture isn't going to be possible. The only communication the program will have with the DSLR itself is when to open and close the shutter.


Sorry for the long post everyone. Thank you for all the ideas you have provided me with. It sounds like the majority want a tilt/pan and focus option. I will look into these and see how I can make this happen.

 

[Aglet]

Then you have to take into account how to program it. I can make it so it will slow down during a curve but what if someone moved the position of the curve? Now you have it going fast through the curve and slowing down during a straight portion of the track; programmatically it would suck using a system people can manipulate. I would have to remove the ability to slow down during a turn because the program wouldn't know where the turn was; if that makes any sense.

program could use total track distance as measured (or add up pre-programmed common pieces) and then the director specifies a velocity profile for the entire distance, you'll know where the curve is.

As an alt, use a hardware trigger (optical, mechanical) for telling the controller what section of track it's on. Could be used for straight and curved, could even ID them with an IR sensor and markings on track.

It should not be significantly more complex. If someone changes the track, they need to redo the velocity profile with the new track shape & length.

Like I said earlier, being able to adjust aperture isn't going to be possible. The only communication the program will have with the DSLR itself is when to open and close the shutter.

It's just more stepper motor utility channels to control the shaft on a cine lens rig; useful for AE operations.
Again, all these stepper motor channels can be GUI programmed using something like the velocity profile/track length for a simple, consistent interface that's relatively intuitive. (I'm supplying you patentable ideas here! )
Ooops, actually not. Just hosed that by providing published public disclosure here with witnesses.

Use multiple strip-chart type graphs of the computed total track length in x-axis; display them one above the other so you can see how they all line up and compare the functions relative to each other.

Each chart displays an active dolly-camera function.
- velocity profile on one (top)
- shutter activation period profile (if not constant this allows return to real time as I suggested)
- pan profile
- tilt profile
- focus profile
- aperture profile

MORE OPTIONS - auxiliary function profiles
- more motion channels
- lighting control channels
- filter control channels
- anything anyone could possibly think of adding control channels (binary output trigger points, analog control outputs for light intensity controllers, etc.)

you don't have to incorporate ALL these functions initially, but allow for them.
But if you want to build the best, most cost effective rig, then these FEATURES will really set you apart w-o adding much hardware development cost and the software is all parallel structures to manage these using the time-distance baseline.
VERY DOABLE, very simple interface. checkbox for what function channels you want active, define the curve on activated ones graphically with the option of numerical entry fields to be precise. (Again, national instruments SW does this in its sleep)

I'll stop now. When I start inventing (this is partly what I do for a living) I don't stop until I've built a Rolls Royce or Bugatti, then I pare it back down to fit the budget. :}

 

[paul13walnut5]

The ability to save the co-ordinates from a head position.

So the user will pan tilt track to their start position.

Save it.

Pan tilt track to end position.

Save it.

Movement is stored.

Can do fast run through to check.

Can then set as timelapse.

This would let users set positions without getting bogged down with 3d co-ordinates with - co-ordinates etc.

The biggest barrier to me buying a full timelapse set up us not so much cost, but the programmy nature diy element of exdisting set ups.

I don't care much for script or binary switches. The key to this is ease of use, one box arriving from dhl, no soldering.
If its also a good price you'll sell shedloads.

 

[Jrnation109]

The ability to save the co-ordinates from a head position.

So the user will pan tilt track to their start position.

Save it.

Pan tilt track to end position.

Save it.

Movement is stored.

Can do fast run through to check.

Can then set as timelapse.

This would let users set positions without getting bogged down with 3d co-ordinates with - co-ordinates etc.

The biggest barrier to me buying a full timelapse set up us not so much cost, but the programmy nature diy element of exdisting set ups.

I don't care much for script or binary switches. The key to this is ease of use, one box arriving from dhl, no soldering.
If its also a good price you'll sell shedloads.

Thank you for elaborating on that. I love the pan/tilt option since there are very few systems out there that offer them. I will work as hard as possible to add it into the system. Assuming my kickstart campaign is a success, the product will be shipped in nice packaging with some marketing materials included. You will not have to do anything besides put a battery or two in and then turn it on.

I am considering adding the option of letting people purchase a package that has all the parts in a box not assembled but including detailed instruction how to put it together with the software on a thumb drive. My thought is some people may enjoy learning how to build this and how the software works; would anyone be interested in this?

 

[tpatana]

Mostly copying from other people ideas:

-more than 1 axis movement

-rotate / tilt

-individual parameters / programmable curves for each freedom of movement. E.g. rotate camera 1° every 1 minute towards left, same time move with curve x= 0.5t^2 - 0.3t + 5, y = 0.15t

-should support all common mathematical items, sin, cos, ln, log, e, ..etc...

-fast movement preview to verify movement. Hate to see the planned track to fail after hours of shooting (great idea Paul!)

 

[chasinglight]

I have always been tempted to buy a time laspe dolly/rail system, but have been discouraged for 2 reasons. 1) cost 2) bulk (most places I envision using this are not in my back yard!). If you could bring the cost down I would definitely be interested in playing around with one of these systems. Moreover if you could reduce bulk (break down the rails to shorter pieces, but still assemble to a full size kit) without compromising stability then I think you would really be onto something that no one would have an excuse not to try!

 

[Jrnation109]

Mostly copying from other people ideas:

-more than 1 axis movement

-rotate / tilt

-individual parameters / programmable curves for each freedom of movement. E.g. rotate camera 1° every 1 minute towards left, same time move with curve x= 0.5t^2 - 0.3t + 5, y = 0.15t

-should support all common mathematical items, sin, cos, ln, log, e, ..etc...

-fast movement preview to verify movement. Hate to see the planned track to fail after hours of shooting (great idea Paul!)

Not quite sure why I would want to include mathematical logic like that; to be honest. If I had to guess I would say > 1% of the end user would even know what that logic meant, lead alone how to be able to use it effectively. Sure, it is very simple to incorporate mathematical logic like that in programming language but I don't see how it would be a positive contribution to an end product.