Hi all,

I'm a developer for Shutterbug Remote (http://shutterbugremote.com) and I'd like to get a little feedback on a bulb ramping question from the guys who actually do it. Right now, SBR will allow for a 3 point ramp with any number of photos in between each point. The exposures are evenly spaced in time and ramped linearly in time (not f-stops).

Now, I was planning to add capabilities for ramping the exposure in f-stops, which means working exponentially in base 2 - but, most of the bulb ramping applications I've seen are for day-night transitions, and I would imagine that like many things in nature, the amount of light is changing exponentially (base e). So can anybody point me to the most appropriate reference for the "math of bulb ramping"? I'd love to see a few plots of exposure vs time of day and how it changes with seasons, etc. Why are most intervalometers using f-stops and not "e-stops" for lack of a better word?

Thanks in advance for the lessons......

Video from Chris at "Project Chronos" or Jack Ripper on here.
http://vimeo.com/44010815

I've been designing my bulb rampers using f/stops because cameras use that nomenclature, and using stops/hour is a rough approximation for the natural decay of light. There's a couple graphs on this thread of stops/hour vs time.

viewtopic.php?f=28&t=9704

John - great looking gadget and some excellent info in that thread. Thank you.

The higher order exponentials in your exposure plots are interesting - I feel like actually metering each shot in hardware and then adjusting your rate based on the metering might provide good results with little flicker (like a PID loop where the feedback is from a light sensor.)

I was also surprised at the accuracy you are striving for. What is more important - the absolute value of the exposure, or the repeatability of the exposure?

Good luck with your project!

I don't know how good the result would be if you were to use the nth degree formula. It might be just as well to simplify the curves into a few linear segments for ease of calculations. Using a light sensor in conjunction with the pre-defined curves might actually work pretty well.

I think both of those go together. If you have high accuracy with the absolute values then it will have good repeatability.

Micheal I'm wondering how the shutterbug will achieve smooth results with-out feedback? It's been proven many times, without timing feedback from the flash syc the results will be just as prone to flicker as auto exposure.

I agree. I would say that this impossible to do this way.

Because f-stops are always relative to some previous position or starting point, and if I want to change one of my other settings (ISO or aperture), I need to know my shutter speed is going to behave accordingly.

If you're shooting a 1 second exposure, and then you add another whole second, you've added a whole stop.

If you're shooting a 1.5 second exposure, and then you add a whole second, you've added about two thirds of a stop.

If you're shooting a 3 second exposure, and then you add a whole second, you've added about a third of a stop.

If you're up to 30 second exposures and you add a second, you're not making much difference to your shot at all.

For a smooth transition of exposure, it needs to be relative to the previous exposure, the exposure needs to be increased (or decreased) by a certain percentage (whatever fraction of a stop), rather than a fixed amount of time.


The two kind of go hand in hand, they're both as important as each other. Knowing what your exposure should be is nothing without the ability to reliably implement it in your shot. If it's not reliably repeatable, it's really not all that accurate.

Have to say though, I've never felt the need to get PC sync socket feedback from the Nikon bodies to the Arduino, other than, initially figuring out the shutter lag, when sending in signals through the 10 pin port, and managed very consistent shutter speeds without flicker in the playback (using M42 lenses to eliminate aperture flicker). I don't do it in general operation.

This one was with the 70-200mm f/2.8VR wide open at 200mm on the D300s, with the shutter speed controlled by the Arduino. Exactly the same exposure on each shot, but no flicker at all, and there's been no kind of deflickering done in post. As I said, Nikon bodies (or at least the D200 and D300s) are very consistent from shot to shot.

Milapse / Cronix: Agreed that feedback is important for ramping. Bulb ramping for timelapse is also pretty niche as far as "most users" of remotes are concerned. So what I'm thinking about is a second dongle (let's call it a Shutterbug Ramper) that attaches to the hot shoe and has a built in light sensor. Anyway..... thanks for all the feedback!

Quick update:

I threw together a small PCB with a ambient light sensor on it - features very high dynamic range, built in UV and IR filters, and compensation for lux units. I capture the data from this sensor using a microcontroller over USB and recorded the readings every second or so. At the same time, I ran a time lapse taking a photo every minute using shutter priority mode. The idea is to see how to camera metering compares to the sensor values, and how the sensor values change with the scene. Here is the time lapse: http://www.youtube.com/watch?v=hdTIyWxSvy8 and the sensor data is attached.


(Yes, the time lapse sucks. Wasn't meant to be pretty, just needed a reference.)

A few interesting things: The change in light intensity during full sun (at the beginning) due to shadows, clouds, etc. is really large. And on this particular light, the sensor dropped down to about 0.05lux, which is really quite dark.

Conclusion: I think that driving a bulb ramp exposure, where this type of sensor is used to "push and pull" the exposure times as part of a feedback loop incorporating flash sync will work nicely. Should have some results coming soon.