Accurate USB remote sync for single camera shots? - page 3 - Microfunguy's Builds - CHDK Forum supplierdeeply

Accurate USB remote sync for single camera shots?

  • 33 Replies
  • 16407 Views
Re: Accurate USB remote sync for single camera shots?
« Reply #20 on: 16 / May / 2008, 17:24:48 »
Advertisements
We bought an S3 IS and confirmed with a rotating mirror and a laser that CHDK software indeed enables shutter speeds of 50 micro seconds or less.


That is the first 'authorative' confirmation that these fast electronic speeds are real.

Quote
The trigger pulse to take the shot had to be at least 30 milliseconds, the camera did not respond for shorter trigger pulses.

The keyboard is polled every 10 msec so I guess the signal has to be present for three consecutive polls.

Quote
Is the camera hardware capable of being triggered with microsecond accuracy and how can we do it?

It is FAR more accurate to wait for the trailing edge of the trigger pulse.
In StereoData Maker we can sometimes synch two cameras to almost 50 microseconds.

Quote
We are also planning to call and ask someone from canon, I don't know how canon feels about CHDK.


 
Oh dear, I hope that does not cause problems.


David

*

Offline fudgey

  • *****
  • 1705
  • a570is
Re: Accurate USB remote sync for single camera shots?
« Reply #21 on: 16 / May / 2008, 18:00:43 »
...and to clarify a bit: You can get better accuracy with CHDK compared to what you got from your experiments so far. You need to use the SDM build's sync feature. It was written to sync two cameras which get their triggers from a single USB remote wired to both cameras, but it can also help you to trigger a single camera more precisely. I don't know if it's compatible with the S3IS, though.

The delay from your remote trigger pulse to the shutter opening doesn't actually get any shorter using the sync feature, but (unlike in the standard Allbest build) the delay will be pretty constant (to almost 50 us for some cameras, quoting Microfunguy above, which could be enough for you to capture some nice images with a greater probability).

So, this will only help you if you can predict your plasma events by a couple of tens of ms or so (or whatever the delay is) i.e. if you for instance are actuating them electronically.

*

Offline BB

  • ***
  • 164
Re: Accurate USB remote sync for single camera shots?
« Reply #22 on: 16 / May / 2008, 18:05:45 »
And since this is all open source... You could probably trace the "shoot" code and optimize it for your exact application (you may have to load the firmware and trace that routine too...).

-Bill

Re: Accurate USB remote sync for single camera shots?
« Reply #23 on: 16 / May / 2008, 18:23:41 »
I don't know if it's compatible with the S3IS, though.

Actually, the s2is and s3is were not recommended for use with SDM.

We could not get the s2is to work and the s3is has a widely varying delay before firing the shutter.

The shoot sequence is interrupted at the last possible stage and held-up until the signal on the USB connector drops to zero.
For many cameras there is then a fairly repeatable delay before the shutter fires.

With a bit of electronics you could create the plasma at the same instant.


David


Re: Accurate USB remote sync for single camera shots?
« Reply #24 on: 17 / May / 2008, 12:18:59 »
With a bit of electronics you could create the plasma at the same instant.


In fact, you can also do it in software.

In 'StereoData Maker' (SDM), the delay after the USB signal goes to zero is a menu-settable value (100 microsecond units, but you could make them smaller).
We use it to ensure the single flash fires mid-way through the exposure and illuminates the scene for both cameras, thus avoiding double shadows.


David

Re: Accurate USB remote sync for single camera shots?
« Reply #25 on: 24 / May / 2008, 21:58:36 »
...and to clarify a bit: You can get better accuracy with CHDK compared to what you got from your experiments so far. You need to use the SDM build's sync feature. It was written to sync two cameras which get their triggers from a single USB remote wired to both cameras, but it can also help you to trigger a single camera more precisely. I don't know if it's compatible with the S3IS, though.

The delay from your remote trigger pulse to the shutter opening doesn't actually get any shorter using the sync feature, but (unlike in the standard Allbest build) the delay will be pretty constant (to almost 50 us for some cameras, quoting Microfunguy above, which could be enough for you to capture some nice images with a greater probability).

So, this will only help you if you can predict your plasma events by a couple of tens of ms or so (or whatever the delay is) i.e. if you for instance are actuating them electronically.


Thank you so much guys for all the swift responses. I first tried the SDM sync.bas on the s3is and was disappointed since I couldn't get it to trigger even as good as 10 miliseconds through the usb. Then I got a hold of a A700  and tested it with SDM sync.bas for triggering accuracy with 1/40000 sec shutter and WOW! I could trigger it down to 50 microseconds over and over again. Even down to 25 microsecond however reproducibility was only about 50 percent for this case. I am hoping to get nice plasma images and implement this into main diagnostics for the experiments.  I also borrowed a sd700 and will test that too.  What camera would you guys recommend for the fastest shutter speed and most accurate triggering ?

I did the triggering test with a photo diode and I had two independent triggers. For triggering A700 I used a 2 second long pulse. Although the camera is  in manual focus it opens its shutter all the way and does somethings then takes the shot at the end of 2 seconds. Is there a way to tell the camera to just take the shot which could make triggering more stable I guess.

Thank you so much again, I think the SDM software could catch on pretty quickly as a valuable scientific tool. There are already people asking me about this in our lab.











*

Offline BB

  • ***
  • 164
Re: Accurate USB remote sync for single camera shots?
« Reply #26 on: 25 / May / 2008, 00:30:43 »
Not quite an answer to your question on the 2 second delay--but a clarification... Canon P&S cameras use Dark Frame Subtraction for exposures longer than 1 second. So, for example, a 15 second exposure will be followed by another 15 second exposure with the shutter closed.

Normally, the camera does this with every shot--as this gives the most accurate dark frame (typically a purple "fog" from the sensor amplifiers--"amp noise". Will also remove hot pixels). Amp Noise is  typically related to the temperature of the sensor.

If you need better than 50% duty cycle for taking long exposure photographs (~2-60 seconds max), you can turn Noise Reduction (DFS) off in the CHDK menus. Or, use uBasic to take five shots and then a dark frame and post process it later.

-Bill

Re: Accurate USB remote sync for single camera shots?
« Reply #27 on: 25 / May / 2008, 05:10:10 »
Yes, the s2is and s3is are not recommended for precision stereo synch or precision delay to capture.
I cannot answer regarding shutter-speed but your tests suggest that it is 'real'.
So far, in terms of the difference in delay times between two cameras before shooting, the A630 performed best but this was a one-off test.
Some very inexpensive cameras with modest features (such as the A460) can have very repeatable delays after USB signal goes low.

Quote
it opens its shutter all the way and does somethings then takes the shot at the end of 2 seconds. Is there a way to tell the camera to just take the shot which could make triggering more stable I guess.

The whole point of this method is to let the camera do whatever pre-exposure operations it needs to and then pause the shoot operation until the USB signal goes low.
At that point, the camera presumably has very few other operations to do before firing the shutter.
So, the delay should be very consistent from one shot to the next.

In SDM you can programme a delay after USB low to a resolution of 100 microseconds.
We could increase that resolution to about 0.1 microseconds.

I assume that you used a fast-response photodiode ?

It will be interesting to hear of any further work.

Incidentally, SDM 1.70 is released today.


David


Re: Accurate USB remote sync for single camera shots?
« Reply #28 on: 25 / May / 2008, 05:13:05 »
Not quite an answer to your question on the 2 second delay

In this case, two seconds is the time the shooting-sequence is paused for, it is not the shutter speed.


David

Re: Accurate USB remote sync for single camera shots?
« Reply #29 on: 18 / June / 2008, 16:22:13 »
Yes, the s2is and s3is are not recommended for precision stereo synch or precision delay to capture.
I cannot answer regarding shutter-speed but your tests suggest that it is 'real'.
So far, in terms of the difference in delay times between two cameras before shooting, the A630 performed best but this was a one-off test.
Some very inexpensive cameras with modest features (such as the A460) can have very repeatable delays after USB signal goes low.

Quote
it opens its shutter all the way and does somethings then takes the shot at the end of 2 seconds. Is there a way to tell the camera to just take the shot which could make triggering more stable I guess.

The whole point of this method is to let the camera do whatever pre-exposure operations it needs to and then pause the shoot operation until the USB signal goes low.
At that point, the camera presumably has very few other operations to do before firing the shutter.
So, the delay should be very consistent from one shot to the next.

In SDM you can programme a delay after USB low to a resolution of 100 microseconds.
We could increase that resolution to about 0.1 microseconds.

I assume that you used a fast-response photodiode ?

It will be interesting to hear of any further work.

Incidentally, SDM 1.70 is released today.


David


I have taken some nice plasma images and decided to more work on characterization of the camera (A700). Since we have a data system with so many triggers I decided to use more fast diodes
to characterize the camera so I used 8, I sequentially triggered diodes with various pulses. After extensive testing the camera shutter speed came out to be 70-80 microseconds at 1/40000 setting, and the shot time was drifting and jumping. Sometimes tens of microseconds sometimes much more. I also tested this with various image quality aperture and zoom settings and the 70 -80 micro seconds was reproducible
for image size 2816 by 2112, normal resolution, and f 7.1 and 8 with no zoom. Since the trigger was
drifting to be able to know where the shot is taken, I use three photodiodes right on the lens that sequentially light up 100 microseond apart. Capturing the photodiode on the image simultaneously allows me to know what time the picture is taken. Its working pretty good good so far. If I test any other camera
that turns out to do better I ll post it here.
here is one of the images I took
Image:Picture 2.png - CHDK Wiki
Thanks




 

 

 

Related Topics