Sync Multi Cameras - 3d Modeling

waterwingz

Nestkast
Thanks now I know I don't have to use the inverted remote script anymore.

You probably should not need any script at all if you only want sync'd shots using a USB remote. I think that the menu options cover what most people need/want to do.
http://chdk.wikia.com/wiki/USB_Remote#Multi-Camera_Synchronization

Nestkast

There is one thing I don't get:
"If the Enable Sync option is set in CHDK then each camera will wait for the final transition of the USB signal ( 5V to 0V )."
It seems just to be the opposite, with Sync enabled my camera takes a shot immediately when it gets 5V on the usb port and as reslult the camera's are not in sync. (elph300hs, one push, normal control mode)

Whith sync disabled it works as I would like it takes the shot as soon as the 5v is switched off, just like the inverted remote script.

waterwingz

Nestkast
There is one thing I don't get:
"If the Enable Sync option is set in CHDK then each camera will wait for the final transition of the USB signal ( 5V to 0V )." It seems just to be the opposite, with Sync enabled my camera takes a shot immediately when it gets 5V on the usb port and as result the camera's are not in sync. (elph300hs, one push, normal control mode)
With sync disabled it works as I would like it takes the shot as soon as the 5v is switched off, just like the inverted remote script.

Strange. Just tested it (for probably the 1000th time) on my SD940. Works as expected - with or without sync enabled it takes the shot when 5V goes away. I also enabled a 5 second sync delay to confirm it goes through the sync code.

So I checked the source for the ixus220_elph300hs and it appears the necessary hook for sync shooting is not enabled for that camera. The hook to wait_until_remote_button_is_released is not implemented in capt_seq.c for any of the firmware versions. The code hook to disable noise reduction is missing too.

This means you can still try to take pictures with sync disabled but the accuracy of the sync will be suspect.

This is probably fixable. I won't have time to look at it for a few days (unless someone else volunteers) but I'll need somebody willing to test the fix. Like you ? (if so, what firmware version do you have ?)

Update : it seems somebody already noticed this : https://chdk.setepontos.com/d/6341/754 :😀

Nestkast

It would be very nice if this can be fixed! And of course I am willing to test, I have camera's with firmware version 101E and 101G. And waiting a couple of days is no problem. Thanks for your help.

waterwingz

Nestkast
It would be very nice if this can be fixed! And of course I am willing to test, I have camera's with firmware version 101E and 101G. And waiting a couple of days is no problem. Thanks for your help.

If you PM me a "real" email address I'll send you a message directly when I'm ready for testing.

srsa_4c

waterwingzThe code hook to disable noise reduction is missing too.

I had a failed try on this (see this and the followup posts), I hope you'll have more success 😀

David Ripple

waterwingz and microfunguy, good point about the accuracy of the measuring method. I will probably get a cheap CRT system to test on. However, I don't think I'll need sub-millisecond accuracy. 1-10 milliseconds should work. But I still have to test the multi camera rig system with 123d Catch to make sure.

I'm changing my camera to the Elph 100HS because it has good image quality. I just got 6 of them, so I should test Sync sometime soon.

I hope I don't get the same problem as the Elph 300hs that Nestkast has.

After I run some test on the Elph 100hs, I'll share my findings.

Nestkast

I would also be interested in your findings. And perhaps I should also look for a cheap CRT monitor. I hope Waterwingz can fix the sync problem whith the Elph 300hs.

waterwingz

David Ripple
I hope I don't get the same problem as the Elph 300hs that Nestkast has.

The sync code in capt_seq.c for the Elph100 is there and appears to be correct so you should be fine.

waterwingz

Nestkast
It would be very nice if this can be fixed! And of course I am willing to test, I have camera's with firmware version 101E and 101G. And waiting a couple of days is no problem. Thanks for your help.

links posted here : https://chdk.setepontos.com/d/6341/762

David Ripple

waterwingz, I greatly appreciate how helpful and quick to respond you are. Thank you!

I tested 5x Elph100s (the 6th camera arrived in the mail broken). They all synced at a minimum of 1/30th of a second (based on camera screen refresh). I still need to set up a more accurate time measurement tool. However, this is great news, and I will further my tests.

David Ripple

Currently, I'm triggering the shutter by plugging in and unplugging the power of two daisy chained USB hubs. The problem is when I leave the Elph100s plugged in and the power connected to the hubs, after about 20 seconds it will automatically trigger the shutter. This does not happen on the SD800s. Any ideas how to make it so it does not trigger automatically?

waterwingz

David Ripple
Currently, I'm triggering the shutter by plugging in and unplugging the power of two daisy chained USB hubs. The problem is when I leave the Elph100s plugged in and the power connected to the hubs, after about 20 seconds it will automatically trigger the shutter. This does not happen on the SD800s. Any ideas how to make it so it does not trigger automatically?

There is a "hard coded" 20 second timeout in the code for sync mode.

In sync mode, if the shutter stays "held" too long (20 seconds) then it gives up on waiting for the sync to complete on the assumption that something is wrong. That was done because the sync loop basically freezes out almost all other camera functionality - making it look like the camera is "dead".

I guess in your case you want to be able to wait longer after setting focus & exposure for the actual event that triggers the shot to occur ?

The SD800's should perform exactly the same way - unless you are running the old 1.0.0 version of CHDK on them.

David Ripple

Yes, It would be great if I could override the hard coded feature. There will be many times that the model will be trying to strike a certain poses and it takes longer then 20 seconds. It would be nice to not have to reset the cameras when this happens.

waterwingz

David Ripple
Yes, It would be great if I could override the hard coded feature. There will be many times that the model will be trying to strike a certain poses and it takes longer then 20 seconds. It would be nice to not have to reset the cameras when this happens.

I rebuilt the old 1.0.0 release and tested the timeout in sync mode. With that version the timeout existed but rather than being 20 seconds it was only 10 seconds!

We need to think about what to do here. I guess that I could just eliminate the timeout - not sure if there would be any bad side effects. I don't really want to add a CHDK menu choice to enable/disable the timeout or set its duration.

Microfunguy

David Ripple
I'm triggering the shutter by plugging in and unplugging the power of two daisy chained USB hubs. The problem is when I leave the Elph100s plugged in

That is a very curious way of triggering the cameras that is just asking for problems.

It is not recommended for a number of reasons, it is also not necessary.

SDM users of multicam rigs just use a regular USB switch and a number of gentStereo repeaters :-

http://www.gentles.ltd.uk/gentstereo/sdm.htm

David

waterwingz

Microfunguy

David Ripple
I'm triggering the shutter by plugging in and unplugging the power of two daisy chained USB hubs. The problem is when I leave the Elph100s plugged in

that is a very curious way of triggering the cameras that is just asking for problems.

Now that you mention it, I was wondering about that too. You would really need to look at the waveform of how the power decays using an oscilloscope on the USB line to the camera to determine if you are getting a sharp cut-off or a slow decay. The later case will surely cause variations in sync quality.

Microfunguy

David Ripple
I still need to set up a more accurate time measurement tool.

You will have to point all five cameras at a CRT.

By definition, multicam rigs cannot provide precision synch.

For example, with the S95 75% of shots with TWO cameras have a synch difference of one msec or less.
However, 25% have errors a lot more than that up to 8 msec.

David Ripple

I bought the Gentles solution and used it last night for the first time. For my purpose, and with only 5 cameras both gentles and USD daisy chain provide the same result. Keep in mind with 3d modeling software, I need probably only about 10 msec accuracy.

The gentles is nice because of the easy shutter button. Also, I think as I add cameras it should be better with the power decay. But the only way to know is to test it. Or if someone else has done the test.

The USB daisy chain seems to be better when it comes to all the cameras always taking the picture. In my limited tests with Gentles I found that sometimes if I hit the shutter button at different time intervals, some of the cameras activate the shutter and some do not. I also noticed that sometimes after triggering the shutter at odd time intervals. i.e. shoot, wait 12 seconds - shoot, wait 5 seconds - shoot, wait 2 second - shoot, etc. it makes the shutter speed setting messed up. Meaning, I can no longer set the shutter to 1/30. It does not even have that option in the menu. I have to restart the camera a couple times or re-install the CCHDK2.CFG file in order to have all the shutter speed options again. Having all the cameras always shot the same amount of pictures is very important for my naming convention. Any insight into this would be appreciated.

waterwingz - maybe it would be good to make the timeout 2 minutes.

waterwingz

David Ripple
I bought the Gentles solution a couple weeks ago and received it several days ago and used it last night. For my purpose, and with only 5 cameras both gentles and USD daisy chain provide the same result

So far, so good.

The gentles is nice because of the easy shutter button. Also, I think as I add cameras it should be better with the power decay. But the only way to know is to test it. Or if someone else has done the test.

If they did, they have not posted it. Far too common a human failing once one has the answer I fear. Let us know what you find out please ?

The USB daisy chain seems to be better when it comes to all the cameras always taking the picture.

Okay .. well that's strange.

In my limited tests with Gentles I found that sometimes if I hit the shutter button at different time intervals, some of the cameras activate the shutter and some do not.

Hmmm .. you have all of them setup as "one press" and "normal" ? And do they all go into "waiting for sync" mode when you press the button (lcd off and sync led lit) before you release USB power ? If they all do not go there then that is very strange.

I also noticed that sometimes after triggering the shutter at odd time intervals. i.e. shoot, wait 12 seconds - shoot, wait 5 seconds - shoot, wait 2 second - shoot, etc. it makes the shutter speed setting messed up. Meaning, I can no longer set the shutter to 1/30. It does not even have that option in the menu.

Trying to follow you here - its just not possible for the USB remote to change that sort of thing. Are you telling me that if you go to the CHDK "Extra Photo Operations" menu then the "Override Shutter Speed" menu item is not there ?????

I have to restart the camera a couple times or re-install the CCHDK2.CFG file in order to have all the shutter speed options again.

So multiple resets fixes it ?

« Previous Page Next Page »