GH1 Stereoscopic Rig Collaboration

Very very nice inoue_k3D. I'm building similar stuff, just got my 1st 14mm prime and started doing test shoots, batteries are here etc.

I have an idea of modding a drawer runner (ball bearing type) for a smoothly-adjustable side-by-side mount. There are lots of possibilities there, including telescopic ability for interaxials >= 1m. I've already taken one apart and it's looking promising.

I'll also need to build a mirror box. I wondering, isn't it more convenient to do a horizontal setup (both bodies horizontal, but 90 degrees rotated)? I know inoue has has built this type in the past. It seems easier to make exposure changes that way.

I'm also trying to decide on my monitoring options for a stereo monitor, either 2 720p DIY LCD panels, or just bog-standard 800x480 7" screens. The problem with the 720p panels (apart from finding a small HDMI controller board, I may actually have one) is building a reliable but light-weight case for them.
 
Last edited:
Hi, gl and all

Here is a test 3D video shot on my new GH2 based system.
http://www.youtube.com/watch?v=qj17GoanNAY

The purposes of the video are;
1) Sync test using high magnification shooting of fast moving cherry blossoms as subject
2) 3D Telemacro shooting using 100-300mm Zoom lenses with EX Telecon mode (up to 1560mm equivalent) where the working distance was more than 2m.

I couldn't recognize any missync in these clips.
Any comment will be much appreciated.

P.S. to gl
I like horizontal BS configuration, but it requires larger mirror in case of wide focal length especially at 16:9 aspect ratio.

inoue_k3D
 
inoue_k3D said:
I couldn't recognize any missync in these clips.
Any comment will be much appreciated.
Click to expand...

I can see missynchs occasionally. Maybe the most noticeable is at 1:55 (see the flower on the right part of the picture - moving with a wind). Can you give us the picture of your 3D setup?
 
I have a few questions about your findings with M4/3 lenses, as I've had no luck with 100% reliable sync using M4/3 electrically connected lenses. Hope you don't mind a barrage of questions:

1) do you power-up, allow the lenses to calibrate, then power-cycle again?
2) do you focus AFTER the second power-cycle?
3) do you ever have ann initial offset with the Panny 14-42mm lenses attached?
4 )are you compensating for any consistent initial offset using adaptive power-up delay?
4) how close does Vsync correlate with Wimmer's Sync tester results?
5) does auto - exposure work ??

I think MOST users would prefer to use Panny lenses, if there is a way around the seemingly non-deterministic initial offset that seems to correlate with "power-up lens motor controller calibration".

Thank you for your help with this!!!

-- David
 
Can anyone help me select the right kind of MOSFET for switching the body power lines, and also the remote resistances, under microprocessor control?

I thought it'd be easy, but I've been Googling and reading datasheets for 2 days now, and my head is spinning.

AFAIKS, I need a low-side driver that accepts CMOS logic levels. And 'low-side' means I have to place it after the load, near ground correct?

Or is there a simpler type of solid-state switch for this stuff?
 
David Cole said:
This is the part that I used (however, I really can't remember WHY I selected it) : STP12PF06 data sheet is here: http://www.datasheetcatalog.org/datasheet/SGSThomsonMicroelectronics/mXuvwwt.pdf
Click to expand...

Thanks, but I don't think I could tie this directly to my micros' output pins, as it seems to require 'input voltage + ~3-4V extra' to switch. I think this is a 'high-side' device right? Some low-side ones like this one can switch at 2.4V (or lower), so fine for logic levels.

Or am I missing something?
 
re: trigger voltage... I'm using with an Adruino board which offers standard TTL outputs (+5 volts when high). Sorry... missed your CMOS logic level requirement.
 
David Cole said:
re: trigger voltage... I'm using with an Adruino board which offers standard TTL outputs (+5 volts when high). Sorry... missed your CMOS logic level requirement.
Click to expand...

I see, well then it's official, I don't understand MOSFETs! :). Guess I'll do some more reading tomorrow.

BTW the remote line reads as 2.46V.
 
I don't understand them either (certainly I don't understand the spec sheets). I found the part I'm using in an Adruino application note. Apparently these are known as logic-level MOSFETs, presumably because they switch at TTL. (Check this out: http://www.nteinc.com/Web_pgs/LL_MOSFET.html)

Anyone else shed some light here?
 
Last edited:
David Cole said:
1) do you power-up, allow the lenses to calibrate, then power-cycle again?
2) do you focus AFTER the second power-cycle?
3) do you ever have ann initial offset with the Panny 14-42mm lenses attached?
4 )are you compensating for any consistent initial offset using adaptive power-up delay?
4) how close does Vsync correlate with Wimmer's Sync tester results?
5) does auto - exposure work ??
Click to expand...

1 & 2) I power up the cameras, wait a little until a red sign showing writing/reading to SD card vanishes then focus both cameras by half-pressing using external remote connected to my 3DSLR Master and then I lock focus and exposure (not necessary!). After that I start shooting.

3 & 4) I can't remember exactly what was initial offset, but I add some offset in order to be about 1.7ms out of synch after power up. Then the offset is starting to reduce. After power up I have always the same initial offset (I guess in more than 99% cases).

second 4) After my testings some time ago I could not see any differences between both. Maybe I should make testings again, but it's more or less consistent.

5) You mean if I left auto exposure? Yes, it works with auto exposure and with locked exposure.

By the way, have you ever tried with switching from "NTSC" to "PAL"?
 
g.l said:
Can anyone help me select the right kind of MOSFET for switching the body power lines, and also the remote resistances, under microprocessor control?
Click to expand...

I can't remember exactly which MOSFET I used (will have to check it), but I selected it in order to be able to switch the POSITIVE power on. So, it's probably not suitable for you, since you'd like to switch on/off the ground signal. However, be careful by switching on/off the ground signal, since you'll probably have connected grounds by remote signals. Please, check again before wiring...
 
Here is a colour-corrected version with added music (XView)...
The original files, when Vimeo takes away the highest resolution file, can be downloaded by using links given on Vimeo page.

 
Last edited:
Hi, Damir (#300)

Thank you for your sync check. I admire your checking ability!
I fixed the mis-sync's by means of frame shifting of n/60 Sec in Premiere CS5 time line and confirmed perfect sync in all clips.
This means recording of video frames are well synchronized in both cameras running independently each other after simultaneous power on just by using a single external battery.
I will need to use a clapper in case of no sound cue in order to ensure the post sync hereafter. Here is the corrected yt3d video.
http://www.youtube.com/watch?v=C4fv_XxTdKc

The photos of my twin GH2 rig were already shown at #300 i.e.,
http://v-galleries.com/images/Twin_GH2_Rig/index.htm

BTW, I saw your 3D video at VIMEO. Letting me know where can I find the details about your BS rig and 3DSLR Master system will be much appreciated.


Hi, David (#305)

I still do not fully understand what is your concern with the AF mode of the lenses on the 3D sync.
I am using five kinds of Panasonic M4/3 lenses with AF mode for 3D shooting and having an impression that AF process is not affecting the final 3D sync at all.

On the other hand, I have some finding that I have to shift more frames in post for the LR pair at the very first shooting just after power-on as compared the succeeding shots (normally zero or two frames shift). In both cases the consequent sync at the time line is same, just perfect.
I am not sure this is related to the auto-focusing process of your concern or not, I hope this observation may help you in anyway.

inoue_k3D
 
inoue_k3D said:
I still do not fully understand what is your concern with the AF mode of the lenses on the 3D sync.
Click to expand...

In over 100 sync tests I found that in approximately 30% of the cases where I was using M4/3 lenses (in manual focus mode), I had extreme initial offset (of 20 ms or more). To be clear, I do not mean a full frame of offset - which could be easily fixed - i mean intra-frame offset.

After speaking to a Canon lens engineer, I postulated that the cause might be that each camera might be initializing the positioning control mechanics in the lenses in a non-deterministic fashion and then starting the sensor integration clock. Also, Vitally stressed from the beginning that we should be working with manual lenses to avoid this find of issue.

Most compellingly, once I switched to full-manual lenses, my initial offset dropped below the resolution of the Wimmer CRT sync test that I was using - near 0 offset. I declared victory and stopped using M4/3 lenses.

Now we see that Crunchy is using M4/3 lenses quite successfully. I'm trying to rectify his results with my own. One potential difference might be that I never tried auto-focus and the "half-push" on the remote button. This might reset the sensor clock.

What do you guys think ?
 
With the greatest respect to Vitaliy (who is just amazing), I think he was wrong on several points early on in the discussion. In fairness, we're all just educated-guessing (or 'trying to make sense of the data' :)).

The auto-focus resetting the clock is actually possible! We know that the GH2 actually has a 120fps mode, which is used during focusing (that's how it got so fast) - and I think the GH1 has a 60fps focus mode. Neither modes are used during normal shooting of course, so the sensor mode would need to be changed temporarily, and this might or might not reset the clocks.

On the other hand, the sensor mode is probably changed all the time, eg. when switching from live view to recording, and (as far as we know) that doesn't affect sync, right? Or does it?

Also I don't know if the faster modes are used with video - maybe before you start recording, but obviously not during (when you refocus).
 
crunchy said:
By the way, have you ever tried with switching from "NTSC" to "PAL"?
Click to expand...

No, NTSC only. Are you testing in PAL only? Perhaps we should start a test result database to track these things? We could track, for example:

serial number distance, firmware, lenses, average min and max initial offset, average min and max 5 min offset, Wimmer to VSync correlation, PAL or NTSC, AVCHD vs MPEG, framerate, auto-focus used, etc.

Also, does your Vsync change when you press the half-press remote button for AF ?
 
inoue_k3D said:
I fixed the mis-sync's by means of frame shifting of n/60 Sec in Premiere CS5 time line and confirmed perfect sync in all clips.
Click to expand...

Now it is much better. However, have you checked clip around 1:16 and 1:27? Observe the clip e.g. in anaglyph form and check edges of the flower in wind. Otherwise it seems perfectly (or at least very-well) synched. :)

inoue_k3D said:
This means recording of video frames are well synchronized in both cameras running independently each other after simultaneous power on just by using a single external battery.
I will need to use a clapper in case of no sound cue in order to ensure the post sync hereafter.
Click to expand...

I don't need clapper. All my clips (if they show well synch of video signals during shooting) have the exactly same length in frames and they are synched (at least it seems so).

inoue_k3D said:
BTW, I saw your 3D video at VIMEO. Letting me know where can I find the details about your BS rig and 3DSLR Master system will be much appreciated.
Click to expand...

I am now working on my beam-splitter (it's practically finished). After that I'll continue developing 3DSLR Master. I'll have to add some features I'd like to have as, for example, time-lapse, manually changing parameters, etc.

inoue_k3D said:
On the other hand, I have some finding that I have to shift more frames in post for the LR pair at the very first shooting just after power-on as compared the succeeding shots (normally zero or two frames shift). In both cases the consequent sync at the time line is same, just perfect.
I am not sure this is related to the auto-focusing process of your concern or not, I hope this observation may help you in anyway.
Click to expand...

This is very interesting, indeed. Are you 100% sure that they are always perfectly in synch? You never got any shift? What happens with synch when you finish shooting and would like to continue with another shot without switching off the camera in between?
 
David Cole said:
One potential difference might be that I never tried auto-focus and the "half-push" on the remote button. This might reset the sensor clock.

What do you guys think ?
Click to expand...

This could be the reason (I think that using auto-focus before shooting is very important)!
 
You must log in or register to reply here.
Back
Top