GH1 Stereoscopic Rig Collaboration
- Thread starter David Cole
- Start date
David Cole
Well-known member
The LANC Shepherd manual says, "As long as the shutter sync is less than about one tenth of the shutter speed, disparity of the stereo images related to shutter sync will not be noticeable. So if the camera selected shutter speed is 1/100, then a shutter sync displayed on the LANC Shepherd of 1/1000 or smaller (say 1/2000, 1/4000, etc) will result in images with no noticeable motion disparity."
I'm not sure about that. When using a LANC Shepherd rig, I usually recycle power if the disparity is greater than .05 milliseconds.
In the case of the GH1 - we're getting either damn close to 0 disparity (as close as one can tell from the sync tester software), or, a gross error that is as much as 20 milliseconds.
David Cole
Well-known member
Which buttons? We'll test this immediately!
Can you repeat the experiment several times? Each time check with oscilloscope the voltages as you were doing before (just to check if waking-up appeared at the same time) and report about synch results from Wimmer's Synch Tester? I think that about 20 measurements should be enough to get some relevant statistical data.
It would be great if you can also make one longer measurement to see if time drift is constant or it changes by time.
g.l
Well-known member
That's actually what I said (I know, language barriers
).They will drift, the question is only how quickly. I've heard Lanc Shepherd users say that they can shoot ~30mins until the drift becomes a problem, but this no doubt depends on the model, temperature etc.
This is very interesting! Please, make more measurements and report about results.
David Cole
Well-known member
Those are clearly on the "nice to have" list. I'm sure we've all used rigs where the cameras have to be independently configured, or are controlled via external means (focus and zoom controllers, for instance). Of course, once you put the rig in an underwater housing, "nice to have" becomes damn nice to have.
One problem that we have now ...that may be a GH1 noob thing - is that the focus seems to "rebound" when we power-cycle. That is to say, the cameras have to be refocused after each power cycle. That is a pretty big pain in the butt. Any ideas on that ?
David Cole
Well-known member
I have experience with a few older tape-based LANC cameras, that drifted over time. My recent experience has been with Sony HDR-CX550v camcorders. If you get a good start from the LANC Shepard (or other controller), they seem to stay in tight sync as long as you are recording. I suppose that makes sense, as the standards for frame rates aren't terribly tolerant of drift.
Somewhere I read that composite out can be get by holding down trash button for 10 seconds whilst LCD is open. Note that it won't work from the moment on you start shooting. So it may be used for checking if the cameras are in synch (at least it was my idea how to check synch before shooting).
g.l
Well-known member
Here's some data from a Lanc Shepherd user:
"All these devices work on the same principle <snip>, that two camcorders of the same model will power up very close to one another, and because the camcorders use fairly precise crystal oscillators for master clocks that are used to generate the syncs, the cameras will maintain the initial low skew between the syncs for fairly long time. The power needs to be cycled by the controllers first, to get the initial low skew. Right after power-up, the syncs are free running, and all internal functions, including the sensor samplings are slaved to the vertical syncs. The trick is to get the syncs as close as possible together in phase in the first place. The actual activation of the record button is fairly irrelevant, because the response (recording), will be slaved to the syncs"
So that would be a global sync that begins at boot start - if this is true for the GH1 as well, then the boot up sequence is the critical time, and the memory cards may have something to do with it. After all they may be scanned early for the techs to load diagnostic programs (recovery from bad flashes etc).
More:
"I measured on my camcorders the syncs better than 500 microseconds of mismatch for the first 2 minutes after power-up, and within 2 milliseconds for about 20 minutes, after simultaneous power up. After that you simply cycle the power with the LANC controller to get the initial alignment back."
"All these devices work on the same principle <snip>, that two camcorders of the same model will power up very close to one another, and because the camcorders use fairly precise crystal oscillators for master clocks that are used to generate the syncs, the cameras will maintain the initial low skew between the syncs for fairly long time. The power needs to be cycled by the controllers first, to get the initial low skew. Right after power-up, the syncs are free running, and all internal functions, including the sensor samplings are slaved to the vertical syncs. The trick is to get the syncs as close as possible together in phase in the first place. The actual activation of the record button is fairly irrelevant, because the response (recording), will be slaved to the syncs"
So that would be a global sync that begins at boot start - if this is true for the GH1 as well, then the boot up sequence is the critical time, and the memory cards may have something to do with it. After all they may be scanned early for the techs to load diagnostic programs (recovery from bad flashes etc).
More:
"I measured on my camcorders the syncs better than 500 microseconds of mismatch for the first 2 minutes after power-up, and within 2 milliseconds for about 20 minutes, after simultaneous power up. After that you simply cycle the power with the LANC controller to get the initial alignment back."
Last edited:
Does "rebounding" happens in Manual focus mode or with manual lenses?
Last edited:
g.l
Well-known member
Drift is a known problem, because the crystals are not 100% accurate, there is always some tolerances in the components, quality, temperature etc. The same is true with audio recorders & sound cards. eg. when you record from multiple soundcards, their clocks need to be synced or the minor differences will corrupt recordings.
If you think about it, it doesn't matter if your camcorder is capturing slightly slower or faster than (say) 24fps. The playback will happen at 24fps regardless, so it's always within the standard. In effect your recording is speeded up or slowed down slightly during playback.
David Cole
Well-known member
We have made over 100 iterations of this test and simply cannot find any correlation between sync and the remote pins 3&4 power pulse (as per the scope pict).
We HAVE seen some interesting variation on the USB data line, however. These values are in microseconds:
79.858
53.239
88.731
88.731
79.858
106.447
70.985
79.858
97.604
70.985
301.686
62.112
70.985
97.604
Interestingly, no mater how we rig the test hardware, the LEFT camera always proceeds the RIGHT. Unfortunately, these values do not seem to correlate with sync at all. We'll rerun w/ the sync values to share.
It would be interesting to look at that vertical blanking pulse from the composite video line. Any idea how to get that signal pumping out of the camera?
David Cole
Well-known member
it happens in manual focus mode. Unfortunately, we don't have any adapted lenses here at beampath. We have the kit lenses, Olympus 9-18mm and Panny 20mm pancakes (all micro 4/3 lenses).
I can't imagine it would happen with an adapted lens... as there is no electrical connection between the camera body and lens, right?
g.l
Well-known member
I think that tells us that the unwanted delay is caused later in the boot sequence. What would be helpful from Vitaliy is the exact component initialisation sequence at boot.
Although it would also be cool to have board-level sync and just be done. Maybe we should start a 3D hack fund, so we can buy two bodies (preferably beaten up and cheap) and experiment.
Similar with LANC signal. It cannot be compared with video out signal which has "perfect" timing.
These values are very good for video (1/50s or 1/60s shutter speed). I suppose that those numbers are missynch measurements taken from the CRT screen?
Have you tried with trash button as I mentioned before?
David Cole
Well-known member
Sorry for the confusion. To clarify - the set of timings that I posted was the offset between the power pulses on the data line of the GH1's USB ports.
I forgot to mention that bias can be easily solved by powering one camera slightly ahead of the other (in your case by approx. 80ms) as is done in LANC Shepherd or 3D LANC Master.
What happens if you detach the lenses (and allow shooting without lenses in menu)? Does timing change?
Added: Moreover, have you checked if both cameras have the same settings in menu?
Last edited:
OK. When you have time, give us both results together: offset between USB power pulses and actual difference between both cameras (by using Wimmer's Synch Tester).