sync'ing 3D video - page 7 - Creative Uses of CHDK - CHDK Forum supplierdeeply
« previous next »
Pages: 1 2 3 4 5 6 [7] 8 9 10 11   Go Down

sync'ing 3D video

  • 103 Replies
  • 37623 Views
*

Offline waterwingz

  • ******
  • 12140
    • Waterwingz
Re: sync'ing 3D video
« Reply #60 on: 29 / June / 2013, 01:33:26 »
Advertisements
Wow.  I'm clearly from the wrong planet here.
Logged
Ported :   A1200    SD940   G10    Powershot N    G16
*

Offline ahull

  • *****
  • 634
Re: sync'ing 3D video
« Reply #61 on: 29 / June / 2013, 08:05:11 »
@Hardware-hacker - not sure I followed that either, any chance of citing some detailed references?
Logged
*

Offline Microfunguy

  • ****
  • 4037
    • StereoData Maker
Re: sync'ing 3D video
« Reply #62 on: 29 / June / 2013, 19:16:55 »
Quote from: Hardware_Hacker on 29 / June / 2013, 01:22:51

A 625x improvement in "incremental" comparison speed can be achieved just by using Horizontal sync for [Pal-Sd-Tv] for #1/#2/A/B.

How do you know which lines the horizontal synch pulses are for ?

Your posts are very interesting but very difficult to understand.

Please try to explain things more clearly.
Logged
*

Offline Microfunguy

  • ****
  • 4037
    • StereoData Maker
Re: sync'ing 3D video
« Reply #63 on: 29 / June / 2013, 19:21:23 »
Quote from: waterwingz on 28 / June / 2013, 22:34:58
Hey,  the 1970's are calling and they want their 2N2222's back.

I ended up configuring an LM358  (because  I had one) as a dual Schmidt trigger and connecting the back-to-back led's across the outputs.


I pressed the remote until the led's were at minimum brightness and then checked with a 'scope.
They signals were in-phase.
I will next make the circuit more transportable before testing outdoors.

Logged
*

Offline waterwingz

  • ******
  • 12140
    • Waterwingz
Re: sync'ing 3D video
« Reply #64 on: 29 / June / 2013, 19:34:28 »
Quote from: Microfunguy on 29 / June / 2013, 19:21:23
I pressed the remote until the led's were at minimum brightness and then checked with a 'scope.
They signals were in-phase.
I will next make the circuit more transportable before testing outdoors.
Microcontroller solution becomes the easy part of this exercise.
Logged
Ported :   A1200    SD940   G10    Powershot N    G16
*

Offline Hardware_Hacker

  • ****
  • 322
Re: sync'ing 3D video
« Reply #65 on: 30 / June / 2013, 00:21:08 »
Quote from: ahull on 29 / June / 2013, 08:05:11
@Hardware-hacker - not sure I followed that either, any chance of citing some detailed references?

Yes & some Graphics

in a day or so.

Happy-Hacking By  H-H

but please keep up the good work for sync'ing 3D video.

Every "Sync" method will have, some, advantages and dis-advantages. 


Logged
*

Offline Hardware_Hacker

  • ****
  • 322
Re: sync'ing 3D video
« Reply #66 on: 30 / June / 2013, 01:15:07 »
Quote from: Microfunguy on 29 / June / 2013, 19:16:55
Quote from: Hardware_Hacker on 29 / June / 2013, 01:22:51

A 625x improvement in "incremental" comparison speed can be achieved just by using Horizontal sync for [Pal-Sd-Tv] for #1/#2/A/B.

How do you know which lines the horizontal synch pulses are for ?

Your posts are very interesting but very difficult to understand.

Please try to explain things more clearly.

The post once again was not ment to be a practical soloution.

It was ment to just point out that speed is a critical issue.

Revisting the "Rocket Forum"

Both Rockets would have crashed.
Because of various System Design Flaws.
So Back to the drawing board.

Revisting the "Motor Example"

Absolute [Vert_Sync] Concept = Absolute [Encoder] Concept.

David, some time ago, you drew a "Sawtooth" Vert_Sync" diagram, please re-post it here.

Incremental [Horizontal_Sync] Concept = Incremental [Encoder] Concept.

This works like a set of Vernier Calipers:-

Absolute [Vert_Sync] always is 0.00000000000

David, Please draw  a similar "Sawtooth" Horizontal_Sync" diagram, please post it here.

Ignore interlacing to make it simple.

Just two in simple bits say line 0 to line 4 and same to line 625.

Because of the fixed number of Tv_Lines it has a similar "Sawtooth"  shape.

So lets say:-

Absolute [Vert_Sync] Reference #1 = A_#0 = HOME

Absolute [Vert_Sync]     Camera #1 = A_#1

Absolute [Vert_Sync]     Camera #2 = A_#2

Our new Absolute [Vert_Sync] aim is attempt to make:-

A_#0 = A_#1

AND

A_#0 = A_#2

Also we need to "Track" Incremental [Horizontal_Sync]

Please note that in the Motor example:-

It doesn't just do ONE REVOLOUTION it could do say 10,000 Revs, Forward and Back.

The Incremental [Encoder] keeps track of Direction/Revs.

Additionaly the Absolute [Encoder] keeps track of Direction/Revs AND the HOME location.

But at 1/625 of the count i.e. 1rev. = 1count.

"Tripple Difference" [tracking] Concept:-

1/ Absolute [Vert_Sync] Camera #1, Camera #2, = HOME = Ref#0

2/ Incremental [Horizontal_Sync] Phase_Advance/Retard for Camera #1,

3/ Incremental [Horizontal_Sync] Phase_Advance/Retard for Camera #2,

Hey!! this seem to be Sync-Mk2x

IT IS BUT at 625 times as FAST.

BUT it can GO EVEN FASTER  ,,,,,,,,,,,,,,,,,,,   may be in the next post!

You still need to consider Camera-Key-Polling, Mode-Delay, USB-Jitter and Camera-Lag for pseudo-stable 3D Video sync/gen-lock.

Happy-Hacking By  H-H

but please keep up the good work for sync'ing 3D video.

Every "Sync" method will have, some, advantages and dis-advantages.
« Last Edit: 30 / June / 2013, 01:27:17 by Hardware_Hacker »
Logged
*

Offline thepanoguy

  • ***
  • 138
Re: sync'ing 3D video
« Reply #67 on: 30 / June / 2013, 08:56:11 »
Logged
*

Offline Microfunguy

  • ****
  • 4037
    • StereoData Maker
Re: sync'ing 3D video
« Reply #68 on: 01 / July / 2013, 09:59:28 »
Quote from: waterwingz on 29 / June / 2013, 19:34:28
Microcontroller solution becomes the easy part of this exercise.

Let us say that I can create two 24 Hz square waves  (with resolution of 0.1 msec) that accurately represents the phase of the movie capture 'clock'.
I can drift these waves into synch.
Just before perfect synch, one wave starts 0.1msec before the other.

Is it possible to detect this accurately using a very small microcontroller ?

Could the input from the left camera generate an interrupt which then runs a routine that constantly checks for the right input going low and increments a counter ?
If that count is identical to left camera period count they are exactly in synch.


It should also indicate in some way, with one or more led's, that we are getting closer to perfect synch as well as indicating perfect synch.

Input voltages could be scaled to suit the controller which should be suitable for 12V or lower.
« Last Edit: 01 / July / 2013, 10:20:42 by Microfunguy »
Logged
*

Offline waterwingz

  • ******
  • 12140
    • Waterwingz
Re: sync'ing 3D video
« Reply #69 on: 01 / July / 2013, 13:24:11 »
Quote from: Microfunguy on 01 / July / 2013, 09:59:28
Let us say that I can create two 24 Hz square waves  (with resolution of 0.1 msec) that accurately represents the phase of the movie capture 'clock'.
50% duty cycle?

Quote
I can drift these waves into synch.
Just before perfect synch, one wave starts 0.1msec before the other.
Is it possible to detect this accurately using a very small microcontroller ?
Should be no problem for even the cheapest ( sub $10) Aurdino clone.

Quote
Could the input from the left camera generate an interrupt which then runs a routine that constantly checks for the right input going low and increments a counter ?
If that count is identical to left camera period count they are exactly in synch.
I wouldn't bother with interrupts for something this simple.  Just poll both signals in a tight loop and  count the time between when one goes from 0 -> 1 and the other does the same.  Polarity is then a function of which one toggles first.

Quote
It should also indicate in some way, with one or more led's, that we are getting closer to perfect synch as well as indicating perfect synch.
As I mentioned earlier,   I'd use two LED's.  Slow blinking for way out of phase, faster blinking as phase error approaches zero,  full on at zero.  Phase polarity determines which LED is blinking but they will both be full on when in sync.


Quote
Input voltages could be scaled to suit the controller which should be suitable for 12V or lower.
The digital inputs on typical controller cards look for 0-5V or 0-3.3V depending on the uP Vcc value.


But more interesting to me is automating this. 

Using the logic above, we can use the external controller to generate a continuous USB 5V PWM Digital output signal where a 100 mSec pulse = in phase,  0 = out of phase negative,  200 mSec = out of phase positive, and anything in between is just proportional to the sync error.

We could then use a script ( Lua or uBASIC ) to create a control loop and tweak the register you were playing with based on that signal.  Or add some C code to the keyboard task.  That would give you a feedback loop that would quickly lock sync and adjust for any drift.
Logged
Ported :   A1200    SD940   G10    Powershot N    G16
Pages: 1 2 3 4 5 6 [7] 8 9 10 11   Go Up
« previous next »
 

Related Topics
SimplePortal 2.3.6 © 2008-2014, SimplePortal