CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD - page 52 - Creative Uses of CHDK - CHDK Forum

CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD

  • 704 Replies
  • 174099 Views
*

Offline SticK

  • *****
  • 779
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #510 on: 07 / August / 2013, 09:21:58 »
Advertisements
S90_101a NOVIGNCORR RAW ANALYSIS in RawTherapee

These images were acquired under the same conditions as in my previous post edit.  Images are neutral-processed and zoomed 100%.  Fig 1 is the camera JPG for reference.  Average intensity is ~115/256.  Fig 2 left is the Canon CR2 and right is CHDK DNG (no badpixel correction so bad pixels are visible).  Both RAW images are essentially the same and follow the frame intensity uniformity established in my previous post.  Average intensity of both RAWs is ~105/256.  The ~10% intensity difference between the camera JPG and the RAW images is very acceptable.

Thus now we have thorough validation of all three standard image formats.

*

Offline SticK

  • *****
  • 779
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #511 on: 07 / August / 2013, 09:39:20 »
NOVIGNCORR and CHDKPTP LIVEVIEW

@srsa_4c

It seems to me that novigncorr in liveview might behave similarly to nodistcorr.  The figures show liveview with novigncorr enabled and disabled respectively, and naturally the correction is identical in both.  Although liveview is not mission-critical as was imaging, uncorrecting liveview would help greatly by taking advantage of CHDKPTP's high-res liveview capability for real-time instrumentation alignment.

Would this be relatively easy to do, with the hope that nodistcorr experience could possibly help ?

*

Offline srsa_4c

  • ******
  • 4451
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #512 on: 07 / August / 2013, 15:38:32 »
Would this be relatively easy to do
Unfortunately, not. The still image related corrections were relatively easy to find (mostly due to descriptive strings). The live view related routines are quite complex, and there's a lot of them. This correction was not designed to be switched off, there are (IMHO) no event procedures that could help as with the distortion correction.

*

Offline SticK

  • *****
  • 779
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #513 on: 07 / August / 2013, 16:16:00 »
Quote
The live view related routines are quite complex, and there's a lot of them.
Did you find a similar quagmire for your liveview distcorr hack?  That hack was very important and I thank you sincerely for it because would have introduced positional distortion for step-and-repeat tiling alignment operations, essentially rendering the entire solution useless for mosaic imaging.  Uncorrecting vignette in liveview on the other hand is not so important and I can get by without it.  But if you think you can manage it, it would be very welcome.


*

Offline srsa_4c

  • ******
  • 4451
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #514 on: 07 / August / 2013, 16:19:06 »
Did you find a similar quagmire for your liveview distcorr hack?
Not one, but two event procedures offer control over that, it was quite easy to spot.

*

Offline SticK

  • *****
  • 779
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #515 on: 07 / August / 2013, 18:08:43 »
The live view related routines are quite complex, and there's a lot of them.
Thank you for looking.  Please know that every one of your efforts is greatly appreciated, whether successful or unsuccessful, and your expertise is held in very high regard here.  Your novigncorr is a glorious final solution. 

Hence I can now say the entire system solution has been raised to a completely different level.  I officially declare that we now have a complete functional solution for the S90 Primary Focal Plane Configuration.  :)

There will be a few more minor (I am guessing) but useful hacks I will ask for, probably more in reyalp's department.  Also of interest will be moving to V1.2, since direct USB file transfer is implemented there and I am excited about exploring that new material on the S90.

To everyone else following this recent (likely boring to most) part of the thread ...

Please stay tuned ... soon I will be posting my findings on temperature detection, and in a week or two, to complement srsa's latest elegant hack, I am introducing a new advanced MCU-driven hardware solution that handles external exposure control, eliminating the need for a hardware shutter when the camera is used in dark environment.  Most scientific instrumentation configurations (like mine and aggsquirrel's) fall into this usage class.
« Last Edit: 07 / August / 2013, 19:46:01 by SticK »

*

Offline SticK

  • *****
  • 779
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #516 on: 09 / August / 2013, 07:52:59 »
CAMERA ORIENTATION WORKAROUND ?

The PCB of the 101a when installed in my instrument will be "sideways."  In the 101a srsa builds (including the latest), for some reason the camera orientation icon remains fixed not responding to changes in orientation, in MANUAL mode.  Until a couple of weeks ago the icon was stuck in the normal orientation for many months which was fine, and when I attached the other CCD to search for the on-board thermistor, the icon went sideways, and now it is stuck there resulting in vertical view images.  I didn't notice it until I had re-assembled the imager (put the PCB back with its own CCD).  edit: This could also be related to electrical hacked state of the 101a because changing CCDs made the orientation icon react while the camera was in permanent M mode.  You can see it in the last CHDKPTP screen shots a few posts ago.

While still on the bench, I can get the icon to revert back to normal orientation by going to AUTO, rotating the camera, and then back to M mode.  That fixes the icon, but only while it's powered.  If I PDN/PUP, the icon reverts to the sideways orientation. 

My guess is that orientation must be an EXIF flag or rotation matrix.  The problem is that since I will not have the option of doing the temporary re-orientation trick when the camera is installed in the instrument, is there a way with current CHDK to force orientation to always be "normal" ?  If not, can the srsa build be modified to accept an orientation suppression flag, or be put into permanent normal state ?
« Last Edit: 09 / August / 2013, 09:03:44 by SticK »

Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #517 on: 09 / August / 2013, 09:08:35 »
soon I will be posting my findings on temperature detection

What about the origin of life, that is the big one ?

Also, a DIY kit of parts  :)


*

Offline SticK

  • *****
  • 779
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #518 on: 09 / August / 2013, 09:54:24 »
@Microfunguy

So nice to hear from you! 

Quote
What about the origin of life, that is the big one ?
Yes, always in the forefront.  I very excited about the new imager ... still .. a bit of time to round out corners  before it is actually installed for exploration to begin.

Quote
Also, a DIY kit of parts
Hmmm ?

*

Offline srsa_4c

  • ******
  • 4451
Re: CHDKPTP: S90 Primary Focal Plane Configuration - hacking out the CCD
« Reply #519 on: 09 / August / 2013, 12:15:27 »
now it is stuck there resulting in vertical view images.
I don't think any of my hacks could have influenced that sensor. Don't know what kind of image processing / viewing software are you using. Have you tried to make them ignore the orientation flag in exif? For example, IrfanView has an option for that.
My first try to override the orientation flag propcase failed, it's either not writable, or is sampled in a different firmware routine.

 

Related Topics


SimplePortal 2.3.6 © 2008-2014, SimplePortal