IXUS 115 HS / ELPH 100 HS request - page 34 - DryOS Development - CHDK Forum  

IXUS 115 HS / ELPH 100 HS request

  • 538 Replies
  • 220945 Views
*

Offline dnw

  • *
  • 22
Re: IXUS 115 HS / ELPH 100 HS request
« Reply #330 on: 05 / July / 2012, 03:55:37 »
Advertisements
Swithich,

I don't know if Canon has regressed or not, all the cameras involve compromises of features of processing and costs. The IXUS 115 has some very good features and a good low noise sensor. But my camera cost me A$109.00. The lens I have for my Nikon D40 costs me A$220. It seems that good optics are much more costly that good electronics. Many camera review sites have sample photo's . You can search flicker by camera type (I think).


             I do not understand the problem you are proposing about raw output. Your camera takes a 'raw' image and it is 'developed' in the camera and saved as a jpg file.

If you output a raw image you then 'develop' it using software outside of the camera. The only difference is the software used. That said the software in the camera is very good and customised for that sensor and It takes some work and skill to achieve the same result manually on external software.

Also, If you then save the developed output in a lossless format like tiff of png, you escape the compression problem.

I use rawtherapee to 'develop' my raw images. One of rawtherapee's  features is that you can apply the same conversion profiles to set of images such as a panorama set, that is you minimise color differences between shots. If have used hugin successfully in this way.

Regarding super-fine JPG images - are you aware of the CHDK option Quality Override where you can select super-fine for this camera.

« Last Edit: 05 / July / 2012, 04:56:08 by dnw »
A530 & IXUS 115 HS Firmware version, 1.01C, CHDK status:   user

Re: IXUS 115 HS / ELPH 100 HS request - alpha9 with working RAW
« Reply #331 on: 05 / July / 2012, 08:08:38 »
Some progress at last.  Rewrote the boot.c and capt_seq.c code from scratch - modelled after the elph500 code. Its only for the 1.01b (and 1.01c) as nobody has come forward to work on the other models so far.

ixus115_elph100hs-101b-1.1.0-alpha9

dnw tested this last night and reports that RAW and DNG are now working.  Several of the shooting issues he has been seeing are corrected too.   Here are some questions though :

1) dnw reports that every second JPG is massively overexposed when using DNG. The corresponding DNG images are fine.  Apparently this is not a problem when just shooting RAW ?

2) Can someone test whether script issues with shoot vs press "shoot_half"/press "shoot_full" problems are fixed ?

Ported :   A1200    SD940   G10    Powershot N    G16

Re: IXUS 115 HS / ELPH 100 HS request
« Reply #332 on: 05 / July / 2012, 08:46:25 »
good for saving DNG

However, the problem is same as the previous alpha omega build

the JPG generated is sometimes showing kind of over-exposed, however, it is definitely the previous frame captured.

for DNG, if the JPG is showing that strange color, the DNG saved is exactly the same as the previous DNG file, only differ in the EXIF (the EXIF is stored correctly)

for example, if I captured 6 frame
CRW_0101.DNG (just when the camera turn on)
CRW_0102.DNG portrait orientation
CRW_0103.DNG landscape orientation
CRW_0104.DNG portrait orientation
CRW_0105.DNG
CRW_0106.DNG

of which 0101, 0103, 0104, 0106 are problematic
then 0101 JPG and DNG will be all white
IMG_0103.JPG will be strange color, but from the outline, I can see that it is compressed with 0102 information and will be showing with landscape orientation, because 0103 is shooted on landscape

0103 DNG is exactly the same as 0102 DNG

IMG_0104.JPG will be strange color, but from the outline, I can see that it is compressed with 0102 information and will be showing with protrait orientation, because 0104 is shooted on protrait

0104 DNG is exactly the same as 0102 DNG

0105 is normal, both JPG and DNG are good

IMG_0106.JPG will be strange color, but from the outline, I can see that it is compressed with 0105 information, 0106 DNG is exactly the same as 0105 DNG

same as alpha omega build, to solve this problem, press "playback" button during after shot review to play the image just captured, then capture another frame, all the JPG and DNG will be good.

« Last Edit: 05 / July / 2012, 10:04:45 by larrycafe »

Re: IXUS 115 HS / ELPH 100 HS request
« Reply #333 on: 05 / July / 2012, 08:54:38 »
However, the problem is same as the previous alpha omega build
I think its good new if we now have source code equivalent to the famous "alpha-omega" build !

Quote
the JPG generated is sometimes showing kind of over-exposed, however, it is definitely the previous frame captured.
for DNG, if the JPG is showing that strange color, the DNG saved is exactly the same as the previous DNG file, only differ in the EXIF (the EXIF is stored correctly)
I'd like to understand if this issue is related to just DNG or to both DNG&RAW.   Can you repeat the sequence with just RAW enabled and post results ?

Quote
same as alpha omega build, to solve this problem, press "playback" button during after shot review to play the image just captured, all the JPG and DNG will be good.
If I understand this correctly,  you are leaving shooting mode and going into playback mode after each shot ?  So in essence,  you are repeating the first step from the sequence you posted above each time ? (i.e. the first DNG & JPG are always good - by resetting each time you are essentially getting a "first" shot every time ?)

Ported :   A1200    SD940   G10    Powershot N    G16


Re: IXUS 115 HS / ELPH 100 HS request
« Reply #334 on: 05 / July / 2012, 09:53:53 »
If I am shooting like normal, that is
1. half press and shoot
2. the camera shows review
3. review ends
4. half press and shoot again

I get the problem frame randomly, could be 4 out of 10

if I am adding a playback between each shooting, that is
1. half press and shoot
2. the camera shows review
3. press playback button to enter playback mode
4. half press and shoot again

then I can almost guarantee I get good JPG and DNG

I think the problem is the same with CRW and DNG, however, I do not know what software to open the CRW, can you suggest so I can try? I think someone did a file compare before to compare the problematic CRW with the previous CRW captured, and they are the same in every byte.

Just one suggestion to all, I am using Silkypix to open the DNG and it is pretty good software. the Silkypix for Pentax version works well, I think Ixus 115 HS use the same sony sensor as Pentax Q.

Silkypix recognizes the DNG purely on the EXIF camera name and build  :P

Re: IXUS 115 HS / ELPH 100 HS request
« Reply #335 on: 05 / July / 2012, 09:55:36 »
the first JPG and DNG may not be always good

but if it is bad, it shows an all white frame, both in JPG and DNG

Re: IXUS 115 HS / ELPH 100 HS request
« Reply #336 on: 05 / July / 2012, 10:07:03 »
and to be more precise

if a frame is bad, its JPG will be compressed from the previous good frame but over-exposed colour. DNG will be storing exactly as the previous good DNG.

if the first frame is bad, no previous good frame can be found, JPG and DNG will be all white

Re: IXUS 115 HS / ELPH 100 HS request
« Reply #337 on: 05 / July / 2012, 10:12:59 »
if I am turning off the DNG, allowing the RAW to be saved as CRW, I got every JPG very good.

but yet I need to see if the CRW is saved correctly, give me some time to check


Re: IXUS 115 HS / ELPH 100 HS request
« Reply #338 on: 05 / July / 2012, 10:30:29 »
downloaded rawtherapee to open the CRW

although the JPG are all good, the CRW are not.

like DNG files, some of the CRW files are exactly the same as the previous good CRW.

if I am saving DNG, I can tell from the JPG that the corresponding DNG is bad or not
if I am saving CRW, all the JPG are good, and I should see the CRW to know it is bad and just a copy of the previous good CRW

Re: IXUS 115 HS / ELPH 100 HS request
« Reply #339 on: 05 / July / 2012, 10:50:36 »
Okay - these are good clues.  Thanks !

Sounds like its not tracking the current RAW buffer properly.  That narrows it down to two small functions in lib.c.

It doesn't explain the corrupted DNG file - unless the byte swap process is messing up the image before it gets converted to JPG.  Hmmm  .. that might be it.

I'll play with this some more today and maybe get it right by tonight - alpha10 !!

Ported :   A1200    SD940   G10    Powershot N    G16

 

Related Topics


SimplePortal 2.3.6 © 2008-2014, SimplePortal