Bug Report: IXUS 800 IS, A610, IXUS 60, IXUS 50, A420

koshy

1096

« on: 23 / July / 2014, 17:17:49 »

Advertisements

Shooting JPEG and DNG into the same folder I noticed two problems:

a) DNG time stamps are 1 hour earlier than JPEG time stamps
EDIT: This was no bug but daylight saving time

b) Long story short: Something goes wrong with the very first DNG file name in a folder.
EDIT: Found the same problem for A610 and IXUS 60
EDIT 2: Found the same problem for A420 and IXUS 50

- I had used the camera for a session before the current one
- the previous session ended with IMG_0021.JPG
- I cut the DCIM folder from the SD card
- on this new session the camera started with IMG_0001.JPG but I got a CRW_0022.DNG
- from the second shot on the IMG_0002.JPG got a matching CRW_0002.DNG
- looking at my previous session I also lack a CRW_0001.DNG
- there is no excess number DNG so it must have been between 2 and 21 and got overwritten

I was using this CHDK: ixus800_sd700-100b-1.2.0-3515

« Last Edit: 03 / August / 2014, 13:04:20 by koshy »

Logged

Koshy had a little ELPH which wasn't white as snow but everywhere that Koshy went the ELPH was sure to go. (actually an SD, but that detail ruins the rhyme...)

srsa_4c

4451

Re: Bug Report: IXUS 800 IS

« Reply #1 on: 23 / July / 2014, 17:22:44 »

Quote from: koshy on 23 / July / 2014, 17:17:49

a) DNG time stamps are 1 hour earlier than JPEG time stamps

That can be perfectly normal if "Daylight saving time" is enabled in the Canon menu. CHDK is not able to take that into account.

Logged

koshy

1096

Re: Bug Report: IXUS 800 IS

« Reply #2 on: 23 / July / 2014, 17:32:18 »

Quote from: srsa_4c on 23 / July / 2014, 17:22:44

That can be perfectly normal if "Daylight saving time" is enabled in the Canon menu.

O.k. I checked and that is what got me. So a) is solved which leaves b)

Quote from: srsa_4c on 23 / July / 2014, 17:22:44

CHDK is not able to take that into account.

Could the JPEG be analyzed to detect this kind of thing?

Logged

Koshy had a little ELPH which wasn't white as snow but everywhere that Koshy went the ELPH was sure to go. (actually an SD, but that detail ruins the rhyme...)

koshy

1096

Re: Bug Report: IXUS 800 IS, A610

« Reply #3 on: 23 / July / 2014, 18:40:37 »

On A610 using CHDK a610-100e-1.2.0-3524 I found the same problem.

I was missing my CRW_0001.DNG for IMG_0001.JPG
I retested and found that on an empty card as described in the initial post I got CRW_0014.DNG and IMG_0001.JPG for the first shot after the previous session ended at CRW_0013.DNG

Logged

Koshy had a little ELPH which wasn't white as snow but everywhere that Koshy went the ELPH was sure to go. (actually an SD, but that detail ruins the rhyme...)

koshy

1096

Re: Bug Report: IXUS 800 IS, A610, IXUS 60

« Reply #4 on: 23 / July / 2014, 18:56:22 »

Same for IXUS 60 IMG_0001.JPG and CRW_0020.DNG after the previous session ended with CRW_0019.DNG

I thought of a new test and here it gets weirder. To check the one shot I had naturally turned the camera off and removed the card. This time I re-inserted it without removing the images. Powered the camera back on and guess what: Vice versa! IMG_0021.JPG and CRW_0002.DNG
O.k. again. Third time gives CRW_0022.DNG IMG_0022.JPG once it's in sync it won't break anything. I took three more shots three times turning the cam off in between and that was continuous through CRW_0031.DNG. Of course as soon as we clear the card the trouble starts anew.

Logged

Koshy had a little ELPH which wasn't white as snow but everywhere that Koshy went the ELPH was sure to go. (actually an SD, but that detail ruins the rhyme...)

srsa_4c

4451

Re: Bug Report: IXUS 800 IS

« Reply #5 on: 24 / July / 2014, 18:10:38 »

Quote from: koshy on 23 / July / 2014, 17:32:18

Quote from: srsa_4c on 23 / July / 2014, 17:22:44
CHDK is not able to take that into account.
Could the JPEG be analyzed to detect this kind of thing?

Well, the firmware could be analyzed, but that costs time and the devs may want to spend their time on stuff that they find more interesting.

About the counter issue: the image counter is updated by a Canon task that runs parallel to the task we're modifying. Note that most models do not support RAW and only need to number JPEG images. If the counter is updated after CHDK grabs the raw data, we're seeing the previous counter value. This is what happens on most early Vx ports: we're guessing what the next image number will be, and that guess is (as you can see) mostly right. This might again be something that could be improved, but that would - again - require more or less research.

I personally find the 2 above issues not major enough - one can be worked around, the other happens rarely. For more info, check reyalp's sig

Logged

koshy

1096

Re: Bug Report: IXUS 800 IS, A610, IXUS 60

« Reply #6 on: 25 / July / 2014, 07:10:57 »

Quote from: srsa_4c on 24 / July / 2014, 18:10:38

For more info, check reyalp's sig

It is a hack alright. I don't have any expectations here. What I try to do for the community is documenting what I encounter opposed to just ignoring it / working around it. Often when I identify bugs in our software I think "c'mon how may folks noticed this before I did and didn't say anything".

Anyway, what the devs do or don't pick up is theirs to decide. Issue b) above destroyed images I would have wanted. Mine could be reproduced. That might not always hold true.

Knowing it exists any user can circumvent the problem by taking two blank frames on each empty card on the affected cameras. Interestingly few camera models have this.

Logged

Koshy had a little ELPH which wasn't white as snow but everywhere that Koshy went the ELPH was sure to go. (actually an SD, but that detail ruins the rhyme...)

srsa_4c

4451

Re: Bug Report: IXUS 800 IS, A610, IXUS 60

« Reply #7 on: 25 / July / 2014, 14:14:21 »

Overwriting files is bad. There is currently no check for pre-existing files in the raw/dng routines (AFAIK). This can affect some other (newer) cameras too, due to the unreliability of file number handling (some ports are mostly protected against this, they delay raw file saving until the firmware's file counter becomes updated).

Logged

koshy

1096

Re: Bug Report: IXUS 800 IS, A610, IXUS 60, IXUS 50, A420

« Reply #8 on: 03 / August / 2014, 13:05:54 »

Today I found that IXUS 50 and A420 suffer from the same naming / DNG overwriting issue

Logged

Koshy had a little ELPH which wasn't white as snow but everywhere that Koshy went the ELPH was sure to go. (actually an SD, but that detail ruins the rhyme...)