Show posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Show posts Menu

Messages - Sapporo

Camera-specific Development / Re: Canon EOS R5 / R6
April 02, 2023, 09:48:29 PM
A1ex mentioned about some columns not saved in the raw file for 5D III
Camera-specific Development / Re: Canon EOS R5 / R6
April 02, 2023, 04:54:29 PM
R6 has 20,1 effective megapixels and 21,4 total megapixels. Including the optical black field to the left and top I get 5568x3708 pixels. Total 20646144 pixels. Where are the others? To the right and trashed before saving to card like 5D III?
Reverse Engineering / Hot pixel suppression at ISO 6400?
November 24, 2022, 11:17:39 AM
My two 6D and M5 have less hot pixels at ISO 6400 and above. At 25 seconds exposure hundreds of hot pixels at ISO 3200 and below were reduced 10 times.
Raw samples
Notice that the hot pixels left at ISO 6400 are the ones in pair at ISO 3200.

7D doesn't have the feature. I wonder if newer cameras like R5/RP also have the feature?
Quote from: maxtorine on November 23, 2022, 09:00:46 PM
Thank you! Never heard of that one but will give it a try if I can't find a free solution  :)
Ask the creater of Tornado EOS if the paid version will be able to solve your service menu issue.
#5, but that will cost you more than the camera is worth
Scripting Q&A / Re: Take a photo at each ISO
May 31, 2021, 02:46:46 PM
Works great for 7D even up to H. With 6D it only goes up to ISO 6400 and not 12800, 25600, H1 and H2.

I have tried

function test()
    camera.iso.value = 100
    camera.iso.value = 200
    camera.iso.value = 400
    camera.iso.value = 800
    camera.iso.value = 1600
    camera.iso.value = 3200
    camera.iso.value = 6400
    camera.iso.value = 12800
    camera.iso.value = 25600
    camera.iso.value = 51200
    camera.iso.value = 124000

keymenu =
    name = "ISO photo bracket",
    select = function(this) task.create(test) end,

My goal is to take raw samples for NR in a neural network
Reverse Engineering / Re: Lens serial number in Exif
December 16, 2020, 02:18:36 PM
Quote from: Walter Schulz on December 16, 2020, 01:20:01 PM
Nothing we can do from here.
Only thing I can imagine is trying to contact stolencamerafinder admin and asking him to address this issue in some way. Maybe asking Canon users to look for len's funny number via EXIF and not trying to use a number unable to track.
As you can see, people have figured out before how it works with other serial numbers from Canon.
Reverse Engineering / Re: Lens serial number in Exif
December 16, 2020, 11:48:22 AM
Quote from: Walter Schulz on December 16, 2020, 09:46:24 AM
Sorry! No correlation found.

Do you have any problem affected by this? Repo, legal, incurance claim, ... ?
The problem may be when a lens gets stolen and the only lens serial number is the one in Exif.
Another thing is the web page stolencamerafinder. If you don't know how it works you will enter the lens serial number that is not the same as the one the web page will search for.
Reverse Engineering / Lens serial number in Exif
December 16, 2020, 08:22:23 AM
Canon added lens serial number in Exif 2004 (28-300 f/3.5-5.6). The tag is a mix of decimals and hexadecimals.

Lenses from 2014 from (16-35/4L IS) have the same lens serial number in Exif as written on the lens.

Samples what Canon has used before

Any idea how to convert lens serial number in Exif to the real lens serial number? I have tried some combinations this morning without result.
Camera-specific Development / Re: Canon 5D Mark IV
October 21, 2020, 09:36:16 AM
Firmware 1.3.0 out now.
Camera-specific Development / Re: Canon 5D Mark IV
September 30, 2020, 01:36:50 PM
C log in 5D IV is just an on/off feature in the firmware. Same as language restriction or WiFi/GPS.
Possible to change with Tornado EOS but I also suppose A1ex could do the same.
Feature Requests / Re: ISO below 100
August 17, 2020, 10:43:30 PM
Search for ISO 77 and 5D III.

Btw, are you using the raw histogram? Then you get feedback how much you have left before you clip the highlight.
Quote from: russellsnr on March 20, 2020, 02:22:37 PM
Hi, In the middle or was in the middle of moving back to the UK but with the problems we are facing that has been put on hold, I normally use the Lightroom plugin to convert my dual iso files but that computer has been packed away but I have a small lap top running Windows 10 so where please can I download the stand alone version of the converting software for a Windows 10 computer. Thankyou.
I have been using Barracuda GUI for Windows
Quote from: kayman1021 on September 29, 2019, 10:29:16 PM
I had some success with the .cr2 file. After dropping it into Adobe DNG converter, I've recieved a 16bit uncompressed .dng file.
After many tries, I've managed to open the data. Stangely the inner data width is 5360 opposed with 5202 that is reported by RawTherapee. The additional data is not black, but very dark rectangle on the left. RawTherapee somehow just crops it. Height is 3465.

I've overwritten the values with vertical neighbour averages, and stuffed the data back into the dng. Black lines are gone, but there are visible artefacts, like above the eye.
Almost gone when using VNG4 in RawTherapee.
Quote from: dim on September 21, 2019, 01:08:41 PM
I tried Switch for the 1st time without problem on a ubuntu studio 18.10. It installed Yad , as planned. With a serie of pictures, dualiso, from the same Eos5DmkII, it doesn't work at all. I look at the log file & there, i don't understand the problem... The Switch files are in my Home directory. I followed the instructions. Can you help me?
The log file:
Got the same yesterday. Worked last time I tried. Ubuntu 18.04.
Camera-specific Development / Re: Canon 90D
September 16, 2019, 11:05:40 AM
Quote from: Greg on September 02, 2019, 11:06:40 PM
DR analysis based on a lossy compressed CR3 file is not a good idea -
I got the same results when I with Hraw checked normal raw files from 90D.
Quote from: a1ex on September 01, 2019, 07:17:35 AM
Hardware issue - one of the 8 readout channels is no longer working.

What we can do, at software level, is to interpolate these dead columns as bad pixels (easiest to do in post-processing, but also doable in the camera, to some extent). Unsure whether Canon's own bad pixel correction feature would work here (didn't try).

In any case, this workaround will introduce some aliasing (somewhat similar to Dual ISO), but at least you will be able to keep using the camera for a while. Whether that's going to be a matter of days, months, or years until it fails completely, I have no idea.
Do you have more details how Canon's own bad pixel correction works? The only thing I know is the "lens cover on, manual sensor cleaning etc thingy", but nothing that explains how it works and why it doesn't always work.
Camera-specific Development / Re: Canon EOS R / RP
August 25, 2019, 08:23:37 AM
Quote from: kitor on August 24, 2019, 11:41:15 PM
This info was published before those firmware were even mentioned elsewhere, so I assumed it was an error.
But yes, that makes sense.
CVE-2019-6001 published 6 August, firmware 1.3.0 published 7 August. 1.3.0 was already ready.
Camera-specific Development / Re: Canon EOS R / RP
August 24, 2019, 10:07:12 AM
Quote from: kitor on August 23, 2019, 12:12:22 PM
R firmware 1.3:
RP firmware 1.2:
No mentions about fixes for recent vulnerabilities.
Not fixed.
"EOS R firmware version 1.3.0 and earlier, EOS RP firmware version 1.2.0 and earlier, EOS RP GOLD firmware version 1.2.0 and earlier"
Quote from: chris_overseas on August 12, 2019, 09:35:01 AM
I doubt this is a concern, the vulnerabilities found aren't related to the firmware encryption. I also wonder if it's even possible for Canon to update the AES key with a firmware update.

"Even though our camera model doesn't support Bluetooth, some Bluetooth-related commands were apparently left behind, and are still accessible to attackers. In this case, we found a classic Stack-Based Buffer Overflow" - maybe those cameras don't have the Bluetooth code in their firmware, or different versions of it that aren't vulnerable?
They are vulnerable according to CVE. 7D together with WFT-E5 isn't on the other hand mentioned.

Possible to encrypt all files. Perhaps a new feature for ML instead of io_crypt?