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.

Messages - histor

Pages: [1] 2 3
The first thing, which comes to my mind, is checking the card serial number (in-camera). If it's out of the Red-range, then it reduces writing speed. There are people, wanting to change these serial numbers but they lack low-level tools for CF. The other way is to find and patch these ranges in camera firmware. This may be a long way or a very long way  ;)

And the last thing to try - switching cards in some way. What if the range check is done just once? When the card is initiated? Can we block the card-door sensor? The latest way of downgrading canon firmware appeared from such card fiddling.

Main Builds / Re: Nightly Builds - try the very latest stuff here
« on: July 31, 2020, 09:52:03 PM »
For 5DII, please, check this thread. And Downloads> Experiments for other latest staff.
Failed auto test is not a problem. There are more dangerous things here  :P

Raw Video / Re: 5D3 full resolution live view 16/9 5K and pre rec
« on: April 16, 2020, 07:32:25 PM »
May I ask you a question: isn't lightning "slow enough" to be caught by motion detection and full-resolution silent pictures? It might be tested with some video at home...

1) No additional sensitivity
2) I guess buffering happpens in RAM, then data slowly pass to the card. Getting free RAM is a good idea - unload all unused modules. Long time ago we even compiled without some features (defishing, etc.) to make ML smaller.
3) Is the bitdepth really critical here? You're shooting just a bright line. Everything else should be blended from a good long exposure frame. Isn't it?

Please, report on you experience. Check CHDK forum on that subject too.

Camera-specific Development / Re: Canon 5D Mark II / 5D2
« on: April 07, 2020, 10:55:11 PM »
a1ex had did some test a long time ago on 5d2 and had the CF card overclock for faster write speed.
It seems, there was a little problem with overclocking. Please, be cautious.

General Help Q&A / Re: Does ML have any log profiles? (Canon 550d)
« on: April 05, 2020, 07:13:52 PM »

Camera-specific Development / Re: Canon 550D / T2i
« on: April 02, 2020, 09:25:42 AM »
All the nightlies found are for middle of 2018.

Look into Experiments section. By the way, fully working adtg_gui module is the perfect start to implement crop_rec video and/or custom adtg gains. Look how Reddeercity did it for 5D2 - lot's of these things are avialable for 550d (try it!).

Or just keep the camera and keep waiting. I use my 550D with great pleasure.

« on: February 26, 2020, 12:47:56 PM »
You have a plenty of options (I've tried all).

Split-screen is useful for dark conditions and fast shooting, when no other option is applicable. You'll waste a lot of shots, it takes time to get familiar to it. It's fun. It's very easy to install/uninstall it yourself.

Zoomed live view and tethered shooting are best suitable for macro and studio stills.
Focus peaking in live view is very effective, but it doesn't work with 100% zoom.

The fastest option is a chipped adapter. It enables built-in phase detection focusing (red dots it the optical viewfinder). I've tried a lot of adapters. The cheapest ones worked best for me. I tuned 5DII inner focus adjustment  for each lens (and wrote down the settings). Works perfect (except macro range). Programmable chips (with focus correction) were so wrong at default  "0", almost useless. 600D lacks inner focus adjustment.

In any case, you need widely opened aperture.

Other experimental builds / Re: Cleaner ISO presets
« on: January 21, 2020, 01:56:03 PM »
Not to interfere with the discussion, just to be a mediator here.

This standard is barely applicable to RAW files in terms of sensitivity. Camera producers use JPEGs for that case.
"ISO speed and ISO speed latitude values shall not be reported for raw images, however, because with raw images processing that affects the values has not been performed" (ISO 12232:2019).

The author of the software - Alex Tutubalin is one of the leading developers of LibRaw (not just the commercial part of the project). Fastrawviewer blog is very informative.
Looks like these two Alexes are the best experts on the question in the open-source world.
But both usually say, they know nothing. As Socrates did.

As Luther noted, there was a canon trick for intermediate ISOs (like 160) – exposure time was changed behind the scenes. Silently. But really it's not the case now. You can always use some moving (rotating) object to prove the same exposure for two frames.

General Help Q&A / Re: Permanent mirror lockup with EOS 500D/T1i?
« on: January 08, 2020, 12:28:01 PM »
You can still fix the mirror up without any software (just discussed here). I used Π-shaped piece of black velvet textured cardboard for that. So you can use USB for remote control and download CR2 files while shooting.

Share Your Videos / Re: an early xmas present for 5d3 owners from TNT!
« on: December 23, 2019, 08:28:44 PM »
but pleaseeeee bring this to 5d2!

Already done: works on all Reddeercity’s builds for 5d2 + (portable module from Alex)
Some fast settings to start with - from my previous posts. Not consider as spamming  ;)

And the whole CMOS/ADTG/Digic register investigation on ISO thread to study.

Academic Corner / Re: Depth map generation using focus stacking
« on: December 17, 2019, 08:52:59 PM »
Welcome and... you have some real giants to compete with. Zerene Stacker is definitely the first one to mention.

Reverse Engineering / Re: CMOS/ADTG/Digic register investigation on ISO
« on: December 06, 2019, 08:27:57 PM »
I promised here this summer to provide better statistics on SNR while using reduced gains on 5D2.
A set of cloudy sky flats repeated twice. (Notice also, how much SNR is lost after 6 consequent shots, as the sensor becomes 7 degrees hotter). Numbers calculated with a fast PixInight script.

Code: [Select]
SNR = 26.42 db, K IMG_7671iso100_DFE111
SNR = 21.04 db, K IMG_7672iso200_DFE111
SNR = 16.76 db, K IMG_7673iso400_DFE111
SNR = 15.92 db, K IMG_7674iso100
SNR = 13.52 db, K IMG_7675iso200
SNR = 12.07 db, K IMG_7676iso400

SNR = 23.85 db, K IMG_7677iso100_DFE111
SNR = 19.55 db, K IMG_7678iso200_DFE111
SNR = 15.68 db, K IMG_7679iso400_DFE111
SNR = 15.27 db, K IMG_7680iso100
SNR = 12.92 db, K IMG_7681iso200
SNR = 11.63 db, K IMG_7682iso400

RawDigger produces a bit different results (and I cannot explain why), but the leader stays the same.
You can download the samples and check my math.

So it works, as expected. The only problem now: it's damn cold to tweak registers outdoors  ;)

There is some open source project with Optical Flow - Interpolating missing frames may be implementable here too...

No, that looks like actual data from the sensor. I get such bias frames adjusting analog iso register (between the main steps) (and I successfully use them as bais frames). Maybe you managed to set different iso for this part )

Academic Corner / Re: Underwater photography
« on: November 18, 2019, 09:08:10 PM »
Then more it is advertised, then further from science it becomes.
“Once trained, the color chart is no longer necessary,” – they say. The water in my bath has different color every morning. ;) So does the ocean, I believe.
It’s a good research but it's going to become a button push process for gopro or something alike. That’s a pity.

And on the other side, there are also brilliant “unscientific” old ways to solve the problem like trashing the Blue channel (and reconstructing it from the rest) and guessing the depthmap (Kolor and Adobe used that for Dehaze feature).

Raw Video Postprocessing / Re: raw2mlv processing
« on: November 15, 2019, 11:31:44 AM »
- post process the mlv file, ie merge with mean or median statistics

Already implemented in mlv_dump. Simple average only. Not yet included in the GUI.
Code: [Select]
-- Image manipulation --
 -a                  average all frames in <inputfile> and output a single-frame  MLV from it
 --avg-vertical      [DARKFRAME ONLY] average the resulting frame in vertical direction, so we will extract vertical banding
 --avg-horizontal    [DARKFRAME ONLY] average the resulting frame in horizontal direction, so we will extract horizontal banding
 -s mlv_file         subtract the reference frame in given file from every single frame during processing
 -t mlv_file         use the reference frame in given file as flat field (gain correction)

Post-processing Workflow / Re: Agressive vertical stripes?
« on: November 15, 2019, 10:58:58 AM »
Banding is always there. At iso 6400 its simply hidden under aggressive amplifier noise. There was a great experiment to correct it in camera.

From my photo experience banding is partly corrected with bias and dark frames and finally goes away after applying flats.
There are special tools for that like Superbias and Canon Banding Correction scripts in PixInsight.
Video case is a bit more complex as we don’t have optical black areas in the frame.

According to the Calculator, it should. Test the card, read its specification and find the way to recycle it safely :)

General Help Q&A / Re: Dual ISO processing strangeness
« on: September 26, 2019, 09:32:47 PM »
May the --force be with you  ;)

General Help Q&A / Re: Dual ISO processing strangeness
« on: September 26, 2019, 08:34:01 PM »
I referred to windows version of cr2hdr-20bit

Sometimes it ignores the files
Code: [Select]
Doesn't look like interlaced ISO
ISO blending didn't work

Running it with the --force option you get blended DNG plus information from the log like this
Code: [Select]
offset too largeYou may get stripes at some lines finally (in the highlights, I believe).

General Help Q&A / Re: Dual ISO processing strangeness
« on: September 26, 2019, 08:03:26 PM »
Have you tried some --force?
Code: [Select]
cr2hdr.exe --force IMG.CR2

You haven't removed the previous build before copying the latest one. There is no in that branch. Remove ML folder and copy one from the downloaded archive. Does it work?

So much discussions about the black window...
Here are two more programs to hide cmd window -

And a C++ sample  here.

Script languages like Autoit/Autohotkey also run hidden processes without a problem.

Fantastic resolution! 3968x1800. Unfortunately the top right corner stays dead.
Frame difference, exaggerated

So, what was the trick? Was it fair? :)

Pages: [1] 2 3