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 - Marsu42

Pages: 1 2 3 [4] 5 6 ... 58
76
Wanted: your results

You keep updating the module in the repo - should we wait some more, or is this the one you want the tests to be run with? If I take the time to cover the 6d, I really don't want to do it twice :-p

77
General Chat / Re: Help choosing camera (6D vs 5D2 vs 5D3)
« on: June 03, 2014, 10:16:19 PM »
what i like to shoot is fast motion. such as: motocross

Forget 5d2/6d, terrible tracking af, get a 5d3 or - if money really isn't an issue - 1d4/1dx.

78
I didn't do systematic tests, just noticed borked greymax awb with undexposed shots. If you didn't look into that much yourself I'll try to find the time and do a proper test series with greymax and greymed.

I ran a test series with the dual_iso samples I uploaded earlier (https://bitbucket.org/Marsu42/ml-pull/downloads/dual_iso_samples_6d_60d_FIXED.zip) and graymax works ok with underexposure on these. Obviously it's really the algorithm failing on non-studio shots when a plain color area cannot be found. Knowing this, I'll use greymed in the future on my shots...

... but as an idea, maybe an smart auto-detection greymax/greymed would be possible if the tint is off the scale with greymax.

79
So, if you set the camera on a tripod and shoot the same test scene at different exposure settings, you get different auto WB results?

I didn't do systematic tests, just noticed borked greymax awb with undexposed shots. If you didn't look into that much yourself I'll try to find the time and do a proper test series with greymax and greymed.

80
By whom and by what method is it known?

Sorry for not providing an exact quote, I remembered you saying as much after introducing the dual_iso awb so I thought it's not necessary to provide exact proof. For my shots, greymax starts failing with ~2ev underexposure and results in tint off the chart, usually magenta.

The graymed algorithm is a bit closer to the traditional gray-world hypothesis; maybe it's worth trying that too.

Ok, I'll try that for a change. Probably the reason for the failure really is that it fails to find color areas which might not be surprising in nature shots.

81
Concerning greymax: It's known the algorithm completely misses if the shot is underexposed. Now, some very smart people would say "just don't underexpose!" but for simple /me and probably others who expose for highlight safety: Probably there is a way to adjust the awb detection for these cases? If required, I can provide some dual_iso samples.

82
Archived porting threads / Re: [UNMAINTAINED] Canon 100D / SL1
« on: May 29, 2014, 10:41:08 PM »
Great work, sorry for not replying I've been busy with school and I just graduated and got a job so I am busy with that all day now.

I fully understand and sympathize as I'm in about the same situation, working on ML needs some coherent time in peace and quiet or nothing will come of it. But it's a good reminder to stick with the working ports even though some maintainer has been found, life changes too fast to make promises for unpaid work for the next year or so :-)

I promise to work on this in the near future

83
Personally, I would segregate with folder management.  Originals in one folder, processed in another.

I'm doing that right now with cr2 dual_iso originals on a backup hd and converted dng on a live hd right now. The nice thing about this solution that I just need to back up the non-dual_iso files from the live hd and am not duplicating data.

But folder management doesn't work if cr2hdr insists on overwriting the source .dng and you cannot rename the target or set an output folder :-\ ... cr2hdr simply needs an output path option. Or am I missing something here?

84
There is an option to not overwrite existing data:  https://bitbucket.org/hudson/magic-lantern/commits/7fad96785821644a05dfb9d4212791e6e1bc7706

But even if you have to opt-in so you have to be already aware of the problem, and then it's no good because it would refuse converting the dual_iso original dng... the .cr2 rename method is imho rather hackish. I'd rather rename them to .mld or something :-p

I'm getting lower case output here from Adobe, on a windows (8.1) machine.

In theory, you can configure any windows ntfs volume to be case-sensitive, I did this because it helps mingw before I finally moved to cygwin. But very few users do that, usually windows is case-insensitive and .dng overwrites .DNG and vice versa.

85
On Windows, be careful because they have the same extension, so the unprocessed one will probably get overwritten. [...]
To solve this on Windows, one way would be to rename the unprocessed .dng back to .cr2, since dcraw doesn't care about the extension. Ugly, but gets the job done.

Yes, I already noticed that, quite a catch since it destroys the original data w/o even a warning. Couldn't you just add an auto-rename or output path option to cr2hdr to solve this once and for all for everyone (yes-i-could-submit-a-patch)?

86
Probably because so many things rely on the embedded thumbnail.

They do? Other than in-camera preview I never found anything that uses it, it's just a waist of storage space. The one reason to keep it around is because you can extract it from the cr2 if the raw data is broken for some reason, but that's hardly any good with dual_iso shots.

Thanks for the answer concerning cr2/dng though, in this case I'll convert my cr2 to dng as any other cr2-processing programs like dxo cannot handle dual_iso files anyway.

Edit: fyi: the dng files are nearly 20% smaller than cr2.

87
Prey, is there any drawback if converting the dual_iso unprocessed cr2 to dng (non-lossy of course) and *then* put them through cr2hdr?

Storing images as dng instead of cr2 always saves a few mb (mostly because the jpeg preview is dropped), and this can add up to a lot of saved space looking at my growing cr2 dual_iso archive. Alternatively, I'm still looking for a tool that erases the thumbnail from a cr2 and leaves the rest untouched :-p

88
General Help Q&A / Re: Micro SD possible for ML boot ??? ( 5D3 )
« on: May 29, 2014, 05:06:32 PM »
do you think a class 4 card is fast enaugh to boot ml from? im sorry if it sound stupid or anything, i just want to be sure ;)

This:

ML doesn't care what card it is.  If Canon accepts it, it's fine for ML.

89
I'm thinking at some backend where you would write the dependency checks only once, so you would have a variable to display in the menu (that you can toggle), and another variable to use in the worker code.

Well, at the end of the day imho the current way is working rather fine, it's not like ML is the firmware for the next mission to Mars. There are not that many inter-dependencies (that would come to my mind), and you just have to be sure to check for enabled/disabled features in the code. Much more important to get shiny new features that actually improve the photography result. Or save the camera's shutter. Like full res silent dng :->

90
Note that graying out is done only on the user interface; it has no effect of the behavior of that code. So, right now, the checks have to be done both in menu (to gray out the option) and in the worker code (usually to make sure that option has no effect, or to print an error message).

+1 for checking the latter, it often happens to me that a feature or part of it still does something even if it has been deactivated in the menu, so don't forget to double-check if individual menu options are on/off and then if the whole main menu item has been disabled :-p

91
Archived porting threads / Re: [UNMAINTAINED] Canon 100D / SL1
« on: May 28, 2014, 04:51:37 PM »
Probably. Hello world was only achieved yesterday, I don't there's been a whole lot else done at the point, although I can't speak for count.

I really don't want to spoil everyone's day (again :-p)), but for people deciding what camera with ML support to get I cannot help to mention that "Hello World" is essentially next to nothing and only shows Canon didn't disable the boot code loader. "Hello World" does *not* mean finding some Canon fw stubs and copy/pasting the ML code from another camera type will immediately lead to a working ML, usually there is a lot of time-consuming and tiresome tuning to do.

92
Hello, so what do you suggest me to do!?... is this a mistake of Photomatix?

I wouldn't mention the word "mistake" :-) but since you're in customer e-mail support contact with them anyway, there's no harm telling them that they're not observing the dng specification as there are default values for the tags in question when they don't exist in the file. Falling back to the default value would be the proper thing to do and easy to accomplish as well.

93
When reading an optional tag, the proper thing to do is to use its default value, as described in the DNG spec. Do you disagree with this?!

Right, if it does have a default value (which I didn't know, if I'd read long lists I'd start with the exiftool help page :-p) then clearly the photomatrix guys are to blame - I'd suggest the user in question replies to their "support" e-mail as such, suggesting to add the default fallback to their program. In reality of course, adding these to cr2hdr might be the realistic solution :-\

94
These tags are optional according to the DNG spec, so... yeah, better let the free software guys work around problems in nonfree programs.

Well, you cannot really call it a problem since they're free to depend on official, but optional tags if they require them for one reason or another and all camera out there excluding ml write them - which seems to be the case or they'd have added a workaround by now.

Still, it unfortunately shows that ml still seems to be considered a "hack", not widespread and not to be taken seriously enough to add consideration for ml features (dual_iso, mlv) into commercial, closed-source software :-\

Anyway, adding these tags is quite easy, you only have to edit chdk-dng, fill the missing tags

I would have thought adding static exif tags would best be done in cr2hdr esp. since as I just learned the silent dng don't contain much exif data anyway?

95
Duplicate Questions / Re: Map the Main Dial to change ISO?
« on: May 28, 2014, 10:36:48 AM »
Try looking at the SET buttons options.

At least on 6d, it's possible with stock Canon fw to set iso with the back dial when SET is pressed which makes it more or less unusable for me since I'm no circus artist and am not able to do this with right-hand only operation while keeping the camera perfectly still. Lifting this requirement (and adding the feature for older cameras) so any chosen dial would modify iso w/o pressing set would be welcome, though not really crucial. It's a part of the "map buttons/dials to various Canon/ML functions" really.

96
Feature Requests / Re: GPS in silent dng
« on: May 28, 2014, 10:33:51 AM »
I think silent pics needs to refactored first, as no exif data is added, let alone GPS information.

None at all? My, that shows how seldom I use silent pix since there's no matrix/full res mode in the dng version :-\ ... I've updated the feature request accordingly.

97
Feature Requests / Exif data (inc. gps) in silent dng
« on: May 28, 2014, 08:47:07 AM »
It would be nice if the (well, 6d for now) would add [edit: the usual suspect lens/camera exif data inc.] the current gps information to silent dng pictures so you can skip adding the geo tags from the tracklog later on. Currently, chdk-dng says: "Code stripped down a bit, since we don't care about GPS and advanced exif stuff (at least for now)"

Not a very pressing issue, but I'd like to add a feature request here as a reminder if someone works on gps or dng anyway.

98
If one has trouble identifying the language from a public source code repository, or if he asks things like where you can find boot-hack.c, the skill set of that candidate becomes quite obvious

Well, in all fairness he was just asking for some basic information to get his son involved who is supposed to know (even more :-)) stuff...

My son is a developer and I have a 6d? I would be happy to talk to him and see if he would be willing to do a little side work on ML if it would be helpful. Are there devs who would be willing to talk or assist him with getting up to speed?

... it doesn't work like that, getting started with embedded ML is not like setting up MS Visual Studio and code some gui apps, it's much more time consuming and trial and error. But on the other hand, coding for ML can be really fun if you have enough time on your hands because you can make your dslr do what *you* want.

It'd be great if your son lends a hand, but he would have to have some interest of his own in the project to have enough motivation to get started - it's not paid work, after all.

99
Now I face this huge problem it's called "fixed pattern noise" it literally ruins the image and I want to fix it.

You've just discovered the main problem with the 7d :-\

Fyi: the fpn problem has been somewhat lessened by Canon on newer crop models starting with the 60d. The problem is that this noise type invalidates a good part of the potential dynamic range so you always have to expose correctly or ettr. In postprocessing, there's also much, much less potential for raising shadows than on the best Canon sensors (1dx & 6d).

Unless some dev discovers the holy grail, you either have to live with it, get a more current model (70d) or for video the 5d3 I'm afraid to say, though the latter also has some fpn.

100
General Help Q&A / Re: 2nd curtain sync - Canon
« on: May 27, 2014, 09:53:23 AM »
Not sure if the rest of the ETTL remote triggers bunch (Phottix, Pixel King, Yongnuo, RadioPopper, Hahnel, etc. pp.) are up to the task.

Yn can do it with their st-e3 clone trigger, but only in good ol' a:b:c ratio mode and only if you set all flashes to m = no e-ttl. Probably not much of a loss, there was an interesting discussion on CR which demonstrated that 2nd curtain does make little difference most of the time like dancing people except for *very* long exposures with movement in *one* direction.

Pages: 1 2 3 [4] 5 6 ... 58