Messages

Do what I did and see what event it generates.

I also want to test some but noticed that dm.spy.o was already in the makefile, but cant seem to get the cache hack to work correctly. I included #cache_hacks.h in boot-hack.c (noticed there's a lack of defined HIJACK_CACHE_HACK) also noticed that there's some problems with the debug_intercept ()

Patching the stuff A1ex posted was ok though, but cant compile.. any hints would be great.

This was rather surprising. More tests should be done before we can make any claims (Sigma image stabilizator, 2010-year old model)

http://imageshack.us/a/img585/8712/21963082.jpg

For longer delays, use the self-timer option (2s or 10s) - only one click.

That's true and with a rather annoying blinking light as a bonus.

I know.. I can do like Canon and put some scotch over it.



Did some more tests yesterday and was able to get decent handheld shots at 1/60 iso 6400 800mm f/8 evening/night photos from across the street. I was also able to get pretty decent 1/4 iso 100 50mm f/2 indoor shots without much shake.

Two examples:

Iso 3200, 1/100 f/8.0 800mm mirror lens, hand held.

http://imageshack.us/a/img26/2183/img1468acr.jpg


Iso 12800, 1/160, f/8.0 800mm mirror lens, hand held.

http://imageshack.us/a/img853/6108/img1395acr.jpg

There's Canon's proper MLU for that. This is for hand held shooting.

Yes, but as already noted, ML is superior the Canon mirror lock up since it only requires one click. With a delay of more than 1 seconds, it could actually become a replacer for the Canon function.

So I don't really understand why the function should be limited to hand held shooting only.

Done some quick tests with a 800mm mirror lens and shooting at a target around some 5 metre away and come to the following conclusion.

No matter what shooting style, live view gave a slight edge of sharpness in my test...

1 ) Photo
2 ) Photo using live view
3 ) 2 second delay
4 ) 2 second delay with live view
5 ) Remote control photo taken
6 ) Remote control photo using live view
7 ) 10 second delay
8 ) 10 second delay with live view
9 ) Mirror lock up (Canon) 2 shutter presses
10 ) Mirror lock up (Canon) live view
11 ) MLU - 1 second delay
12 ) MLU with live view - 1 second delay

The 2, 4, 6, 8 are all sharper than their non live view counterparts.

I perceive the differences between 9 - 10, and 11-12 as somewhat smaller than the differences between other counterparts, this despite contact with the camera in order to activate the functions.

I got the impression that the remote shooting with live view gave the sharpest result, followed by a 2 second delay, and the mirror lock up and all the other live view photos. It's harder to tell which of the ordinary photos that was sharper but I perceived 1 and 9, 11 as quite similar and probably amongst the sharpest ones, most likely due there's not much difference from pressing the shutter one time, or two times (as long as there's long enough pause between the presses).

One thing that bothers me is that the maximum pause is only up to 1 second.

Shooting with long lenses (moon) or macro lenses (extremely small objects), sometimes it can be better to have a pause closer to 4 s, or 10 s to let any camera shake die out. Could it be implemented?

I've also noticed that the feature does not work well with changing shooting mode from one shot/continuous shooting to remote control/2 second remote... Not sure why, but it seems it does not work then...

Yes, that's true. Canon requires two presses, which makes ML superior.

Canons model works in an old manner, you press once and decide when to take the picture. I suppose wind , lighting conditions and other stuff might have been the reason for two presses because that way you could measure the scene, and pretty much just wait until the breeze has settled or that "right light" is there/bird in the frame, or whatever...

As long as there's a pause value large enough on ML, it should pretty much everything one can ask for.

Stretching the feature further is probably not possible due the mirror lockup/lack of measure possibilities..

I've just done some quick tests, with a sigma lens that has image stabilisation and without the stabilisation and the MLU feature and somehow got the impression the MLU is better. But needs more testing.

Well, it has mirror lock up. Not very hand-holdable, but...

There are quite a few articles on it. I stumbled on it when I first started shooting with mirror lenses...

5D2, C.Fn III: Autofocus/Drive Mirror lockup 0/1 - Disable or Enable
550D, C.Fn III: Autofocus/Drive Mirror lockup 0/1 - Disable or Enable

Random links:

http://www.cameratechnica.com/2011/04/26/dslr-mirror-lock-up-worth-the-effort-or-not/
http://www.bobatkins.com/photography/tutorials/mlu.html

Thanks. From the few tests I've tried on the new builds, it even works better than the Canon version on the 550D...

I've some similar findings as the original problem posted:

* The Magic Lantern Bracketing takes a bit longer to perform than compared to the Canon firmware. 3 different exposures, 3 different shutter settings. Full iso control has not been able to boost speeds.

* I've also noticed that Bracketing speed differs substantially compared between my 5D2 (slower) to my 550D (faster). No matter what setting, the 550D always performed faster than the 5D2 for some reason unknown.

I've been running the latest builds and also tried the nightly builds but found similar results.

Tried taking some photos on the fly. It's a bit smoother than the built in Canon feature, but I noticed that sometimes the picture is not taken.

When using a Canon 550D, 0.1 seconds and trap focus, I noticed that sometimes the mirror lock-up would occur, but no picture would be taken until i pressed the shutter.

When changing to 0.2 or even 0.5 seconds the mirror would lock up and a picture was taken and then returned to normal mode without further interaction from the shutter button.

I also noticed that the mirror lock-up would not occur, unless I turned autofocus off from the lens and no picture could be taken, until the autofocus got turned to manual focus on the lens.

Thanks for this!

Looks very interesting. Will try it out!

After reading the Dx0Mark article on the 650D, I was kind of unimpressed. Rather stick with some older 550D, or other model, but that's just me...

Nowadays there are old 5d2's for pretty good money and also seen people selling their 7d's for quite a low sum (almost as expensive second hand as the 5d2's though, so... )

http://www.dxomark.com/index.php/Publications/DxOMark-Reviews/CANON-EOS-650D-Strictly-Status-Quo

Voted. Good luck!

I would prefer a few buttons. Touch screens are generally quite bad when it rains, for example...

I've tried some older Carl Zeiss Jena Biometar, Arsat, Tamron lenses. They have all been good fun, especially the Pentacon 6 lenses with a tilt adapter. The adaptall-2 lenses can also be extremely fun.

Of course nothing comes without drawbacks. The older lenses can sometimes have slow/rusty apertures and other problems.

In case one does not want to go down the used/vintage lenses route, Samyang recently released a couple of so called Vdslr cine lenses, which is the same as their manual lenses but designed to fit focus pull systems. 

Other than that Nikon still produces some great manual focus lenses. The adapter is quite cheep and in case escaping the Canon route, the lenses will still work on ... Nikon...

Looks good. The clarity is really remarkable in your example.

I've found the exposure values a bit confusing (I assume 1 exposure step is 1 aperture step)

Very nice!

Ah. That's excellent! Havn't run Bracketing and Focus stacking at the same time.

From what I've learned faulty white balance can be tricky at times to save in post processing. Especially pink/purple flowers tend to overexpose quickly in post production for some weird reason. Might be that Adobe Camera Raw dont handle it too well or due the general curves are not symmetrical. I'm not sure if one would benefit from multiple exposures in such a case.

I fully agree. It would make things difficult and the problem with manual lenses is still there. As of now HDR bracketing still works with manual lenses without having to change the aperture (shutter and iso only). Adding DOF to HDR bracketing complicates things, but it sure looks pretty interesting... Combined with White balance bracketing and focus stacking, one would get a very powerful bracketing tool.

Thanks for that article.

But I'm wondering if there's a possibility to see what happens if the aperture changes the exposure values as a extra feature and/or combined with shutter/iso?

Here's two manual examples with aperture and Shutter combined.

In the first example I went from a bright scene with a wide aperture, closing down the aperture and compensating somewhat with a slower shutter not to underexpose too quickly, gradually changing it into a dark scene. Merged in Photomatix.
http://imageshack.us/a/img9/7109/img9942and9moretonemapp.jpg

The other example i went from a pretty dark scene, with a wide aperture and fast shutter, slowly narrowing down the aperture and compensating with a slower shutter to gradually change it into a bright scene. Merged in Photomatix.
http://imageshack.us/a/img600/9576/img9931and10moretonemap.jpg

And for reference normal DOF bracketing merged in Helicon Focus:
http://imageshack.us/a/img834/1378/focusresult3a88.jpg

Exposure (Tv) bracketing with shutter and iso changes, aperture stuck at f/1.4. Merged in Photomatix:
http://imageshack.us/a/img100/6192/img9922345678tonemapped.jpg

At f/1.4
http://imageshack.us/a/img834/3776/img9904acr.jpg
At f/22
http://imageshack.us/a/img685/7884/img9912acr.jpg

[Edit]
Another example, this one running from bright with a narrow aperture, slowly opening up the aperture and also also adjusting shutter speeds not to overexpose but rather to darken the image:
http://imageshack.us/a/img713/2672/img9969and8moretonemapp.jpg


Also notice how fake photomatix images tend to look.

I experimented some this morning just to get a hang of the bracketing. There are of course endless possibilities but I'm wondering if the aperture should affect the exposure in similar way as the shutter does?

I used Helicon Focus to stack a bunch of images ranging from f/1.4 to f/22. The result is painterly, but dull.

50mm f/1.4 http://imageshack.us/a/img138/4493/img9815acr.jpg
50mm f/22.0 http://imageshack.us/a/img89/9545/img9822acr.jpg

Helicon Focus (depth map)http://imageshack.us/a/img811/4339/focusresult2b88.jpg
Helicon Focus (weighted average) http://imageshack.us/a/img546/9328/focusresulta88.jpg

That was really fast, Alex!

Already tried tweaking some with the new exposure tricks and it seems promising. Aperture and exposure bracketing in combination with focus it would be a killer...

Trase: That's how I tried some stacking as well, but also tried it with Helicon Focus. Then there's Zerene stacker, which might come in handy.

Sounds like an interesting feature.

I did a couple of image stacking tests using mirror lenses in order to get sharp tele pictures and have also tried a couple of shots with changing aperture on some manual lenses with some pretty interesting result.

What image software do you stack the images with?

I think the best way to deal with this option is to choose f/5.6 for most of the time, and keep your camera in AV or Manual mode setting (in manual mode, you could select auto iso and still tell the camera to keep a example 1/100 shutter and f/5.6 aperture. simple!)

I personally shoot a lot using TV mode (for mirror lenses and other longer lenses and when I want to make sure that my shutter is fast enough) or I choose AV mode (for landscape, every-day-situation snapshots, etc.) and at times M mode for those times when I have time to play around and want to have a bit more control.

All those stuff stays the same whether I am using Magic Lantern or not.