Messages

Sorry to continue with something that seems already discussed and finished. Just want to give my opinion.

I have now a 5d mk III and I am pretty satisfied with the results it gives me with in-camera HDR and I really wish I had it also on my 60D. If Magic Lantern could do that I would surely appreciate very much, even doing it with JPG. In fact I am doing it with JPG in the 5d mk III as usually I don't need the RAW quality. Usually I use images only on the web, so JPG is ok.

Bests.

Hi Dolec,

as far as I know you can make any video "look" like slow motion using any video editor capable of changing the speed of the video, like kdenlive, adobe premiere and probably many other. The thing is that usually when you do this you make the film with a bigger fps rate, so that when you reduce the speed you can still have something resonable, like 24 to 30fps showing the slow motion. The top fps that your equipment is capable to make depends on its hardware, so ML can't increase the fps. You can still make the normal movie and make a slow motion in an external editor, but as long as you will have less frames it can look really bad, depending on how much you change the speed.

I guess that usually people use the fps override of ML to reduce the fps rate, so it is the opposite of what you want, allowing, for example, longer frame expositions.

Cheers, Cartola.

Hi,

I have made this night picture using magic lantern for the exposure bracketing:
http://cartola.org/360/en/2012/08/19/day-and-night-on-the-roof-at-leblon-rio-de-janeiro-brazil/

I have set up an initial exposure of 2s and 2EV between exposures, have set up exposure order to 0 + ++, so the 2s exposure would be the fastest one, then I've set up 4 shots. The resulting times have been 2s, 10s, 42s and 170s. Following this steps you can extend times much more, as you can start with 30s, make up to 5EV between shots and up to 9 shots with ML auto bracketing. You can also notice that I've gotten those times using f16.

Regards, Cartola.

More 2 cents: the sony sweep feature is good, but not as perfect as that video shows. It many times let defects in the final image, like lines that don't match very well, ghosts and others. The thing is that to make a perfect panorama you should rotate around the nodal point / no parallax point of the camera and they not only don't mention or teach you that but it is also almost impossible to be made without a tripod and an appropriated panoramic tripod head. Having movement in the scene also causes many problems and sony also don't solve them, in fact that usually need to be solved using masks that would really be difficult to set automatically.

It will work many times, mainly when you shoot nature landscapes, and probably this is the way it is most used.

Another option to stitch the pictures in post processing is Hugin, and one great advantage of it is that it is also free, as ML.

There are softwares that can also make a panorama from videos (ex 1, ex 2). As sony algorythm probably does, it is only a question of choosing some frames and stitch them, as all panorama stitching software does. It would be easy to make a unix script to do this with free software like ffmpeg and Hugin.

What I can agree with the first post here is that it could be a good feature for the amateur or those who don't need super stitch quality and don't want to learn a new software and spend time stitching panoramas. Probably that is why every Sony owner I new like this feature very much.

Cheers, Cartola.

Just some newbie thoughts: many times my only post processing to combine different exposures is only run enfuse. Sometimes align_image_stacks (from hugin) and then enfuse. I usually don't like the artistic / artificial effects done by the tonemapping usually used in Photomatix, Luminance HDR and many other tools.

Ok, enfuse is even more trivial, but when you do this several times it takes time. I surely would use an enfuse like fusion in-camera.

Cheers, Cartola.