Messages

Really been enjoying the new features so far.

A question on rack focusing. I'm using a 17-55 2.8 IS, and I'm finding the results of the rack focus to be less than optimal.

So I focus and set this to 0 (to be the end/destination of the rack focus operation).

I then use the scroll wheel to set to, say -20 steps from 0, and perform the rack focus. The rack focus is performed, but never ends up on 0 (ie where the initial focus point is) - it's always short (so in the above example, it would probably end up in visual terms of more like -7 steps from 0, rather than 0. So the end point is always out of focus, rather than in focus.

I'll do some playing, and I understand this is very much lens dependent, but I was wondering what the results are from other people (both with other lenses and also the 17-55) to compare? Is this feature working for you? Does it work for you better with some lenses than others? Certain step sizes? Or is the feature a little bit flaky unless you have amazing lenses..?

I'm just trying to work out how to best get this feature to work for me, or if there is an issue with it, etc?

Edit: More detail.
I've got it working better by increasing the step delay from 10ms to 20ms, step size 2 (or 1 for finer resolution), step wait on. It seems to be hitting the end point more accurately now.

One slightly frustrating thing - when altering the steps to here value with the top scroll wheel, moving it fast (which you kinda have to do with a step size of 1 for it to go anywhere, as you want values of +100 or more) often comes off the parameter (especially if the soft-limit is reached, but it will often come off the parameter without that happening) - and then you are now scrolling fast through the menus. Not only that, but there is no way (I can see) of continuing to adjust the steps to here value without first zeroing it out and starting again. This makes setting up a rack focus operation quite painful, currently.

So - could you check what's happening when moving the top wheel quickly to adjust the steps to here parameter, and stop it exiting adjusting the parameter and flying off into menu scrolling - basically, if you hit the soft-limit, continuing to move the top dial should have no effect, with the value sticking at the soft limit, and the camera shouldn't just exit adjusting the parameter without warning.

And if there could be a way to go back into the steps to here parameter and continue adjusting it without zero-ing it out again first, that would be an extra bonus (though less of an issue if the previous issue can be made a bit more stable.)

When I installed it, it worked first. Then when I restarted the camera, i did not see the ML menu anymore. It's like it was not installed.

Yes, alpha versions have to be manually reloaded every time for now.

And it works great! (in alpha2)

g3gg0 and team - thanks for the continued efforts, and of course alpha2, and have a great holiday season. 

Time to dig out the manual for the new features...

Excited for: HDR exposure bracketing, timelapse, rack focus and lots more... awesomeness

btw : for the 3rd alpha , can you guys try to control the audio gain SEPERATLY ? i mean volume on the left channel and volume to the right  .... 

Out of interest, what would be the point in this? If you are plugging in external mics and need independent control, really you are requiring a level of audio that's beyond what a DSLR provides, and should be thinking of either a professional video camera with XLR's, phantom power etc, or an external recording device, like a Zoom H4N or some other audio box that gives you proper preamps and level control.

If you are not plugging in external mics, then separate levels is largely pointless as the mic is essentially receiving the same signal for left and right channels (it's tiny and in one position). The *only* reason I can think of for independant levels is for safety, to let one side be -12dB down than the other, as a safety backup channel if levels get really loud and distort - the DLSR preamps are poor and thus can't be run at a large headroom to account for this - in this case, I can see it would be useful (and I do this on other cameras quite often).

The internal mic on a DSLR is really only for quick and dirty stuff, or reference audio. Any further requirements to that and you should probably be upgrading your audio gear.

Note - I was actually an audio guy a long time before being a photographer/videographer, and I've also got a Sony professional video camera. DSLR audio is (currently, it will probably change) poor at best.

i know, but this is the only thing that preventing me from using it on a daily basis.

Sure, but the alternative is to not have it at all until the ML team release a stable, thoroughly tested solid version, which could be a year away.

Be thankful that you at least can test it, rather than moan that it's not yet what you want it to be - there are (good) reasons for this, and we need to respect the developers about this stuff.

every time i line up a shot, change location, giving directions to the cast the camera shuts the shutter and i need to reflash

Hmm You don't need to reload ML if you just come out of live view. You only need to reload it if you turn the camera off. So in your case, stop turning the camera off. Batteries are cheap enough and last long enough that leaving the camera on (screen off) between shots is perfectly viable for most uses - I do this a lot.

i am currently running stability tests and cleaning up stuff i prepared for alpha 2.
expect the alpha 2 to arrive on xmas or before if there isnt any show stopper like the last time

Nice! 

Does the bug happen without ml too? If so, yes, we should notify Canon ASAP

Yes, all of the basic testing with this bug from us non-developers were done without ML loaded.

when I review a photo and hit the button to zoom in it would immediately zoom in all the way instead of incrementally.  It is reproducible every time I try it.  Is this intentional or just something not quite right?

This is a feature, not a bug
It enables you to check critical focus quickly, which is what it is intended for.

"Zoom In: quick image magnification: a single click can zoom all the way in, on center point, on AF point etc (configurable)."

Prefs -> Image Review Settings...

"Quick Zoom

Faster zoom in PLAY mode, for checking critical focus:

OFF
ON - zooms faster than Canon firmware.
SinglePress → 100%: a single press of Zoom In will zoom all the way in (to 100%) - on center point. Next press will zoom out (full screen image).
Full zoom on AF point: similar, but it will zoom on currently selected autofocus point.
Full zoom on last position: similar, but it will remember the last position of the zoom box."

Do read the documentation when playing and learning ML, there is a lot of stuff in there...

thanks binky38, thanks beej!
that is a workaround for the problem.
its not a "real fix" yet.

Yeah, good news, it's a step towards understanding what's going on and therefore what the proper fix would be.

I guess the thing to do is to look at what the code is doing in that specific instance, and try to determine which bits of code is responsible for the issue - once you have that, understanding what is going on should hopefully lead to a proper fix. But you know what you're doing with this, I'm sure!

but at least we now have a chance to recover from that problem

Indeed!

hover over the eos utility shutter button (camera performs exposure and focus)  but do not press the button
Open card door

Camera shuts down every time.

Good spot. Confirmed. I also tried this in manual focus mode, with back button focusing, to take focusing out of the equation.

This would suggest, as per my earlier hunch, that the issue is caused in the metering/exposure calculation stage.

Now, in Manual mode, the camera's metering/exposure calculation isn't contributing the final picture exposure, but there is metering and exposure calculating going on to determine where the exposure needle is going to be displayed (indicating possible under/over exposure with your settings).

So - I then changed the camera's half press from "Metering" to "AF Lock", to remove metering from the shutter button.

So, to break it down:
- I can turn the camera on, settings as above, with EOS utility running and metering taken off the shutter button
- take a photo (from EOS Utility)
- open the card door
- camera hangs.

- Close the card door
- half press on the camera (some metering is still done as the exposure needle is displayed)
- open the card door
- the camera still hangs.

- Close the card door
- Half press (ie mouseover) on the EOS utility shutter button
- open the card door
- The camera shuts down ok.

[AGC disable hack]

Is there or is it possible to have the AGC disable hack like the canon 550D? It was AMAZING!the canon update did not do an entirely clean audio and Im really curious about ML.

The 7D already has manual audio controls? Set it from "Auto" to 'Manual", pick your recording level, and there you go. Nothing to do with ML, the 7D already has that feature with version 2.x.

Or are you saying that there is a problem with the 7D's manual audio recording?

Quick check reveals the camera will only power off ( and disconnect from computer) when the shutter speed is 1 second or faster.
1.3 seconds and slower than that, the camera will not power off.

Had a quick try (on Mac OSX).

Here, anything down to 0"8 will turn the camera off / disconnect the computer, but from 1 second and onwards, the camera won't disconnect on opening the card door. So, it seems to be dependent on some other factors, rather than a hard limit, if you can shoot at 1 second without triggering the bug, and I can shoot at 1 second and trigger the bug...

So:
ISO 6400, no shutdown starts at 1 second and slower
ISO 320, no shutdown starts at 0.5 seconds and slower
ISO 100, no shutdown starts at 1/5th of a second and slower

All F2.8.

So it seems to be affected by exposure calculation, perhaps..? I'll see whether the aperture changing the exposure calculation has any affect next...

However, I don't think these numbers are hard and fast, as I just did some more playing, and went back to my known baseline I started with, 6400 2.8 0"8, shouldn't lock the camera up, but now does - so there seems to be a bit of varience based on some factor. I just now, at 6400 2.8, triggered the hang at 1/8 and slower, rather than the 1 second and slower it was to start with...

So I was rocking here at 6400 2.8, triggering the hang at 1/5th and slower, so the 1/5th was the hang point. Changing the aperture from F2.8 to F11 did indeed cause the camera to no longer hang at 1/5th - so preliminary tests suggest that the aperture value is also affecting when the bug is triggered.

I haven't got any predictable pattern yet, but perhaps some of those results might point towards some possible avenues of exploration to find out where the issue is occurring.

Don't see it mentioned in the thread, if it was then my bad... When I'm playing video files on the camera itself, I dial down the volume bar (displayed on the screen) and set it to zero but I can still hear recorded audio. I can increase the volume but not completely mute it when using ML.

Ok, I can confirm this. Without ML, sound can be dialled to zero, with ML, sound is always present and while the dial does change the volume, it won't go to quiet.

Have you got a solution to the issue with processing functions not being terminated when turning off the camera..?

umm.. I cant seem to find the intervalometer.. its my first time using ML, is it just me that cant find the intervalometer or was it not included in this release?

The first post in this thread details what is and what isn't in this alpha version.

i am a video guy ...but  i dont really know whats that mean....
better quality ? no compression ? better compression ? what exactly the meaning ?.....

Typically, DSLR's use fairly heavy compression. This keeps the file sizes down, at the expense of "quality". How that loss of quality appears will depend on the footage. For example, if you have an area in the frame of very high detail, like a sparkly shiny cloth, that will rob the video of much of it's bandwidth, causing other parts of the frame to lose detail and be washed out.

Increasing the bit rate means there is less overall compression and more bandwidth, which means that cases like these don't rob the video of quality and fine details. This is only one example, there are other cases where you can also see artifacts, poor compression, blocky handling, loss of detail etc.

Now - if you are shooting video and for your purposes are fine with the default quality, then you don't need to increase the bit rate.

However, there are some of us that do want this control, and to have the ability to increase the birate if the scenario requires it - so, it's a good feature to have..

* please don't beg for adding feature XYZ, it will be added as soon as it works without issues.

In general, it's best if people don't pressure the devs about what/when etc and just let them get on with it - when there is something to announce, they'll announce it...

just to keep you informed:

Thanks for the update.

at the moment we are checking which license/OS we choose for which developer.
another 500 EUR and both, alex and me will get the HEXARM decompiler for our native OS
we are still discussing the exact configuration.

Ok, time for me to live up to my promise. One donation for the 7D announcement: done. One donation for the first alpha: Will do in a bit. I'll also donate again when the beta hits, and once again when the release hits.

Come on people, let's get these guys the tools they need...

I have a small problem with the software. It installed onto the 7d perfectly and i was able to run it smoothly. But. whenever I turn off my 7d and turn it back on again the program uninstalls itself and its back to the canon V2.0.3 software

can anyone help me?

This is intentional behaviour at this time.

OK, I'm so stoked at this development. I upgraded my firmware and loaded the ML 7D alpha. It comes up but the screen is gibberish, I can go from menu item to item but it is totally illegible. Anyone else here experiencing this?

Is it the same issue I describe here?:
http://www.magiclantern.fm/forum/index.php?topic=3103.25

(Scroll down to my later posts for the cause and solution...)

When I run the update with the ML FIR file, it jumps from the "Firmware update program" screen right to the normal screen.
I checked the checksum and tried over and over again. I tried both my CF cards and with the original canon battery.

On the menu that displays the firmware version, if it just says 2.0.3 then you do not have ML running (check your installation, copied to the right place, download not corrupt etc).

If it says something like "2.0.3-ml..." then it's installed and (presumably) working fine. Press the trash key, read the manual, etc

Non-scientific test: With my two, cheap ebay battteries in my battery grip, ML loads ok.

Upps sorry, I mean Magic Zoom, not Focus Peak

I start LV and next I go to ML menu and enable Magic Zoom. When Magic Zoom is enabled then "zoom window" shows up on the screen and this window still blinking.

Yep, this has already been covered earlier in the thread...
http://www.magiclantern.fm/forum/index.php?topic=3103.msg14738#msg14738