I really don't get why this request has gotten such vitriol from other members here. It's a simple feature and one that makes this camera more useful and reliable.
First off. Why do we want it?
Think of a situation like this:
You have a scene that starts out in the dark, and then you go outside. You want the dark to be... well dark. But when you record with the firmware as is, then it pulls up the exposure so that the inside is as bright as the GoPro sees fit. And when you get outside it spends a few seconds in blinding white light. Before pulling the exposure back down again.
Can it be fixed in post? maybe. If the outside light-levels aren't completely clipped then some fudging can be done with keyframes to get a desired result.
Then we have the problem of noise. When shooting in dark places, the camera doesn't just add exposure without cost. It's applying gain to the signal-chain and that brings up noise. and while I can reduce the exposure in post, I still have unwanted noise... and not only that. It's a noise-floor that fluctuates. If I have an indoor scene panning across a light, we get a noisy scene where the noise-floor is high at the dark times and low when in lit times. In the same clip. This makes it impossible to make a proper noise-profile for my denoise-system because if I make the profile for the noisier frames I kill detail in the lit frames. And if I do the noise-profile for the lit part, then the noise in the dark frames is not cleaned up properly. Only option then is to make two noise-profiles and then keyframe between them in some way. This is totally unnecessary work.
Also. While 2.7K 24fps is very clean even when doing shooting in relatively dark places with this camera (and downscaling it to 1080p averages out even more noise), something like 720p 120fps has only at most 1/120 second to get it's light, compared with 1/24 seconds in the highest resolution. So while a scene might be shot satisfactoraly in 24fps. In 120 fps the camera only gets a fifth of the light to play with per frame. So it will have to increase the gain on the video signal 5 times as much as with 2.7K.
And couple high noise-level with pretty agressive compression like the h264 that the gopro uses we loose detail even if we were to adjust the exposure in post.
Compounding the issue further is that to keep with the 180 degree shutter rule we should actually only be shooting at at most 1/48 while in 24fps and 1/240 while in 120fps to have images that have the same motion blur characteristics as film. That's halving the amount of light once more.
Another scenario to illustrate the point. In a room of pitch black you film a guy that sets off a flair gun that shoots through the room in 120fps to be slowed down to 24 fps.
In the current firmware with auto-exposure it would first try to open up the shutter as much as possible (1/120) and when that fails to produce an image that the camera wants (remember, I as the DP might want it to be pitch black, but I'm not consulted by my camera in this matter) it then starts putting gain on the video signal. And at some point it will get an image that is somewhat right according to the meters, but be just a mess of noise and static that is devoid of any interest. Then when the flare shoots off, the camera is caught by surprise and is totally blinded because it has set itself up to record darkness as mid-grey. This results in a couple of seconds of total clipping and then maybe we'll get an image that I want just as the flare dies out. And all of this mess in 5 x slow-mo.
This is not fixable in post. Not in any way no... time wasted filming and storage wasted on the card.
With an exposure lock the camera would be forced to record black as black (I don't care what it's meter says, it can't look into the future to see what is going to happen, so it can complain all it want), the flare is then already exposed for, so it flies out in a beautiful cascade of fire-hazardous situations. And then it dies out. And the image goes dark again. DP is happy. And my camera now understands that I knew what I was doing.
Or during the winter I actually toyed with the idea to put the GoPro under the ice to film how light seeps down when cracked. Again. It would start out in darkness and only a while into the clip would there be any light. I never did film that, though it would be interesting to see. Because I knew it wouldn't be filmed in a pleasing manner that I could use...
I can see that adding a feature like exposure lock might be a bit daunting since it is a camera that doesn't come with any sort of viewfinder. The LCD-backpack is not included and can't be used with the battery-backpack. And the app lags and drains the cameras battery with the WiFi. And what I think they are assuming is that an exposure lock could make it so that too many users use it by accident and therefore will give GoPro bad word of mouth. Far more people than would actually love the feature. So "The needs of the many outweigh the needs of the few", so to speak. But I think they are underestimating their buyers. And I think that the feature could be hidden enough so that the normal run-gunners wouldn't have to see it (I think most people don't stray too far into the menu-system as it is).
But then again... Why did they add protune, 2.7K and cine-modes if not to cater to those with a bit more needs than the normal users?
I see a few way this could be implemented:
- While recording, press "mode"-button. The button as it is does nothing while recording, so it is open for usability. When pressed, the exposure is locked as it is right now. This way I can start recording in a situation where it is lit as it will be later, and then after locked, I can kill the lights and yell ACTION. I guess this could even be done without having a separate menu toggle. Because... how many people accidentally press mode during recording? And we could restrict it to only function while cine-modes and protune is on.
- A menu toggle that locks exposure at record start. Could be used the same way as ¤1 to shoot dark scenes that should start out dark. When record stops, the camera goes back to auto-exposure. When record starts again, it locks again wherever it was when I pressed record. This will also help keep the exposure consistent when panning over dark places or going through shorter tunnels. And your colorist will love you for it. , and it's probably the simplest one to implement.
- Double-press record. When I record, I press once for exposure-metering. Then once again to start record. A third press stops recording and then the process is looped when continually pressing record. This is so that you don't have to waste storage on black video. Since I have separated the act of exposure and recording by time.
- Long-press record. Like ¤3, but instead of clicking once and then once again. We do a long-press (like on smart-phone, press and hold for a couple of seconds) to do a light-meter-exposure lock. then hit record with a sub-two second press to start recording as normal. once more do a short press to stop recording.
- Only make exposure lock available when using LCD-backpack or the App. This way, even the run-gunners would be able to see that something is not automatically adjusting the exposure. And with the inclusion of zebra and false colors we could get really creative and finetune it all to our utmost desire. You could even start exposing for different zones. But this gets very complicated very fast and we don't want to bog down the release-date of this feature with time spent on un-needed sub-features.
- Full manual mode where I have to specify Gain and shutterspeed in menu, (does it even have a variable iris? if so, then it too should be manual). Or the gui in the app or while having the LCD-backpack on.
- A combination of all the above.
- A separate "pro" firmware that only advanced users would download and use. This off course is the least likely solution since we then get a situation where they need to keep two editions up to date and acceptably free of bugs.
- Releasing an SDK for tweakers to modify as they see fit. Keeping firmware-updating triggering off-limits so that a tweaked firmware won't brick a gopro and make it unable to reload the official firmware. This would be unheard of, of course... but then again. GoPro don't actually have to do the tweaking themselves. the tweakers and adventurous users can do just about anything they want, without having to bog down the official gopro development. And if there's a feature that tweakers come up with that GoPro wants... then they can buy the rights for that snippet of code to be included in the official code. I call these coders "Tweakers", because this would be code-tinkering that is completely sanctioned by the original developers. Not hacking. This would also keep the tweaking from being bogged down with having to reverse-engineer everything like Magic Lantern is forced to do.
I see the problem of security in this too. Since it would open up ways for people to steal the code and then sell it to competitors. But I guess it would be possible to open up some of the code for tweaking without making it too easy to steal the "classified" parts. So I file this one under plain wishful thinking... though I would love it if someone actually tried it. Sort of like the Raspberry Pi. They sell the hardware and then users can modify it with code and stuff to their hearts content.
options 1-4 are basically semi-auto features that I think are the most likely and entails the least amount of extra effort. And they are solutions that uses only the stock hardware. 8 is unlikely, and 9 is just wishful thinking.
Again. I fully understand that exposure lock should be an option, and not something that is auto-applied when latest update is released. All to minimize risks of having tons of mails going to GoPro about "Broken" cameras that won't auto-expose like they expect it to. And this could be included by the same motivation that they included cine-modes and protune.
Another feature I would love that is related to this is some way of limiting the auto-exposure. Since most of the time the problem of not being able to fix in post comes in when shooting in the dark we could have a limiter that refuses to gain the signal after a certain point and refuses to open the shutter beyond 1/48. This would also help out us that does weird shooting and we'd still have normal auto-exposure when shooting in the normal sunny days. Though exposure locking does take priority if there is a choice between the two.
Then I would also like to share a couple of correspondenses I've had with gopro. I have basically asked about exposure-locking twice now by mail. And the first answer I got in march 5th 2013 was a bit puzzling:
Thanks for your interest in GoPro!
The features you mentioned in your email are, like you said, almost a broadcasting standard. As a result, at this time they are not likely to be incorporated into the camera. While we have ideas like this on the radar, to get all of these features packed into a $400, waterproof, lightweight, and durable package is not extremely likely at this point.
We will add your recommendations to the list of requested features. This has been requested before and I am sure will be requested again. Thank you for taking the time to provide us this valuable feedback.
Again, why then include 2.7K and Protune and other production level features?
And then a couple of months later I asked again about the same thing and in june 5th 2013 I got this reply:
Thanks for contacting GoPro Support. Sorry to hear about the problems with your HERO3 camera.
We appreciate the ideas you've proposed. Exposure controls may be something we'll implement in the future.
To be kept up to date on any new product releases, please be sure to sign up for our online newsletter:
Please let us know if you have any further questions
So the possibility has been upgrade from a total no, not at this price-point, to a "Exposure controls may be something we'll implement in the future.". And following this trajectory it could even be done by the next firmware... Again wishful thinking. But at least they have loosened their stance about it and I am glad that I am not the only one that has been thinking about it.
So after this overly long and rambling first post I'll say good bye for now, and please don't strike down on feature requests... sometimes a feature might have been overlooked by the developers as a possibility, and if no-one's telling them about it... then why should they look into it? If it wasn't for feature requests I doubt that gopro would even exist at all.
Also, I know that GoPro's are, despite their name, essentially mainly selling to regular consumers, not pro's. So we aren't the biggest chunk of the market. But the only competitor that has what GoPro has along-side with the ability to exposure-lock is the Novo... and that is literally "just" a tweaked out gopro that isn't waterproof (since they implemented a lens-mount) and rents for twice as much per day as a stock Gopro 3 Black, And I don't even know if the lenses are included in that package... Novo won't sell these suped up versions. They only rent them. So there's not that many others that we can ask this feature of... and it's just firmware... not a full hardware-mod that the Novo is. And it's a simple feature too, most likely. It's not at all like if we would be asking for 300fps 4K continuous shooting. Because we would probably have to shoot not just under water for that to work but in a tank of liquid nitrogen. Exposure lock is just temporarily suspending the auto-features. If anything, it should be even less taxing for the hardware then full-auto. Again... I fear I am rambling. So I'll stop right here.