Will there be a need for cameras at all once we have 4k video? Probably within 5-10 years we will have consumer level 4k video cameras. At that time, what is the point of still cameras? You can pick any frame in the video and it'll be a perfectly crisp image, is it not?
Yes, still cameras can do things 4k video can't. I imagine that future cameras will go in the direction of the Canon 1D C, which has still photography and 4k video in the same body.
From a photography standpoint, the advantages of shooting stills instead of frame-captured video include:
- Higher resolution. 4k video is around 8 megapixels, compared to the 16-36 megapixels we find in standard DSLRs.
- Faster shutter speeds. For video, the shutter speed is normally fixed to twice the framerate, so a 30 fps video would be shot with a 1/60s shutter speed. The effect is that movement will appear smooth in the video, but individual frames will be slightly blurry, which you may not want in a photo. If you increase the shutter speed enough to freeze the motion in individual frames, movement in the video becomes strangely staccato, sort of like watching movement under a strobe light. So even if you use 4k video for both video and photography, you will often need to decide ahead of time whether the recording is intended for video or stills.
- Slower shutter speeds. For a 30 fps movie (for example) you can't possibly use shutter speeds slower than 1/30s. For photography, you will often want longer than that, e.g. for deliberate motion blur, light painting, low light, astrophotography, etc.
- Simpler exposure. Photographers as well as videomakers often use aperture for creative purposes, e.g. to get a shallow depth of field. In still photography, you can simply adjust the shutter speed to get the right exposure. For video, on the other hand, the shutter speed is a function of the video framerate, so in bright light you have to resort to ND filters.
4k video can still be useful for photography. The advantages of video over stills include:
- Higher frame rates. Even the fastest professional DSLRs top out at 10-12 fps, while video routinely gives 30 or 60 fps. Although you trade off some image quality, the quality may be good enough for the purpose, and these kinds of speeds are not available in normal cameras.
- "Micro expressions". A wedding photographer who did an early review of the Canon 1D C found 4k video useful for capturing what he calls "micro expressions", tiny changes in facial expressions that can make two photos taken seconds apart look very different. With 4k video, he could simply record a longer segment and pull out the frame where the bride and groom look their best.
In conclusion, there's no reason for cameras to disappear, and photography and video/filmmaking are still very different processes.
But 4k video integrated in cameras can be an additional photographic tool that will be useful in some circumstances.
2In the note about "micro expressions", it makes it sound like it would be simple to pull the best frame out of 4k video. I can't image looking at a 30 second clip and trying to pull the best frame out, I have enough trouble when I'm in burst mode on my DSLR and get 5fps!– dpollittMay 3, 2013 at 14:52
@dpollitt I don't think it would be too hard; I assume a photographer would mostly use the 30fps option when things are happening too quickly for burst mode to keep up. Watch the 30 second clip, find where something interesting happens (the tennis player hits the ball, the bride and groom start kissing, ...), and you may be down to a couple of seconds and 50-60 frames to look at in more detail. A bit of work of course, but hardly infeasible. May 4, 2013 at 16:06
4K video cameras will not be able to replace still cameras.
First, 4K video resolution is only about 9 megapixels. Megapixels aren't everything, but in general more is good, and modern mid-grade to high-end still cameras clearly take advantage of more than that.
Second, and more importantly, video cameras are optimized for video, and that has needs which are different from those of still photography. Each frame will have a shutter speed as needed for the video framerate, and furthermore digital video will be compressed across frames. And while film-production-level digital video cameras may record in some form of RAW, that's unlikely on consumer video cameras within any near timeframe. Plus, a video camera user interface generally has still photography only as an afterthought. An ultimate convergence device may make it possible to set the basic photographic parameters of shutter speed and aperture in a still photography mode, but there's a lot of non-overlap which would be hard to cover.
This really is an interesting and rather loaded question. For one thing, 4k video is HUGE compared to photos. Full quality 4K video that would really maintain quality at a per frame level takes around 250GB to 500GB per hour of video. Taking photos only require one particular frame to be captured and can capture at much higher quality. (4k UltraHD is only 8.3 megapixels which is already rivaled by an average camera phone for still images.)
As far as the devices used, sensors that are used for still cameras can also be used for video and we're already seeing a lot of convergence between video and photography across the market, from both consumer gear and smartphones all the way up through professional gear like the Canon 1Dc. The main reason they still operate in separate video and photo modes is simply that of user needs and quality.
Across the entire spectrum, much higher quality can still be achieved by using a sensor and image processing hardware for a single frame at a time. The image quality as well as the actual number of pixels are both much higher and the time to process and store that much information is too long for video. Even a high end professional still camera normally is limited to about 11 or so shots per second at full resolution which falls fall short of the 24 - 60 shots per second needed for common video formats.
More specifically on the professional side of the spectrum, the kinds of things you need to be able to do and adjust for are very different between video and photo. The concepts are similar enough that you see cross over devices like DSLRs with high end video capability, but the way in which you shoot is still very different.
Video has a fixed shutter time and requires keeping a constantly good image while adjusting on the fly for focus on moving subjects without hunting for focus or lowering the mirror. Still shots on the other hand are concerned about getting a particular moment just right and getting really tight focus and shutter timing control to get exactly the shot that is desired. They can benefit from higher end AF sensors that require the mirror to be lowered.
Is there necessarily a reason that a lot of the technical limitations can't eventually be overcome? No, there isn't, but the artistic choices that professionals have to make and the selection of particular images for stills that anyone needs to make are going to continue to make video and still photos require separate modes.
No, video is shot with a fixed shutter rate, which limits the speed. Sometimes a photographer wants a long exposure, you cant do that with video.
Still and video are two separate arts.
1Sure you can do a long exposure with video: just average n frames to produce an n/30-second exposure. It's not a perfect solution because there are always little gaps, but it's darn close. If actually done this, so I can say for a fact that it works. As a bonus, you get much higher pixel depth because the noise smooths out. A few weeks ago, I took a 30-second video with an iPhone in near darkness. After averaging 900 frames and converting to 32-bit depth, a simple curves/levels adjustment brought incredible detail out of the photo. May 29, 2013 at 4:10
That's fascinating. What software would you recommend to merge the frames in the video into a photo? Would it work handheld? And, last, could you post the video and the resulting photo? Thanks. Dec 13, 2013 at 13:56
We've already seen a lot of convergence between video and still cameras. Still cameras have started to really be useful for serious video, and video cameras have been able to take still photos for ages. Some cameras (Sony Alpha, I think?) have started to use video-like capabilities to make taking photos easier by recording frames before and after you trigger the "shutter", so that if you hit the button a little early or late, you can select a different frame and "never miss a shot."
There's no reason to believe that this convergence has reached an end: it seems inevitable that the video capabilities of still cameras will continue to improve and vice versa, and also likely that some features on each side will bleed over into the other. That said, video and still photography are different, and it seems likely that we'll have cameras that are optimized for each application for a long time to come. Just grabbing a frame from video is not the same thing as taking/making a photograph.
For snapshots of the kids, yeah, video captures just might work if you don't care too much, which is good enough for most people. But, if you are shooting wildlife, sports, landscape, architecture, people, low light, long exposure, etc... you'll really need the creative control that shutter and aperture give you.
So the 4k might be ok for a lot of snapshots, it isn't even close enough to what a craft photographer needs.
You can totally do low-light and long-exposure with video. All you need to do is post-process the video. It's actually very easy. May 29, 2013 at 4:13
By long-exposure I mean on the order of 1/2 second to 4 minutes... May 29, 2013 at 10:36
Ya, that's totally easy to do with video. A 4-minute long exposure? At 30 fps, that gives you 4*60*30 = 9600 frames to average. Quite easy to do using software. As a bonus, you can blend them nonlinearly if that gives a better effect, which is completely impossible with a traditional long exposure. As another bonus, you end up with an extremely high color depth, which smooths out sensor noise amazingly well. The only thing you sacrifice with video when doing a long exposure is the xy image resolution. May 29, 2013 at 18:22
ahhh, I see, yeah, that's cool. May 29, 2013 at 18:56
Your concerns in photography: shutter, aperture, ISO.
- If you want to freeze action, you need fast shutter.
- If you want to blur action or use available light, need to use slow shutter.
You do not have these on today's video cameras.
But there is one additional thing that high resolution does not handle: higher pixel count on the same capture area (e.g. Full frame or APS-C) means that each pixel color value will come from a smaller sensor area. Signals from the smaller area need to be amplified more (because the amount of photons per area is much lower) so basically with more pixels you end up with more noise per pixel.
Obviously, you can remove that noise, but at the very same time you are blurring the image, ending up with much less effective resolution...
Many times, a photographer does not need high resolution, he needs less noise. E.g. I do snapshot (non-posed) portraits, with very shallow DOF, in the dark. Obviously, I need fast enough shutter because the woman is just moving, talking, smiling. My aperture is wide open. So I have to dial up ISO as much as possible. Can you do photos with 4K at 1/200, 1:1.4, ISO 6400? For me, life starts there :-).
I do not think that 4K will replace cameras in general, but you might find it useful in certain situations. Take a footage with a normal camera about your area of interest, and try to cut one image from that video. You will see that most of the time it is impossible, because unless your model is still and your camera is still and everything is still, motion will be encoded as motion in the video (see the MPEG-2, MPEG-4 details). You will see blocking, optimizations for encoding, etc. Unless you are able to record raw (uncompressed) video, a feature you will not see in any consumer- or prosumer-category video camera...