A hotly contested arms race in the mobile world is the number of megapixels in a phone camera. This has generally tracked the megapixel race in the stand-alone digital still camera market.
The fact that people always carry their phones makes them an extremely popular device for spontaneous photography. Hence, back in 2005, Nokia became the worlds largest digital camera manufacturer.
But how many people use the camera in their phone for serious photography, and what do they do with the images? Print them? Blow them up to a poster size?
Not at all often in either case would be my guess, much less so the poster option.
Conventional wisdom on digital still cameras says that anything above 6 megapixels is comfortably capable of producing a 10″x8″ (large size) print.
Viewed on a TV or monitor, the number of pixels displayed is much lower (around 1-4 megapixels) – and when only viewed on the phone or camera itself, typically around a quarter of a megapixel. Past a certain point, what determines image quality far more than the number of pixels in the sensor is the quality of the optics (i.e. lens) and the image processing algorithms applied to the output from the sensor.
Many professionals are perfectly happy with the images produced by 6-10 megapixel digital SLR models.
When squeezing a high resolution sensor into a small device the individual pixels need to be made smaller. Smaller pixels mean less light getting to each one and the result is much noisier images in low-light conditions.
So if you like to take pictures indoors, or on a night out, then a really big megapixel number on a phone camera probably isn’t a good thing for you.
Also, there’s usually limited room and budget for really good optics in a small device. For that reason, most phone cameras have to do without an optical zoom. More megapixels also mean more processing and thus more battery usage, as well as larger file sizes which mean more storage required and greater time to transfer. None of these are good things for the typical user.
So, with all of this well known by mobile device manufacturers, why do we continue ever upward in the race for more megapixels, with the likes of the Sony Ericsson Satio at 12.1 megapixels?
It should be noted that the engineers at Sony Ericsson (some of whom I had the pleasure of meeting in Stockholm last week) appear to have done an excellent job with the optics and the software supporting their sensor, such that one detailed review claims it has “potentially the best camera ever fitted to a phone”.
That said, it clearly suffers from low-light noise, and the default image capture setting is actually 9 megapixels, not 12, suggesting that they aren’t really all needed most of the time! The major reason for adding more megapixels is the “more=better” algorithm in the mind of many consumers when it comes to the numbers on the box.
What does all this have to do with Symbian? Well, to the extent that the platform could define software for the camera sub-system, should it be optimized for a pragmatic megapixel range and an attempt made to educate consumers about the pros and cons of technology choices?
Or, alternatively, should the OEMs that use that platform to build products decide what they think they can sell and drive the architectural requirements and consumer messaging accordingly?
Since the platform is primarily driven by OEM contributions, the latter seems much more likely than the former. However, different OEMs may have different product plans and policies in this area which are likely to lead to some fragmentation. Is this an area where fragmentation is inevitable and if so, how does that affect application developers and the services they can offer to end users?
I think one of the key things for a smartphone’s camera is to take photos quickly, something that the satio does very well. Phones are used for spur of the moment photography when a camera is most likely never available in normal circumstances.
Having spent a week or two with the Satio, it’s close to being a compact camera replacement. If the low light issues were addressed (indoor, party photos, red eye reduction are key areas for a compact) then it’d be a sure fire winner.
I also think that the camera is an easy element for a manufacturer to sell, evidence can clearly be seen just by taking a photo how good a product is- whereas battery life, exchange support etc. can only be partly conveyed.
One day, when someone can cram in a big sensor to catch more light, things might look up, but for me anything above 5MP I don’t care about the MP’s.
Thanks, this really proves my point almost totally. With the great camera software and optics in the Satio, I believe it would actually be a better product if the camera were 5Mpix rather than 12Mpix. Perhaps we’ll get to test this theory on some of SE’s future Symbian products.
The problem is of course one of marketing and differentiating the product.
Thing is, I caught myself going round saying, ‘ooh, it’s got a 12 megapixel camera!’ even though I actually think it’s kinda irrelevant.
Social comments and analytics for this post…
This post was mentioned on Twitter by kasidit: what’s more is not always better – pixels – http://blog.symbian.org/2009/12/15/why-more-is-not-always-better-part-ii-megapixels/...
one of the reasons i like cameraphones are their sharing abilities. and high megapixel images are not good for sharing . typical size of high details photo in my N86 is ~ 1.5mb. Now if i want to quickly twitpic a pic or just share to facebook its annoying in typical edge connection. thats why i usually resample my photos to 1024×768 and upload them . and in flickr they dont appear any worse than full 8mp. (my flickr page : http://www.flickr.com/photos/45548273@N02/ )
i think symbian should develop easy sharing feature in the camera ui that will automatically do the down sampling according to the service used (eg. 2mp resolution if uploaded to flickr , 800×600 for facebook etc.) . this would help a lot for many casual user.
BTW, all the greatest cameraphones in the world runs on symbian platform , (N86/innov8/satio/) . only exception is pixon12 which is not smartphone. cameraphones are symbians strongest point so dont ignore it.
>> i think symbian should develop easy sharing feature in the camera ui that will automatically do the down sampling according to the service used (eg. 2mp resolution if uploaded to flickr , 800×600 for facebook etc.)
This is a really good idea. Why not post it at ideas.symbian.org – it could well get implemented in the platform.
I’m curious, what does ‘optimized for a pragmatic megapixel range’ practically mean?
One factor you don’t cover (I think) is how the underlying hardware market works. There is strong competition between suppliers to handset manufacturers for every single hardware component and these manufacturers will not be the sole targetted customers. The margins in hardware are always going to be greater at the high-end.
I do expect the megapixel war to continue for a little while, and Symbian will need to work as well if not better than it’s competitors at the high end I suspect.
I’m not disagreeing with your point, but it’s worth considering that very often the hardware determines the software rather than careful consideration of the practicalities of the usecase.
>> what does ‘optimized for a pragmatic megapixel range’ practically mean?
That would take some working out, however, there are a lot of camera features that are quite sensitive to the size of the images produced. Supporting things like smile detection, or just multi-shot capture takes a massive amount of RAM at 12 Mpix. If you’re only looking to support 3-5 Mpix you can get more advanced image processing in there. There’s an image editor in the platform too, what features can realistically be supported at different image sizes?
>> I do expect the megapixel war to continue for a little while, and Symbian will need to work as well if not better than it’s competitors at the high end I suspect.
I agree, but I don’t think it has much to do with component suppliers having other customers. The cameraphone market dwarfs the DSC market to such an extent now that I’m sure the suppliers would do a one-off for the likes of Nokia. People will pay more for a high megapixel count, so it’s worth the OEMs pushing that particular “more=better” myth and making high margins on that high-end hardware.
As I said in the post, I expect the platform software to be primarily driven by the OEMs product plans (and thus hardware choices) and as a consequence there’ll be quite a lot of differentiated adaptation in this area, meaning that 3rd party developers won’t be able to easily target a significant market for a more advanced camera application or image processing software.
There a market to build a device around a sharing camera concept, lower megapixel, but inbuilt camera to facebook/twitter etc one touch sharing. A device that is purely built around the social web and is not bolted on.
I loved my Sony Ericsson which I think had about 5 mega ixels. I swapped it for an iPhone when my company decided we should al use Apple but here’s the thing for me – I use my phone extensively as a camera because I travel. I take pictures that I mms or email to my children. With the iPhone after 4 in the afternoon that is almost impossible even on a good day. What I really want from a phone is that digitisation of everyday life poetry that Nokia talked so well a few years ago – bring it bac!
Yes, Apple did a particularly lame job on the iPhone camera. I think it’s there more for a box ticking exercise than anything.