I would love to be able to record 1+hour of audio using existing onboard codecs. This would be very so much simpler to create, and yet would still have a lot of usefulness, so the cost/benefit ratio is much better than MP3. I think pushing the MP3 angle is destructive - sure, it is somewhat better, but it is obviously not triggering the interest of the developers.
So I would like to emphasize that a small tweak to extend the lengths of my cameras existing audio recording features should not be very hard, and would be immensely useful. Take my example below:
Extended audio recording is the feature I really miss from my Pentax Optio S4 - I was able to hold down the power button for 2 seconds, then it goes into audio mode. Then I could hit the shutter and it will record till battery or SD runs out - with display off that can be hours! It is so super handy for recording lectures or presentations or meetings. Sure, the onboard mic and the audio quality (8khz PCM) are not great, but for recording voice conversations it doesn't matter, and hey - I have a Dictaphone/voice recorder in my pocket!
Not all that many MP3 players have this option - at least not the MP3 players I know/own, so it is not simply a matter of "use a more appropriate/readily available device". And in response to the argument that "this is not really what a camera is for", I would say that it is now assumed that a camera's role is also to record video (and voice memos), so I am proposing a subset of this existing feature, not really a new feature.
And the clincher is that my Canon camera (a720) can already record audio (11khz, 8bit, mono (20kbps)), but the voice memo is cut off at 30 seconds, and the video comes with much larger file sizes and battery drain. So surely it would not be hard to extend the voice memo length, or make an option to record video that is is black (and ideally 0.1fps or 1x1 pixels).
Extending voice memo would be preferable, and probably easier, though it might not play back onboard the camera. Hacking the video to record black may be not too hard, and there are already video recording hacks. Black video would also play back on board the camera. Ultra-Low-FrameRate or Ultra-Low-Res video might be harder to implement and might not play back on board the camera.
Maybe it could also be looked at like this:
Step Feature Difficulty Benefit
1 Record longer audio using existing onboard codecs Easy Very High
2 Play back camera-recorded audio Easy-Moderate Low-Medium
3 Play back third-party audio in onboard codec Easy-Moderate Medium-High
4 Play back third-party audio in other codec (MP3) Hard Medium-High
5 Record audio in other codec (MP3) Hard Low
Looking at this breakdown, I have a few observations: each step could be done by itself and still provide value; step 1 definitely has the best cost/benefit ratio; step 5 only has a benefit of Low because I suspect the hardware will only support lower quality audio audio anyway which already has a small file size, so compressing it is not that important.
