Le jour éloigné existe, celui qui ne viendra pas n'existe pas

[bobn2]

Title thanks to Pierre Sottas.

This is the promised follow up to the 'changing the paradigm thread': http://forums.dpreview.com/forums/readflat.asp?forum=1018&thread=36912461

I'm not recapping the thread here, if you weren't in it, have a look. If you don't find it fascinating, this threads not for you. If you did, then this thread is to discuss how a camera would be designed around the new paradigm. If you don't like the new paradigm, start another thread denouncing it and I'm sure many of its proponents will be happy to argue with you. If you have queries, start a thread querying it and I'm sure someone will clarify. This one's just about brainstorming the design of a new paradigm camera. I'm designing the thread to be used in threaded view, so the discussion can be kept on one of several topics at the same time.

Firstly the discussion's split into two subthreads, design ideas that could be executed in firmware on current cameras with existing controls and displays (if the ideas are good enough, who knows maybe someone will pick them up). The second subthread is how a ground-up design for the new paradigm would work. In each sub-thread there will be four sub-sub threads to cover.

1. Capture controls
2. Operator feedback
3. Processing controls
4. Auto modes

In each one, I'm going to post a starter (unless someone gets to start a sub or sub-sub thread before me, in which case they'll post the starter, which will be cool.

So, that's it, I'm signing off this post and structuring the thread now. See you all down-thread.
--
Bob

 

[bobn2]

This sub-thread is to discuss how a firmware could be designed for existing cameras which would make the new paradigm easier to operate. The available controls and displays are:

• A shutter release button with a half-way position
• Two control wheels
• A VF display which can display f-number, shutter speed and a bar graph with, say 12 segments. If these aren't enough you can add other VF displays, as long as you can show that they actually exist on current cameras.
• A mode dial, labelled PASM.
• A button labelled 'EC' or +/-
• A metering mode button
• A button labelled AE lock (or equivalent)
• Top panel display which displays the same information as the VF display
• Rear LCD. You know what they do

plus, again, any other controls that you can show they actually exist on current cameras.
--
Bob

 

[bobn2]

This sub-thread is to discuss how a camera designed from the ground up to work with the new paradigm would work.

You can propose any displays or controls you like, the only condition is, they must actually exist (and if you're proposing something really off the wall, you should post a link to demonstrate that it does exist and is feasible). Probably worth assuming that this is a DSLR alike with an eye level VF - you can make the VF electronic, if it helps your case along.
--
Bob

 

[bobn2]

My starting proposal for the capture controls:
The front wheel controls the shutter speed
The back wheel controls the f-number

Neither wheel operates until there is a half press on the shutter release, then the displays operate (details under Operator Feedback), the AF is activated (subject to existing AF modes) and the control wheels can be operated to adjust the shutter speed and f-number for a few seconds.

The photograph is taken by fully pressing the shutter release (subject to existing drive modes)
--
Bob

 

[bobn2]

My starting suggestion for operator feedback.
The VF display shows the selected shutter speed and f-number.

The bar chart (what was - and still is - the meter display shows the exposure such that all segments lit is the total light necessary to saturate the sensor produce max count in every pixel (this is something to discuss - should max on the meter actually be some distance below saturation?). Thereafter each EV lower exposure turns off one bar so:
|---|---|---|---| means sensor (will be) fully saturated
|---|---|---|-- | means 1 EV less than full saturation.
|---|---|---|- | means 2 EV down, and so on.

In addition should the exposure be such that some of the pixels will be over saturation (blown) the exposure display flashes
if 1% or less will be blown it flashes evenly (dot-space-dot-space-dot)
if 1-2% will be blown, it flashes dot-dot-space-dot-dot-space...
if 2-4% will be blown, it flashes 3 dots space 3 dots space
if 4-8% will be blown, it flashes 4 dots space 4 dots space
if more than 8% will be blown it flashes dash space dash
The display adjusts in real time as the user adjusts the exposure
--
Bob

 

[Great Bustard]

My starting proposal for the capture controls:
The front wheel controls the shutter speed
The back wheel controls the f-number

Just like M mode on my Canon 5D.

Neither wheel operates until there is a half press on the shutter release...

Why not? The wheels operate in M mode, half-press or not.

...then the displays operate (details under Operator Feedback), the AF is activated (subject to existing AF modes) and the control wheels can be operated to adjust the shutter speed and f-number for a few seconds.

I'm not sure where you're going with this. By the way, I use the * button for AF, and the shutter for exposure lock (half press) and taking the pic (full press).

The photograph is taken by fully pressing the shutter release (subject to existing drive modes)

Of course. By why would the wheels not operate for shutter speed and f-ratio until a half-press?

 

[bobn2]

My first suggestion:

The 'processing' controls need to deliver the user's processing requirements to the in-camera JPEG engine and to a raw processor.

After capture the camera evaluates the range of values in the raw file. A user mode controls how this evaluation drives the processing. Options are:

• The greatest value in the raw file is mapped to full white in the output file, and the selected tone curve applied downwards from 'white'.
• Some fixed point in the tone curve is matched to some fixed interval below the maximum raw pixel value - so for instance 18% grey on the output is mapped to two EV below full white.

Those who know more about raw processors than me will probably come up with more sensible suggestions.

The resultant evaluation causes a value to be generated which controls the processing (which might, co-incidentally, correspond to the ISO Exposure Index)
--
Bob

 

[bobn2]

My starting proposal for the capture controls:
The front wheel controls the shutter speed
The back wheel controls the f-number

Just like M mode on my Canon 5D.

Neither wheel operates until there is a half press on the shutter release...

Why not? The wheels operate in M mode, half-press or not.

...then the displays operate (details under Operator Feedback), the AF is activated (subject to existing AF modes) and the control wheels can be operated to adjust the shutter speed and f-number for a few seconds.

I'm not sure where you're going with this. By the way, I use the * button for AF, and the shutter for exposure lock (half press) and taking the pic (full press).

Yes, a lot of people seem to do that, perhaps I should try it.

The photograph is taken by fully pressing the shutter release (subject to existing drive modes)

Of course. By why would the wheels not operate for shutter speed and f-ratio until a half-press?

Interesting, that's how it seems to me, because I tend to keep my camera switched on, so it needs re-activating by touching the shutter button most of the time. Of course it operates without half-pressing the shutter after you switch it on....you realise, I'm admitting I've never read the manual.

The point about this section is, as we know, nothing needs to change (probably) but we can't just assume - someone might have a radical and better proposal.
--
Bob

 

[bobn2]

I'm going to propose:
M - manual, user controls DOF and motion blur

S - this mode sets a shutter speed according to the FL of the lens, assuming a hand held camera, to prevent motion blur (I know that S generally means 'shutter priority' you set the shutter manually, but I want to change it round to mean S - shutter auto. User controls f-number manually.

A - this mode extracts a frame from the LV feed and analyses it to determine what kind of generic scene this is, it then sets the f-number to what it believes is a generally appropriate DOF for that kind of scene.
P - S and A together.

All the auto modes are subject to backing off their selected shutter speed and f-number if the image noise appears to be likely to be too high. How high is too high is controlled by the EC control.

Note: I don't think I'd use any of these Auto modes - I'm not totally convinced there should be any.

--
Bob

 

[bobn2]

I've got to catch a flight tomorrow, so I need to pack, so I'm turning the clean sheet design over to you. maybe there's something here to be done with a touch pad (operated with the thumb of the right hand) to control f-number in one direction and shutter speed in the other.
--
Bob

 

[bobn2]

Bar charts, histograms - what would be best?
--
Bob

 

[bobn2]

Just making up things quickly here, because I have to go and pack - maybe a touch screen on the back with a 'curves' control?
--
Bob

 

[bobn2]

I'm thinking here if you have gyros and accels in the camera, and maybe some analysis of the LV feed, you could have an adaptive shutter speed automation which selects a speed to stop shake or subject motion depending on the actual conditions. As for DoF, something like Canon's DEP mode?
--
Bob

 

[ak1999]

A relative or absolute DR meter will be a good tool for judging the effect of setting high shutter or minimum aperture setting on image DR.

 

[binary_eye]

Metering modes complicate this somewhat. But I'll offer a suggestion from the perspective of someone who uses spot to meter highlights.

How about a number indicating the distance, in EV, from saturation? For example, -10.3, 0, +5.5, etc. While it wouldn't be as quickly readable as a bar graph or analog needle, it would allow more information (and more precise information) in a smaller space.

This would allow the photographer to spot meter the brightest highlight to be recorded with detail, set the distance from saturation to zero, and be assured that those highlights have detail when processed. The spot meter could then be aimed at different areas of the scene to determine how far away they are from full saturation; in essence an indicator of "quality" or how much latitude those areas will have when processed.

 

[aardvark7]

I think a histogram is enough
to judge exposure of the shot
but perhaps a coloured bar
to indicate the noise you've got?

Starting at green up to ISO 800,
at 1600 becoming a shade of gold,
and then on to turn bright red
if the noise is likely to be bold!

Similar to a hi-fi amp,
these can only be a guide,
because personal taste is a factor,
and certain noise we can hide.

However it is done,
the issue is mainly display.
so that the LCD/viewfinder,
shows exposure the correct way.

--
2010 : My new year's resolution - to be a year of poetry!

 

[bobn2]

A relative or absolute DR meter will be a good tool for judging the effect of setting high shutter or minimum aperture setting on image DR.

Just trying to be precise about what you meant by 'DR' - an indicator of the ratio of the highest value in the capture to the noise floor (reed noise)? How would you estimate likely shot noise?

--
Bob

 

[bobn2]

Metering modes complicate this somewhat. But I'll offer a suggestion from the perspective of someone who uses spot to meter highlights.

How about a number indicating the distance, in EV, from saturation? For example, -10.3, 0, +5.5, etc. While it wouldn't be as quickly readable as a bar graph or analog needle, it would allow more information (and more precise information) in a smaller space.

I think I tend to agree. I think one can separate two things (as one can in metering)

• what is displayed
• precisely what it represents

So a display as you say would be good, and the use could (as at present) select what was being metered:

• spot
• average
• peak
• intelligent pattern

etc etc.

This would allow the photographer to spot meter the brightest highlight to be recorded with detail, set the distance from saturation to zero, and be assured that those highlights have detail when processed. The spot meter could then be aimed at different areas of the scene to determine how far away they are from full saturation; in essence an indicator of "quality" or how much latitude those areas will have when processed.

Sounds good to me
--
Bob

 

[bobn2]

I think a histogram is enough
to judge exposure of the shot
but perhaps a coloured bar
to indicate the noise you've got?

Starting at green up to ISO 800,
at 1600 becoming a shade of gold,
and then on to turn bright red
if the noise is likely to be bold!

Similar to a hi-fi amp,
these can only be a guide,
because personal taste is a factor,
and certain noise we can hide.

However it is done,
the issue is mainly display.
so that the LCD/viewfinder,
shows exposure the correct way.

That is a main concern
though another which is muttered
is need for us to learn
how to keep the display uncluttered.

People think of many
things the viewfinder should display
but less, whether any
resultant clutter will dismay.

Your thought is good, I think
that display colour which changes
displays info with less stink
than many different ranges.

--
Bob

 

[aardvark7]

Nice one bob,
just the job!

Keep it neat:
makes life sweet!

Less said,
easily read.

I'm off now,
time for chow...

--
2010 : My new year's resolution - to be a year of poetry!