K5 - Exposure bracketing for HDR

bold There'd be absolutely no point whatsoever in doing that. The only reason to create multiple images is to capture information that you couldn't capture with one capture alone. i.e it exceeds the dynamic range of the camera

And given you can recover up to 5 stops with the K-5 there's little point in taking multiple images 1 stop apart. I could see the point in taking images 3 or 4 stops apart though


It's wouldn't even be 'pseudo' HDR, it would still be HDR HDR is simply the reduction in dynamic range of an scene to produce the commonly known effect. The number of images used to produce the result doesn't determine whether it's HDR or not.

It's not a technical matter, Smeggypants, it's an aesthetic one. Doing all these things gives different pictorial effects, some of which can be very attractive.

1EV or eben 0.5EV steps, usually seven exposures, processed through Photomatix can give some very powerful results.

This is an example, using a preset called "grunge".
https://www.pentaxuser.com/photo/user-5/house-of-fantasy-48967

What you're talking about there is NOT taking a single image and copying it multiple times as Pentaxiphile was talkign about, you're talking about multiple exposures. You'll gain 6 stops more dynamic range with 7 exposures 1 stop apart. Which on a k-5 would equal 20 bits.

You cannot squeeze 20 bits from a 14bit K-5 RAW. It's against the laws of physics to create information where there is none. That's why you can't recover blown whites or blacks below the dynamic range of the sensor. So it's still pointless from a dynamic range POV to split a simple exposure into multiples.

I'm not arguing against what individual pieces of software can achieve like in your example. Personally I find the HDR effect as in your example not to my taste. It looks cheesy to me and is a cliched effect now. Each to their own though

Actually, I was talking about all these things, but no matter.

As an aside, 'bits' do not = EV steps of dynamic range. It's coincidence that the K5 raw files are 14bit and can record a dynamic range of 14EV.

I've just finished reading a book called 'Practical HDR' by David Nightingale - recommended, good examples, nice explanations of practical purposes behind differing HDR techniques (just as we're discussing here) and an overview of using the main software products.

He talks about single image HDR towards the end; he regards it as mainly a means to 'unify the tonal range of an image' (ie reduce it !)or to allow creative HDR effects and increased detail in an image when it had not been possible to shoot a number of bracketed exposures.

He specifically says '...you only have a single exposure to work with, so you ideally need to make sure that the dynamic range of the scene is lower than your cameras range from the outset - any shadow or highlight clipping at the exposure stage will be carried through to the final image.'

The implication from this is that if the DR is already full across the histogram, then just producing additional underexposed & overexposed copies for HDR processing is of no benefit. The point of the bracketed exposures is that each one is optimised for a certain tonal range, allowing the HDR software to use the best bits of them all. The different exposures from a single shot situation will just compromise the quality in some of the tonal areas.

Lee Frost is a big fan of using multiple images taken from a single RAW file to add punch to monochrome images, especially portraits. You can create effects from a single RAW file processed this way that are difficult/impossible to create otherwise. HDR can be used for more things than simply reducing the dynamic range. Smeg, you're wrong that it is pointless.

K.

Just for clarity, my post above does support the points made by Smeggy in relation to multiple processings of a single shot for further HDR work.

The author of my book does say that you can use a single image file (with a good DR) and get good HDR effects from it. You probably don't need specialised HDR software for this now - the Topaz 'Adjust' plugin provides this sort of option, and I've just seen that a new release of my favourite RAW converter, Sagelight 4, now includes an 'HDR' module, but only for working on a single file. The Sagelight creator says that 'HDR' has become so synonomous with the highly stylised overblown look, that is now what people think of when thinking 'HDR'

But real HDR is not this at all - it is simply the merging through the 'tone compression' technique of differing exposures of the same scene.

Yes, I'm sure the Lee Frost has been able to get these good results that way - what would be interesting to know is, did he use the multiple processings simply to extend the DR of the photo he had to normal limits, or did he take it beyond this (ie into clipped shadows and highlights).

If the former (the shot was within the cameras DR anyway), could he not have more simply stuck with the single original image and extended its DR in the usual manner (levels or curves) and then been able to get more or less the same good results from his HDR processing?

If the latter (the image was blown from the start) then what would be the point in pulling down the clipped highlights or pulling up the clipped shadows? There's no detail there to contribute anything of quality to the HDR output.

Looks like I'm with Smeggy on this one.

As to my understanding 6dB = 1 f stop = 1EV step. 6dB = 1 bit

OK, it's not exactly 6dB, it's about 6.02dB to be more exact.

If I'm missing something please clarify

here's some info that show 1 bit = 6.02dB = 1 f stop = 1EV

http://www.productionapprentice.com/featured/the-truth-about-video-gain-and-how-...

http://www.video-editing-made-easy.com/camcorder-light-sensitivity.html

I think these 'bits' and EV matters are important when shopping for a new camera, but what counts in deciding what approach is realistic in HDR processing, then the key indicator is the histogram - for the original shot this will tell you if the camera has 'got it in one', and if not then you could start spot metering around to check the extent of the range of tones present.

The camera's EV range won't come in to it here - just need to know how many stops each way it takes to get nice highlight and shadow detail at each extreme.

By the way, Smeggy, what is the dB unit - sounds like decibel? Haven't hear of it in photography before.

Whatever it is, it's nothing to do with bit depth or dynamic range. A bit is not a measure of dynamic range. I believe more bits are better, in terms of capturing more colour and tone information, but not in a simplistic 1bit = 1EV way.

Yes dB is decibel ( 1/10th of a Bel ). It's used for signal calculations, including electromagnetic signals ( which light is )

You won't see it used for anything to do with photography though, but I used it to illustrate how 1 bit = 1EV = 1 Fstop

When you change one whole F stop or 1EV you either double or half the amount of light captured.

Adding 1 bit or subtracting 1 bit from a digital signal either doubles or halves the maximum signal. In decibel terms this equals a 6.02dB change.

Therefore if you add another bit to the system you add another stop or an extra EV of dynamic range. That's why the K-5 has a max dynamic range of 14 stops, it's 14bit. it's no coincidence.

When you increase ISO you are increasing the amplification of the signal from the sensor. Doubling the ISO = a 6,02dB increase in gain. i.e doubling the signal.

But when you do this you double the level of the noise floor of the sensor, and this reduces the dynamic range by 1 bit, or 1 stop or 1 EV.

And is why images get more noisy as you increase ISO.

Hope that helps

Whatever what is?