PDA

View Full Version : E-M5 ISO bracketing


David Morison
12th November 2012, 01:48 PM
I am not sure whether I am doing something wrong but if I set ISO bracketing at 3 exposures at + or - 1EV I get 2 correct exposures and one exposure at + 1EV. this occurs whether I choose an ISO value or the camera does it via Auto ISO.

Even stranger effect occurs when I download these to my i-Mac - I just get three identical exposures.

What is going on? I can find no reference to this anywhere on the web - what am I missing?

Help please

Thanks

David

jdal
12th November 2012, 02:28 PM
I have no idea what's happening re the i-Mac.

However, i think the problem when you're shooting is that it needs to reduce the ISO to achieve an EV of -1, but it can't because your ISO setting is too low. This may happen with Auto as well, because auto uses as low an ISO as it thinks it can get away with. Try it with the ISO at 800.

Ulfric M Douglas
12th November 2012, 04:54 PM
As Papa Pig would say ' "I'm a bit of an expert at olympus ISO-bracketing..."

Let me enlighten you.
Assuming ISO-bracket is chosen for 1EV ;
If you use it with Auto-ISO (which often chooses minimum ISO = 200) or minimum ISO (200 again) the feature can only give you at the most minus 1EV and plus 1EV until it reaches either end of the ISO range.
At ISO 200 it gives you these three ISOs : 200 200 400
At ISO 400 it gives you these three ISOs : 200 400 800

Did I explain it enough?

Hence losing the ISO100 Jpegs from the older Pens is a terrible thing for those of us who like ISO-bracketing!
Obviously the execs who made the decision (came with the e-pL2 first) are not sufficiently knowledgeable.

Even stranger effect occurs when I download these to my i-Mac - I just get three identical exposures.
That's because the RAW files are identical, the Jpegs are 'simulated' for the top and bottom ISOs.
For RAW the feature is completely useless.

andym
12th November 2012, 05:02 PM
I've never used ISO bracket on any of my cameras so I am interested in what people use it for or any advantage over exposure bracket.I understand it will give a constant shutter speed but most of my bracketed shot would be on a tripod.

Come on someone give me a reason or am I just being naive or a doughnut:confused::confused::confused:

David Morison
12th November 2012, 05:16 PM
Thanks John and Ulfric, that explains it all, looks like I will need to use jpegs when I need to ISO bracket.

Andy, I find the ISO bracketing essential when photographing wildlife, especially birds in flight which are difficult to expose for, as I need to catch the moment and there is only time for one shot. In addition 99.99% of my photography is carried out on the hoof without a tripod so trying to get 3 identically framed images for HDR processing is impossible with AE bracketing hand held, if I'm correct.

Regards

David

PeterBirder
12th November 2012, 10:06 PM
Thanks John and Ulfric, that explains it all, looks like I will need to use jpegs when I need to ISO bracket.

Andy, I find the ISO bracketing essential when photographing wildlife, especially birds in flight which are difficult to expose for, as I need to catch the moment and there is only time for one shot. In addition 99.99% of my photography is carried out on the hoof without a tripod so trying to get 3 identically framed images for HDR processing is impossible with AE bracketing hand held, if I'm correct.

Regards

David

Hi David.
This thread raises some interesting points regarding our understanding of how our digital cameras actually work.

I don't believe AE bracketing will be any more difficult than ISO bracketing. The camera still has to take three consecutive exposures whichever parameter is being bracketed. I would think if you shoot in burst mode at 9fps you should get three frames identically framed.

The other option surely is to "shoot" raw and simply process the raw file three times with different exposure settings.

What Ulfric says about the in-camera jpeg processor "simulating" the higher and lower ISO is entirely correct. Camera manufacturers fool us into believing that you can change the ISO sensitivity of the camera sensor by giving us an ISO adjustment. This is not true! The sensitivity of the actual sensor is essentially fixed (equivalent to about ISO 200 in the case of the E-M5) as is the amount of noise it produces. When you change the ISO setting to a higher value the in-camera processor amplifies the signal from the sensor (mainly digitally), unfortunately it also amplifies the noise too which is why pictures appear increasingly noisey as you increase the ISO setting. What matters here is the signal (the picture data) to noise ratio. If the signal is small either due to low light or under exposure the signal to noise ratio will be small and noise will be visible in the final picture. The reason the E-M5 has such superior "high ISO" capability compared to previous Olympus cameras is that the new Sony sensor used has a significantly lower noise generation characteristic.

Your questions certainly made me think about these issues so thanks for that.

Regards.*chr

jdal
12th November 2012, 10:47 PM
...
I don't believe AE bracketing will be any more difficult than ISO bracketing. The camera still has to take three consecutive exposures whichever parameter is being bracketed....

Not quite. In the case of ISO bracketing it takes one exposure. It creates 3 jpegs from it and 3 identical raw files. With the other bracketing types 3 exposures are necessary.

I'm loth to say ISO bracketing is all done in software, it may be some fancy readout of the sensor thing, but in good old fashioned terms it only takes one photo and produces 3 different jpegs and 3 identical RAWS (if setting is jpeg+raw).

This is different to changing ISO 3 times, which produces different RAW files.

Interesting topic!

Zuiko
12th November 2012, 11:04 PM
Not quite. In the case of ISO bracketing it takes one exposure. It creates 3 jpegs from it and 3 identical raw files. With the other bracketing types 3 exposures are necessary.

I'm loth to say ISO bracketing is all done in software, it may be some fancy readout of the sensor thing, but in good old fashioned terms it only takes one photo and produces 3 different jpegs and 3 identical RAWS (if setting is jpeg+raw).

This is different to changing ISO 3 times, which produces different RAW files.

Interesting topic!

If it only takes one exposure but provides two others by essentially processing in-camera, surely it is no different from processing 3 different exposures from one file in Photoshop? In fact, I would have thought doing the same from one raw would give you more control and theoretically better results. :confused:

Zuiko
12th November 2012, 11:06 PM
Hi David.
This thread raises some interesting points regarding our understanding of how our digital cameras actually work.

I don't believe AE bracketing will be any more difficult than ISO bracketing. The camera still has to take three consecutive exposures whichever parameter is being bracketed. I would think if you shoot in burst mode at 9fps you should get three frames identically framed.

The other option surely is to "shoot" raw and simply process the raw file three times with different exposure settings.

What Ulfric says about the in-camera jpeg processor "simulating" the higher and lower ISO is entirely correct. Camera manufacturers fool us into believing that you can change the ISO sensitivity of the camera sensor by giving us an ISO adjustment. This is not true! The sensitivity of the actual sensor is essentially fixed (equivalent to about ISO 200 in the case of the E-M5) as is the amount of noise it produces. When you change the ISO setting to a higher value the in-camera processor amplifies the signal from the sensor (mainly digitally), unfortunately it also amplifies the noise too which is why pictures appear increasingly noisey as you increase the ISO setting. What matters here is the signal (the picture data) to noise ratio. If the signal is small either due to low light or under exposure the signal to noise ratio will be small and noise will be visible in the final picture. The reason the E-M5 has such superior "high ISO" capability compared to previous Olympus cameras is that the new Sony sensor used has a significantly lower noise generation characteristic.

Your questions certainly made me think about these issues so thanks for that.

Regards.*chr

Hi Peter,

Deja vu! I remember you explaining this to me last Friday! *chr

PeterBirder
13th November 2012, 12:10 AM
Not quite. In the case of ISO bracketing it takes one exposure. It creates 3 jpegs from it and 3 identical raw files. With the other bracketing types 3 exposures are necessary.

I'm loth to say ISO bracketing is all done in software, it may be some fancy readout of the sensor thing, but in good old fashioned terms it only takes one photo and produces 3 different jpegs and 3 identical RAWS (if setting is jpeg+raw).

This is different to changing ISO 3 times, which produces different RAW files.

Interesting topic!

Thanks John, I just checked what happens on my camera again, I did check before posting but got slightly different results.

You are partly correct it does only take one shot. This time I shot raw only in ISO bracket mode. When you then view the preview on the camera (which is of course jpegs) you get three clearly different exposures and the info display shows three different histograms and ISO figures.I then loaded the raw images onto the PC and viewed them in DxO Optics Pro 8 (with all processing presets turned off) and got three identical Raws with identical histograms. HOWEVER the three raws each had a different ISO in the EXIF. This shows I think that the ISO bracketing is being done in camera entirely by the camera raw conversion firmware and nothing else.

I agree with John (Zuiko), you might as well take one raw shot with better quality data and process it with different exposure compensations rather than take "ISO bracket" jpeg's.

All this exercises the brain doesn't it.:)

Regards.*chr

PeterBirder
13th November 2012, 12:18 AM
Hi Peter,

Deja vu! I remember you explaining this to me last Friday! *chr

Hehe. Funny how the same thing pops up in different ways.:)

Nick Temple-Fry
13th November 2012, 12:34 AM
If it only takes one exposure but provides two others by essentially processing in-camera, surely it is no different from processing 3 different exposures from one file in Photoshop? In fact, I would have thought doing the same from one raw would give you more control and theoretically better results. :confused:

Well JDAL is right - ISO bracketing only takes a single exposure. It's easy enought to test, just take a picture with it on and see how many times the shutter clicks. But it does produce 3 Raw files. (and OK I've only checked on an e-pl1). I suspect the EM-5 manual is less well edited than the E-3 was - though both are easy to misread.

Effectively ISO bracketing freezes exposure time but maintains the same judgement as to exposure length, and as the sensor would gather the same info with the same exposure length there is no point in actually opening/closing the shutter 3 times.

EV bracketing varies the exposure time (effectively you are telling auto exposure that the white/gray point has moved by the ev value), so it takes 3 photos with different exposure times.

I'd hope that in ISO bracketing a slightly different raw curve would be applied. If adobe et al can't see the difference, it's because their software is not set up to read whatever flag Olympus sets for iso bracketing in the exif.

But yes for ISO bracketing there seems to be little advantage if you shoot Raw.

(I see Peter has replied whilst my beer addled brain was still trying to compose words into sentences)

Nick

Zuiko
13th November 2012, 01:00 AM
Well JDAL is right - ISO bracketing only takes a single exposure. It's easy enought to test, just take a picture with it on and see how many times the shutter clicks. But it does produce 3 Raw files. (and OK I've only checked on an e-pl1). I suspect the EM-5 manual is less well edited than the E-3 was - though both are easy to misread.

Effectively ISO bracketing freezes exposure time but maintains the same judgement as to exposure length, and as the sensor would gather the same info with the same exposure length there is no point in actually opening/closing the shutter 3 times.

EV bracketing varies the exposure time (effectively you are telling auto exposure that the white/gray point has moved by the ev value), so it takes 3 photos with different exposure times.

I'd hope that in ISO bracketing a slightly different raw curve would be applied. If adobe et al can't see the difference, it's because their software is not set up to read whatever flag Olympus sets for iso bracketing in the exif.

But yes for ISO bracketing there seems to be little advantage if you shoot Raw.
(I see Peter has replied whilst my beer addled brain was still trying to compose words into sentences)

Nick

What I would find more useful is to be able to fix shutter speed and aperture by selecting manual exposure mode, but be able to bracket exposures by shifting the ISO value.

David Morison
13th November 2012, 07:56 AM
I started this thread because I suddenly noticed some strange results for the first time after switching to a Mac. I think this is because Windows does not recognise the E-M5 ORF so I used Olympus Viewer 2 to sort images before dragging and dropping into the pp programme. Naturally Viewer 2 is set up to continue the illusion that I had three differently exposed Raws and interestingly Photmatix also does that. Mac i-photo doesn't set out to fool me in that way and all three Raws look the same.

I will now do some trials using ISO bracketing on Raw+jpeg to see if I can retrieve the same highlight/shadow detail from the Raw as the three bracketed jpegs.

David

Bikie John
13th November 2012, 07:57 AM
What I would find more useful is to be able to fix shutter speed and aperture by selecting manual exposure mode, but be able to bracket exposures by shifting the ISO value.

I think you've hit the screw on the head there John. I think more goes in with exposure at different ISOs than simply how the camera interprets the raw data - maybe it puts a bias voltage on the sensor or something.

Although referring to your earlier comment

If it only takes one exposure but provides two others by essentially processing in-camera, surely it is no different from processing 3 different exposures from one file in Photoshop? In fact, I would have thought doing the same from one raw would give you more control and theoretically better results.

I can see that the camera could be more clever in optimising its JPEG production because it knows you are pretending to change ISO. If you are doing it in PS or whatever you are doing a general-purpose conversion.

I'm glad this discussion has come up. My conclusion is that for a predominently raw shooter there is no advantage to be had from ISO bracketing.

Ciao ... Other John

Ulfric M Douglas
13th November 2012, 08:03 AM
As explained (multiple times!) ISO-Bracket is only for Jpegs.

However : the amount of highlight recovery from the top of the RAW file is very similar to the Underexposed Jpeg of the three ISO-bracketed results. (For bright skies for example)
Using this method allows very easy use of RAW headroom while completely avoiding doing any RAW processing yourself.

When feeding in the three Jpegs to something like Photomatix there is no aligning to be done since they are identical frames with different exposures and colour ranges, and it is also very quick to choose 2 frames out of 3 for the candidates as its all there infront of you.

Proper EV bracketing can give way more range if the camer's EV-bracketing features are sensible enough.

andym
13th November 2012, 08:35 AM
I'm glad this discussion has come up. My conclusion is that for a predominently raw shooter there is no advantage to be had from ISO bracketing.

Ciao ... Other John

Pheww.I though I was missing something.I won't bother then.

jdal
13th November 2012, 09:12 AM
Ulfric's right, it's designed for Jpeg shooters, where it's very useful. I shoot Jpeg+ RAW and it's still useful if I'm using the Jpegs mostly and keeping RAW as a backup.

Interestingly if you load these images in Lightroom the initial image you get is the ISO altered Jpeg one so you see 3 different images,, then when Lightroom has rendered previews out of the RAWs it switches to the RAW versions so you see 3 identical images.

Ian
13th November 2012, 10:42 AM
Not quite. In the case of ISO bracketing it takes one exposure. It creates 3 jpegs from it and 3 identical raw files. With the other bracketing types 3 exposures are necessary.

I'm loth to say ISO bracketing is all done in software, it may be some fancy readout of the sensor thing, but in good old fashioned terms it only takes one photo and produces 3 different jpegs and 3 identical RAWS (if setting is jpeg+raw).

This is different to changing ISO 3 times, which produces different RAW files.

Interesting topic!

I have feeling that there are two ways of adjusting the 'sensitivity' of the sensor; one is, as discussed here, to let the image processor adjust the RAW data to produce the correct tonal levels for the ISO sensitivity desired and another is to change the electrical gain on the sensor. But I am not sure that the gain adjustment happens at all selected ISO values. It may only happen at a small number of stages and in big increments.

Ian

David Morison
13th November 2012, 01:35 PM
A quick test in our rather dark church on a cloudy day. Using the ZD 7-14mm at 7mm, evaluative metering, MF and the two forms of bracketing.

3 jpegs at -1,0,+1 EV with both ISO bracketing and AE bracketing processed in Photomatix and saved from default image.

ISO bracketing:


http://e-group.uk.net/gallery/data/500/HDR11.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/54386)


AE bracketing:


http://e-group.uk.net/gallery/data/500/HDR2.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/54387)


Can't see any difference myself, so if only 3 images are required ISO bracketing seems to be the way to go, only one shutter actuation required and therefore handholding is possible.

What do folks think?

David

Nick Temple-Fry
13th November 2012, 02:22 PM
A quick test in our rather dark church on a cloudy day. Using the ZD 7-14mm at 7mm, evaluative metering, MF and the two forms of bracketing.

3 jpegs at -1,0,+1 EV with both ISO bracketing and AE bracketing processed in Photomatix and saved from default image.

ISO bracketing:


http://e-group.uk.net/gallery/data/500/HDR11.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/54386)


AE bracketing:


http://e-group.uk.net/gallery/data/500/HDR2.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/54387)


Can't see any difference myself, so if only 3 images are required ISO bracketing seems to be the way to go, only one shutter actuation required and therefore handheld is possible.

What do folks think?

David

There is little if any difference.

But I'd like to emphasise the key difference between iso and ev bracketing.

ISO bracketing (as it only takes a single image) has the same exposure time for all 3 shots, so it can't expose more detail in dark areas or get the details in over exposed areas (except perhaps on borderline overexposure).

EV bracketing varies the exposure for each shot, so you end up getting better exposure for details in dark areas and (provided you shoot over a wide enough range) details in highlights. You are effectively telling the camera to move its idea of good exposure up/down the scale.

Nice church chots.

Nick

Ulfric M Douglas
13th November 2012, 02:50 PM
What do folks think?
David if your example had more bright highlights or crushed shadows then there'd be more to recover from both ends of the upper and lower RAW files : making AE bracketing RAWs the best choice.
This example seems comfy enough, but I see a difference in the two images in the dark rafters top left and right.