PDA

View Full Version : EM1 Fireworks


birdboy
2nd November 2013, 10:31 PM
I just got the EM1 in time to try some fireworks pics tonight. Having not had livetime shooting before I thought I would give this a try.

The following pics were taken with 12-40 pro manual focus f7.1. Good use of this exposure mode as you can watch the picture develop. I did use low ISO.

http://e-group.uk.net/gallery/data/500/PB020060.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67228)

All appeared well until you pixel peep. Urgh black spots:eek: They seem to only appear where there is little light. Could they be lazy photosites that need a good bashing with light to get to work.:mad: The following is a 100% crop showing the random black spots.

http://e-group.uk.net/gallery/data/500/PB020060-1.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67227)

I know Carlos1 has experienced something similar but has anyone else seen this or better still explain it?
Thanks John

Phill D
3rd November 2013, 07:40 AM
John I think that's a lovely fireworks shot without the pixel peeping. Not sure what is going on in the crop though, isn't it just the way the noise shows up in the dark areas at that level of crop. Surely at some point the pixels will have to be black? Wouldn't any camera shot like this look similar at such a close inspection? I'm sure someone more knowledgeable will be able to explain.

pault
3rd November 2013, 08:44 AM
Hi,could black spots be debris from the fireworks powder explosion?

OlyPaul
3rd November 2013, 09:42 AM
Hi,could black spots be debris from the fireworks powder explosion?

I'd tend to agree with this as they appear the wrong shape for pixels, peeping at 2000% to see if they are pixel shaped would confirm if this is the case or not.

jmunkki
3rd November 2013, 09:48 AM
Overlay two frames from the same shoot and check if the spots stay in place. Sensor noise is stationary, but firework debris isn't.

birdboy
3rd November 2013, 10:34 AM
Hi,could black spots be debris from the fireworks powder explosion?

Thanks for suggestion but its not firework debris as I can reproduce this away from a firework display:(

It happens with long exposures I shall take some raw pictures tonight to see if its in the jpg software.

birdboy
3rd November 2013, 10:38 AM
Overlay two frames from the same shoot and check if the spots stay in place. Sensor noise is stationary, but firework debris isn't.

I took two frames at different ISO last night and the spots appear to be in the same place. It may be that this has something to do with the incamera noise reduction so I will try again tonight with that turned off.

birdboy
3rd November 2013, 10:45 AM
I'd tend to agree with this as they appear the wrong shape for pixels, peeping at 2000% to see if they are pixel shaped would confirm if this is the case or not.

Zooming in to 11:1 they are clearly square and pixel shaped and appear to me to be something of an artefact you get associated with noise.

The good thing about this is that it is reproducible at the moment. So hope to get to the bottom of it.

pault
3rd November 2013, 12:38 PM
Hi again,
It would interesting if Olympus could advise. Just a thought, wonder if these are pixels that are "mapped out" because they were "hot"

jmunkki
3rd November 2013, 03:50 PM
Simplifying a bit, there are two ways you can map out hot pixels: by interpolating from surrounding pixels or by subtracting the value you got when the shutter was closed. If I was responsible for writing the noise reduction software, I would try to do a combinatin of both. If you just subtract, you can sometimes end up with pixels that are darker than they should be. If you just interpolate, you lose detail that might otherwise be there.

If the black spots stay in place, then you are seeing hot pixels where the noise reduction has been hard at work. It's possible that a firmware update from Olympus could improve the processing to make these less noticeable.

Remember back in the days of film though... Film had grain and it was understood to be part of photography. Digital sensors have hot pixels and noise. It's not really that different.

If the E-M1 has a pixel mapping function, you could try running that. My E-520 has it and it's a good idea to run it every few months. They should have run it at the factory at Olympus, but maybe they forgot?

birdboy
3rd November 2013, 04:06 PM
I've done a bit more testing (the darkness that lurks under my desk:eek:) using raw and noise reduction turned off. The black pixels are clearly visible but it seems only in darkish areas nearby pixels seem to have significantly higher luminance. The black squares are slightly less visible with the raw file and I can only assume that the NR in camera software makes things worst because it darkens adjacent pixels hence it appears as visible black dots at 100%. I have carried out 3 pixel mappings in total and this seems to make no difference.

The only thing I can think of is that I am seeing the PDAF pixels but there does not seem to be regular pattern and my understanding from the promo details is that these had been masked out.

Its time for an email now to Oly technical support.

jmunkki
3rd November 2013, 04:12 PM
It's just a limitation of sensor technology. If you want cleaner images, you probably have to consider a full frame sensor or wait a few years while technology improves. On even smaller sensor cameras, you can see artifacts like that even in normal lighting conditions. In general, it's not a good idea to expect 1:1 crops from consumer digital cameras to look perfect.

birdboy
3rd November 2013, 04:25 PM
It's just a limitation of sensor technology. If you want cleaner images, you probably have to consider a full frame sensor or wait a few years while technology improves. On even smaller sensor cameras, you can see artifacts like that even in normal lighting conditions. In general, it's not a good idea to expect 1:1 crops from consumer digital cameras to look perfect.

With respect I have the E3 and E5 cameras and have taken many long exposure pictures. I have gone back and inspected these and none of them show any signs of the black dots I see now or of limitation you think are present. And this is not a consumer digital camera but a pro spec, the top of range for Olympus.

Peter_Hartland
3rd November 2013, 10:34 PM
Great firework pics again showing how nice the EM1 is.

bredman
3rd November 2013, 10:54 PM
They are not pdaf pixels.

http://www.43rumors.com/curious-e-m1-sensor-picture-and-first-olympus-12-40mm-review/


Also, this may be useful

http://www.mu-43.com/showthread.php?t=53621&highlight=hot+pixels

birdboy
3rd November 2013, 11:52 PM
They are not pdaf pixels.

http://www.43rumors.com/curious-e-m1-sensor-picture-and-first-olympus-12-40mm-review/


Also, this may be useful

http://www.mu-43.com/showthread.php?t=53621&highlight=hot+pixels

Thanks for the links Pete had not seen the second link, it seems there is an issue:(.

Yes I agree they are not PDAF pixel because they extend outside the range of those pixels. On my pics there are a lot and I mean a lot.

It seems that this is a fault/feature of the EM1 that is present on some EM1's :eek: It might not have been seen by most current EM1 users because they are not using long exposures. But once you know its there I can see in on normal exposed pictures where there is darkness. The "dead pixels" if that's what they are easily spotted pardon the pun.

bredman
4th November 2013, 01:57 AM
Thanks for the links Pete had not seen the second link, it seems there is an issue:(.

Maybe not. That hot pixel thread i linked to was just to show the M5 and M1 process RAW differently, and to show what the M1 hot pixels will look like. I don't think your issue is hot pixels. Looks like the consensus over there is that the EM5 cooks the RAWs and that hot pixels are to be expected (that thread may be updated yet) on the M1 -- see posts #22 onwards.

Perhaps those black spots are related to the issue Carlos posted about. His seemed to just be short lived though, and there is a disparity on length of exposure.

birdboy
4th November 2013, 06:24 PM
Maybe not. That hot pixel thread i linked to was just to show the M5 and M1 process RAW differently, and to show what the M1 hot pixels will look like. I don't think your issue is hot pixels. Looks like the consensus over there is that the EM5 cooks the RAWs and that hot pixels are to be expected (that thread may be updated yet) on the M1 -- see posts #22 onwards.

Perhaps those black spots are related to the issue Carlos posted about. His seemed to just be short lived though, and there is a disparity on length of exposure.

So far no response from Oly customer services:(

I think they are dead pixels but there are an awful lot of them I estimate in the 1,000s and they cover the whole sensor.:eek:

I have tried with different lens and get the same black dots appear.

Despite a camera reset and more pixel mappings these black dots remain. Either the pixel mapping is not working or I am doing it wrong.

Does anyone know if there is a correct way of pixel mapping. i.e with the lens cap on/off or in day light?

Here are two more examples.
From Saturday night

http://e-group.uk.net/gallery/data/500/Dead_pixels_sky.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67270)

And taken a few moments ago.


http://e-group.uk.net/gallery/data/500/Dead_pixels.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67269)

What surprises me is that given Oly sell this camera on the claim that it has the best image quality why has nobody else picked this up. Either I have a dud camera or am a dud photographer if so what am I doing wrong.

Showing images at 100% is what I do with my pictures having burnt them to BluRay discs often cropping to 1900 pixels wide to fit our wide screen tv. So these black spots matter to me and I would have thought should matter to other EM1 users The dots are there to a lesser extent with faster shutter exposures you just need to look into the shadows.

brian1208
4th November 2013, 07:02 PM
those look like typical JPG artefacts to me, rather than pixels,

What does it look like if you process it from the RAW image?

If they are pixels you should expect to see them in the same place, if they aren't there then they are artefacts

birdboy
4th November 2013, 07:10 PM
those look like typical JPG artefacts to me, rather than pixels,

What does it look like if you process it from the RAW image?

If they are pixels you should expect to see them in the same place, if they aren't there then they are artefacts

The raw pic are the same the black dots are in the same place. I am not taking that many raw pics because I manly use LR4 Viewer 3 shows the same results.

If they are not dead pixels (I feel a Monty Python sketch coming up here) then they must be lazy pixels some that need a higher threshold of light to be give an output.

Did you take any long exposure circa >7 secs pics with your EM1 if so have a close look at any shadow area for black dots.

brian1208
4th November 2013, 07:27 PM
haven't done this with my EM-1 (still waiting for its return) but I have with my EM-5 which is where my thoughts on artefacts came from.

I'll plough through my files to see if I can find an example of what I'm talking about but the difference I see between the .jpg file and ORF files there are that the 'jpg noise looks blocky and the orf is more typically "grainy"

andym
4th November 2013, 07:31 PM
Though I'd have a play.This is my E-M1 with the 12-60 at 12mm F2.8 and ISO1600,out of my front bedroom window just now.

http://www.e-group.uk.net/gallery/data/500/PB040821.jpg (http://www.e-group.uk.net/gallery/showphoto.php?photo=67271)

http://www.e-group.uk.net/gallery/data/500/PB040821crop1.jpg (http://www.e-group.uk.net/gallery/showphoto.php?photo=67272)

http://www.e-group.uk.net/gallery/data/500/PB040821crop2.jpg (http://www.e-group.uk.net/gallery/showphoto.php?photo=67273)

This is with no noise filter and noise reduction set to auto so as the exposure is 5 sec it's probably done dark frame subtraction.Taken in Raw and converted to jpeg in Silkypix.
One crop from the centre of the frame and on from the far LHS.
As you can see just a bit of chrominence noise.

Interested to hear other opinions but looks like you may have a sensor problem.

Dred to think what my neighbors though.

andym
4th November 2013, 07:36 PM
One other thought, have you tried another Raw developer apart from Lightroom?

brian1208
4th November 2013, 08:04 PM
http://farm6.staticflickr.com/5494/10676431074_4f3f9fd4a0_o.jpg

ISO 1600 jpg crop




http://farm3.staticflickr.com/2865/10676431014_9581c74900_o.jpg

ISO 1600 orf (converted to .jpg crop)

The differences aren't as stark here (I didn't want to enlarge the images) so these are 100% crops taken from more or less the same spot on the same shot, the first as a .jpg file and the 2nd as a .orf file

I think its still possible to see that the .jpg has a black, block component to the noise whereas the .orf crop has a coloured grain

no noise reduction with either of these EM-5 files

(having said all that and seeing the results from Andy's example, I'm not quite sure what the heck is going on with yours :( )

birdboy
4th November 2013, 08:17 PM
One other thought, have you tried another Raw developer apart from Lightroom?

Thanks for doing this Andy at least it looks like it is not a common issue.

I have LR 4 which does not recognise the EM1 raw files so the only way of looking at raw files is with Viewer 3 at the moment.

I am convinced it has nothing to do with the raw image processing in camera or post camera. I have taken a series of pics in raw, noise filtering on (low standard and high) noise filtering off. The black dots are present in all the pictures.

birdboy
4th November 2013, 08:22 PM
Though I'd have a play.This is my E-M1 with the 12-60 at 12mm F2.8 and ISO1600,out of my front bedroom window just now.

This is with no noise filter and noise reduction set to auto so as the exposure is 5 sec it's probably done dark frame subtraction.Taken in Raw and converted to jpeg in Silkypix.
One crop from the centre of the frame and on from the far LHS.
As you can see just a bit of chrominence noise.

Interested to hear other opinions but looks like you may have a sensor problem.

Dred to think what my neighbors though.

The main difference between yours and mine is that of shutter time I see yours are 5secs whereas I think the problem seem to really show above 7secs. I know I have a bit of cheek but could you close down the aperture to say f8 and get the time about 10secs. Thyanks John

andym
4th November 2013, 08:34 PM
Probably really in trouble with the neighbors now and with a cold.

Here you go, I did a bit better with my focus this this also.Twenty second exposure.

http://www.e-group.uk.net/gallery/data/500/PB040822resize.jpg (http://www.e-group.uk.net/gallery/showphoto.php?photo=67326)

http://www.e-group.uk.net/gallery/data/500/PB040822crop3.jpg (http://www.e-group.uk.net/gallery/showphoto.php?photo=67324)

http://www.e-group.uk.net/gallery/data/500/PB040822crop4.jpg (http://www.e-group.uk.net/gallery/showphoto.php?photo=67325)

birdboy
4th November 2013, 08:39 PM
Probably really in trouble with the neighbors now and with a cold.

Here you go, I did a bit better with my focus this this also.Twenty second exposure.



Thanks Andy that is very helpful for me.

John

birdboy
4th November 2013, 08:50 PM
Just to show that these are not artifacts here are two pics both using Olympus Viewer and copied into Paintshop Pro with the joined up dots copied from one to another.

I think this shows clearly that there is something wrong with my sensors pixels.:(

The first one was taken raw12-40pro 11.5 secs 21mm ISO100
http://e-group.uk.net/gallery/data/500/Image110.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67337)


The second taken LSF 12-40pro 15.1 secs 17mm ISO100
http://e-group.uk.net/gallery/data/500/Image210.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67338)

What is interesting is that some of the dots do not show in the first image. This leads me to think that they are lazy pixels and not quite dead. Anyone got a better idea?

PeterBirder
4th November 2013, 09:22 PM
Just to show that these are not artifacts here are two pics both using Olympus Viewer and copied into Paintshop Pro with the joined up dots copied from one to another.

I think this shows clearly that there is something wrong with my sensors pixels.:(

The first one was taken raw12-40pro 11.5 secs 21mm ISO100
http://e-group.uk.net/gallery/data/500/Image110.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67337)


The second taken LSF 12-40pro 15.1 secs 17mm ISO100
http://e-group.uk.net/gallery/data/500/Image210.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67338)

What is interesting is that some of the dots do not show in the first image. This leads me to think that they are lazy pixels and not quite dead. Anyone got a better idea?

This could be "Hot Pixels" which you get with long exposures which cause the sensor to heat up.
In an earlier post you referred to using different "noise filter" settings but the Noise Filter is intended to deal mainly with high ISO noise and only affects JPEGS.
The facility for dealing with hot pixels in long exposures is "Noise Reduction" which is different and takes a second exposure of the same duration but with the shutter closed ( a "Dark Frame" ) and subtracts it from the actual exposure to cancel out the affected pixels. This does affect the raw file and needs to be set to "Auto" so that it only operates when you make long exposures.

I may have missed any reference you may have made to Noise Reduction in earlier posts, if so please ignore my comment, just trying to help.

Regards.*chr

birdboy
4th November 2013, 09:34 PM
This could be "Hot Pixels" which you get with long exposures which cause the sensor to heat up.
In an earlier post you referred to using different "noise filter" settings but the Noise Filter is intended to deal mainly with high ISO noise and only affects JPEGS.
The facility for dealing with hot pixels in long exposures is "Noise Reduction" which is different and takes a second exposure of the same duration but with the shutter closed ( a "Dark Frame" ) and subtracts it from the actual exposure to cancel out the affected pixels. This does affect the raw file and needs to be set to "Auto" so that it only operates when you make long exposures.

I may have missed any reference you may have made to Noise Reduction in earlier posts, if so please ignore my comment, just trying to help.

Regards.*chr

Thanks Peter I think I understand. I did try and see if the noise reduction off on and Auto made any difference, but once again I got the black dots irrespective of the settings.

I have seen this article from Olympus which leads me to think they are dead or nearly dead pixels
http://www.olympusamerica.com/crm/oneoffpages/ask_oly/crm_e_ask_oly_03_09.asp

brian1208
4th November 2013, 10:12 PM
I was thinking about this and suddenly remembered there being a problem with one of the major makers (Canon?) where they were seeing lots of "black spots" on images. After much debate about dust spots etc the final resolution was a firmware upgrade.

I can't for the life of me remember more that that (it was about 3 years back I think) can anyone else fill in the missing details as it may relate to what birdboy is seeing.

(I'll keep on digging and see if I can kick start my creaky old memory)

birdboy
5th November 2013, 11:08 AM
I think I have narrowed this down and it appears to do with using Livetime rather than dialling in the shutter speed.

This one was taken using Livetime and stopping about 8secs in Black dots in the shadows

http://e-group.uk.net/gallery/data/500/Black_dots_Livetime.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67352)

This one taken by dialling in 8secs No black dots in the shadows.

http://e-group.uk.net/gallery/data/500/Black_dots_Manual_time.jpg (http://e-group.uk.net/gallery/showphoto.php/photo/67353)

Now all I need is Oly technical support to respond to my email of Sunday*zzz*zzz

brian1208
5th November 2013, 11:33 AM
A most odd effect, it will be interesting to hear the outcome.

Glad it doesn't seem to be a "Rogue Pixel" problem though