PDA

View Full Version : Help!!! Strange purple haze patches, no CA...


Ro881e
7th April 2008, 11:31 PM
Hi all,

Im new to the forum, and just though someone may be able to help me out. Iv been experimenting with some night photography and have discovered a strange purple flare on the left edge of the frame in most of the pictures. Iv been using more than one lens (18-55 kit canon, and 50mm 1.8) with the same problem. Is this my camera?? I own a 350D, about 6 months old, so should be in top condition still.

Any ideas??

Thanks.

Rob.

PHIL THE FRAME
7th April 2008, 11:50 PM
im no expert but on the evidence you give i would say its not a lens issue its maybe a sensor problem,if possible try using the same lens on a different 350,i know that might not be possible,you could also print out the photo you have uploaded here and try you local photo shop also if the camera is only 6 mts old try the warranty if there is one hope this helps,but as i said im no expert, hope you get it sorted

Pabloverdes
8th April 2008, 12:20 AM
Hi Robbie

I think the previous answer is probably the way to go. If the purple haze is occurring when using both lenses then it's more than likely the camera. If its still under warranty get it fixed professionally.

Paul


(Is this the Hendrix J syndrome?)

Dotty.c
8th April 2008, 07:29 AM
does it appear on daylight shots........are you using the timer and slow shutter speed, it looks abit like light spill of some kind.

Ro881e
8th April 2008, 09:25 AM
Thanks for the info guys, the mystery is looking more and more like a camera problem i guess, confirmed my suspicions. Im gonna have a dig around for the warranty i think, not sure where i left it! :wacko:

Although, Dotty.c, i have only noticed it recently on long (30+ secs) exposure shots with timer, and have never noticed it on daylight shots?? Could this mean something else? When you say light spill, is that when light leaks through the open viewfinder? Because to be fair, i didnt cover it, but i presumed it would be ok as it was pitch dark.

PS. I just got the Hendrix J syndrome pun, Haha! Im a bit slow today...

Robbie

.Ray
8th April 2008, 09:57 AM
Maybe you can put the cam on manual, leave the lens off totally and cap the body then do a 30s exposure to see if the haze appears. Light from the viewfinder shouldn't get to the sensor during exposure as the mirror is in the way.

Dotty.c
8th April 2008, 10:43 AM
The fact that your using 30sec plus in an uncontroled enviroment would suggest an ambient light source creeping into your pic. do as Ray suggests and also try it in controlled conditions, where there is no chance of any extra light, even the internal lights in a car would spill over, what where the other settings you used.

BarryM
8th April 2008, 10:50 AM
try putting ur eye piece over the viewfinder on your long exposed shots, it should be attached to the strap.

Ro881e
8th April 2008, 06:11 PM
Wow! Thanks all, lot of responses here. Erm, ok...

I see what you mean Dotty.c about the possible ambient light source, although i may have failed to mention that it has appeared in exactly the same place on a wide range of different pictures of different subjects. So i might be able to rule this out (oh dear, means the problem is more serious!).

Ray, i think i know what you mean by capping the camera, taking a long exposure of "total" dark (ie covered) right? Ill give it a try later and report back the results.

I suppose at this stage it is a process of elimination, viewfider cap will also be tried.

Appreciate the responses guys, hopefully i can fugure this one out without having to sell my 350!

Rob.

PS. Iv attached another example from different photo (purple haze is at bottom here as camera was tilted), dont know if this helps, but i thought i would anyway.

karl-tkd
8th April 2008, 08:41 PM
try putting ur eye piece over the viewfinder on your long exposed shots, it should be attached to the strap.

I was going to suggest that Baz.....a few months ago i couldn't understand why some of my moving water captures on long exposures were comming out dark, i later realised that light going through the view finder was under exposing the shots, i now cover the view finder with my hand and it has made a difference.

neutron2k
17th April 2008, 08:57 AM
ok this is NOT a camera problem.

What you are seeing is noise from the cameras sensor as you are shooting a long exposure. I had this problem with my D200 and it took me a while to figure out what it is.

All you need to do is turn on the 'Long Exposure Noise Reduction' feature of your camera. That will get rid of it - but it will double your exposure time.

Basically the radiation from the sensor intereres with the exposure causing the pink noise. When you use long exposure noise reduction, a second shot is made of total black (the camera takes a photo with the shutter back down). This second shot is then subtracted from the exposure you made to remove the pink pixels and any resulting noise :)

hope this helps! Very few people seem to know about this.

oh yes light can also get in via the viewfinder but its not that causing this issue:)

george.monaghan
17th April 2008, 09:29 AM
Hi,

Agree above - it is colour shift due to long exposure and it is also "dust",

Try cleaning the sensor as see if that does make a difference.

I might be wrong but that is what it looks like. It is appearing in exactly the same place on the sensor.

neutron2k
17th April 2008, 10:59 AM
think you've missed my point there - its not dust and its not colour shift :) its radiation noise from the sensor itself. The only way to remove it is long exposure noise reduction :)

.Ray
17th April 2008, 11:09 AM
Oh dear radiation noise?! Sounds like a fault on the sensor then.

neutron2k
17th April 2008, 01:36 PM
its not a fault! Its the way they work!

read this for an example :)

http://www.kenrockwell.com/nikon/d200/d200-dark.htm

canonman
17th April 2008, 02:02 PM
its not a fault! Its the way they work!

read this for an example :)

http://www.kenrockwell.com/nikon/d200/d200-dark.htm

Forgive my contradicting you, but I suspect you may have a monitor calibration issue.

The image posted with regard to this issue, has a problem with a very small and randomly concentrated portion of the frame. The diagnosis you have arrived at, refers to the overall pink cast that is in evidence. This however is not the issue in question. Sensor noise does not account for this particular problem !

I fully appreciate that your mis-judgement of this of this is purely a monitor issue, and not one of incompetence or lack of technical understanding.

neutron2k
17th April 2008, 02:17 PM
i am not talking about a pink colour cast! I'm talking about bright pink splodges in the corners and around the edges of an image caused by long exposures!

my monitors (all three of them) are calibrated :)

look at that link i sent and scroll downthe page. you will see EXACTLY what I am talking about and EXACTLY what appears in the posted photo :) I.e. an illuminous ping splodge on the bottom of the frame.

george.monaghan
17th April 2008, 03:26 PM
think you've missed my point there - its not dust and its not colour shift :) its radiation noise from the sensor itself. The only way to remove it is long exposure noise reduction :)


Hi,

Long exposure noise reduction is the same exposure taken without exposing the sensor. The camera then subtracts the stuck pixels from the first image. An exposure with the lens cap on and subtracted from the image will give the same results. (dark frame subtraction)

The image that has been posted does not look like a simple stuck pixels problem.

Thus: http://en.wikipedia.org/wiki/Audio_noise_reduction

I am not an expert but the images posted show dust marking.

Turn off high ISO noise reduction as standard. Takes too long to process in camera.

canonman
17th April 2008, 03:53 PM
i am not talking about a pink colour cast! I'm talking about bright pink splodges in the corners and around the edges of an image caused by long exposures!

my monitors (all three of them) are calibrated :)

look at that link i sent and scroll downthe page. you will see EXACTLY what I am talking about and EXACTLY what appears in the posted photo :) I.e. an illuminous ping splodge on the bottom of the frame.

I apologise for the mis-understandings that seem to be creeping into this discussion, but this is why we need to try and be very specific with our communication skills! You previously referred to "pink noise" (which is evident over all of the image), which didn't accurately describe the specific concentration of the "pink splodges" which you now choose to make reference to.

There still however is a flaw in your diagnostics, which you partly mention above when quoting "in the corners and around the edges".

The generator of the type of noise you suggest being the culprit is often called "dark current" which is inherrent during long-exposures. This type of noise (and there are many others) is a symptom of heat generated by the microchip (more a problem with CMOS than CCD sensors, and the smaller a sensor the worse the problem). The "hotsot" of such will be in the corners (where the heat conduction is amplified at the point of intersection of two edges). From the "corner hotspot", heat will then radiate outwards along the edges.

It is pretty un-likely therefore that the image in question here is suffering from that, as there is no sign of excessive noise in the corners of the frame. The "pink splodge" in this case is generated central to a frame edge.

Another member of the forum mentioned "dust", which you dis-missed out of hand. Any disruption to the uniformity of the light field, will manifest itself in "noise", and as such I think his point (in the absence of a corner hotspot) is a plausible contributor to the problem. Take a good look at reference images posted on line with regard to "long exposure noise", you will see that the initial point of heat generation is in the corners.

I do hope we don't have to enter into any further discussion with regard to Poisson statistics from the random rate of arrival of Photons etc etc, I for one find it all totally boring as I'm sure everyone else in here would agree!?

Oh!! And it isn't curteous to use block capitals in a discussion!

neutron2k
17th April 2008, 09:12 PM
I apologise for the mis-understandings that seem to be creeping into this discussion, but this is why we need to try and be very specific with our communication skills! You previously referred to "pink noise" (which is evident over all of the image), which didn't accurately describe the specific concentration of the "pink splodges" which you now choose to make reference to.

There still however is a flaw in your diagnostics, which you partly mention above when quoting "in the corners and around the edges".

The generator of the type of noise you suggest being the culprit is often called "dark current" which is inherrent during long-exposures. This type of noise (and there are many others) is a symptom of heat generated by the microchip (more a problem with CMOS than CCD sensors, and the smaller a sensor the worse the problem). The "hotsot" of such will be in the corners (where the heat conduction is amplified at the point of intersection of two edges). From the "corner hotspot", heat will then radiate outwards along the edges.

It is pretty un-likely therefore that the image in question here is suffering from that, as there is no sign of excessive noise in the corners of the frame. The "pink splodge" in this case is generated central to a frame edge.

Another member of the forum mentioned "dust", which you dis-missed out of hand. Any disruption to the uniformity of the light field, will manifest itself in "noise", and as such I think his point (in the absence of a corner hotspot) is a plausible contributor to the problem. Take a good look at reference images posted on line with regard to "long exposure noise", you will see that the initial point of heat generation is in the corners.

I do hope we don't have to enter into any further discussion with regard to Poisson statistics from the random rate of arrival of Photons etc etc, I for one find it all totally boring as I'm sure everyone else in here would agree!?

Oh!! And it isn't curteous to use block capitals in a discussion!

apologies for using block captital - I was trying to get the point across that all he needs to do is enable long noise exposure to resolve the issue with the pink splodges and explain why in laymans terms - somthing i'm not great at, but for some reason, like my hello thread, its totally spiralled out of control and I have been taken in completely the wrong way and I just can't be bothered with the hassle.

I was trying to help someone out who had a problem that I have expeirenced as well as several other people and I can't be bothered to sit here and get slated for that.

canonman
18th April 2008, 07:50 AM
apologies for using block captital - I was trying to get the point across that all he needs to do is enable long noise exposure to resolve the issue with the pink splodges and explain why in laymans terms - somthing i'm not great at, but for some reason, like my hello thread, its totally spiralled out of control and I have been taken in completely the wrong way and I just can't be bothered with the hassle.

I was trying to help someone out who had a problem that I have expeirenced as well as several other people and I can't be bothered to sit here and get slated for that.

Apology accepted, many thanks for that.

Why not start afresh with a more open-minded attitude and in a bit more of a light-hearted vain!? I think you have a lot to offer people in here, but we all need to have a little diplomacy when chatting and putting across our opinions. Photography is no different to anything other subject, and especially where science is involved, very few things are absolutely clear-cut (although the scientist likes to think it is). So with that in mind I am always very careful to phrase my wording making it absolutely clear that it is an opinion only (even if I am sure that my opinion is the correct one).

Anyway Neutron2k, welcome again and here's to some valuable contributions from you here in the future !