I'm a bit of a photography luddite so bear with any dodgy terminology or obvious questions!
For a new project I've been playing around with shooting in RAW on my Panasonic Lumix DMZ-TZ100.
The images appear fine on the camera but when uploaded to my laptop some become significantly darker and have shadows around the corners. See the attached screenshot. The thin grey border on the screen shot is just from how I cropped it and isn't part of the problem, just the darkening and black splodges in the corners.
There doesn't seem to be any pattern that determines whether this happens but when viewed in the windows photo viewer the images initially appear OK and then darken and the dark spots appear. I sent one to someone else who opened it in photoshop and they also had the dark spots on that too. When I convert the images to JPEG this effect disappears and they appear as they did on the camera.
Any ideas what's going on here?
Ta.
Oh and as an aside, name the crag!
The camera may be applying some automatic compensation for lens vignetting when it converts to JPG. What f# was the image shot at, vignetting is generally worse with wide apertures.
What happens if you convert the RAW to jpg on your machine (what software are you using for this?). Lightroom also has a feature to provide for this lens compensation.
Also your white balance is way off on that shot. Cameras often struggle with snowy shots. Thankfully you can adjust in RAW easily.
What is happening there is that the camera is automatically applying certain corrections when processing the RAW image internally, whether that is for preview or because you are shooting in JPEG. When you are shooting RAW, you have to apply such corrections yourself. Whatever RAW processing software you use should correct the exposure and vignette for you at least semi-automatically.
Edit: Hardonicus was quicker at typing! :-P
> The camera may be applying some automatic compensation for lens vignetting when it converts to JPG. What f# was the image shot at, vignetting is generally worse with wide apertures.
5.6 I think.
> What happens if you convert the RAW to jpg on your machine (what software are you using for this?). Lightroom also has a feature to provide for this lens compensation.
If I convert to JPEG the effect disappears and the image appears as it does on the camera. I've used a couple of different online and open source programs to do this whose names escape me right now.
> Also your white balance is way off on that shot. Cameras often struggle with snowy shots. Thankfully you can adjust in RAW easily.
The bit I don't understand is why it looks ok on the camera? Something to do with how the various viewers / editing programs process it?
> What is happening there is that the camera is automatically applying certain corrections when processing the RAW image internally, whether that is for preview or because you are shooting in JPEG. When you are shooting RAW, you have to apply such corrections yourself. Whatever RAW processing software you use should correct the exposure and vignette for you at least semi-automatically.
> Edit: Hardonicus was quicker at typing! :-P
Ah OK, so the camera is 'fixing' it automatically but the processing software doesn't? I'd assumed it was the other way round, that the image was right on the camera and the processing software was doing something odd.
Is there anything I can do when shooting to prevent this?
> The bit I don't understand is why it looks ok on the camera? Something to do with how the various viewers / editing programs process it?
Purely because the camera applies processing for you when previewing. That includes vignette correction, white balance detection, exposure compensation etc. The strength of RAW is that it throws away (almost) all on-camera processing and captures the sensor output directly, allowing you to have a lot more control and throw a lot more processing power at the problem than the camera itself has.
One of the main points is that sensors typically have a lot more dynamic range than can be represented in JPEG, and the camera has to apply whatever curve it feels best to transfer that to the reduced dynamic range of the output. When your exposures are perfect, with a subject matter that is not problematic, that's fine. When you have detail that was lost in shadows or highlights, a lot can be done by you later that the camera would have squished out of existence with its on-board processing.
> Is there anything I can do when shooting to prevent this?
Not really. It isn't actually a problem. If the preview is showing that the details you care about are there, then the RAW will contain those details by definition, so what you do with it can only be better.
> Not really. It isn't actually a problem. If the preview is showing that the details you care about are there, then the RAW will contain those details by definition, so what you do with it can only be better.
I think that's what I wanted to know, thanks for your help, and Hardonicus too.
No problem! Just a tip, when shooting snow, set the exposure compensation to over expose the image by some amount. The way the camera measures exposure it will attempt to figure out where the middle grey level is and put that in the middle of its dynamic range. The issue with snow is that most of the landscape is white, so it underexposes to make white grey, which kills off detail where you actually want it. From memory (as I tend to do it by eye), I set it to overexpose by +1EV.
If your using Lightroom if you scroll down in develop there will be two tick boxes for lens correction, tick these. These will apply adjustments that will get rid of distortion and vignette
it is also possible to get presets or use the auto button, the advantage of raw is as you learn more you be able to go back and rework pictures to bring out the best.
this isn’t about obvious “photo shopping” but using the full range of a modern sensor.
the last 5 or so pictures in my gallery would not of been possible if shot in JPEG, accidentally shooting in JPEG is one of the most frustrating things a photographer can have happen to them 😏 especially if the light is challenging.
You are not wrong, but the reality is hard.
The camera does a lot of internally hard-coded design-specific correction in creating a JPG from a RAW: vignetting, CA, distortion, colour balance, noise reduction etc. In 'theory' you can reproduce that outside the camera, but it can be non-trivial particularly for very lens & focal length dependent aberrations like vignetting and distortion in compact cameras. If you want to do it accurately you really need to measure the distortion and vignetting (at a range of focal lengths) and then use that data to correct the RAW image. Of course you may not care about accuracy in many cases (e.g., landscapes), but you shouldn't assume that you can improve on the in-camera RAW-JPEG conversion unless you really know what matters (to you) and what doesn't and what control you have in RAW development.
And yes, you can get camera and lens profiles to use in (for example) Photoshop, but in my experience they are very 'variable' (aka generally not much use) in quality.
You are right of course, the theory and practice do not always align. In my case, I use DxO Optics, which has lab measurements for my camera body and all my lenses, so it improves on the in-body processing at all times bar none.
> The bit I don't understand is why it looks ok on the camera? Something to do with how the various viewers / editing programs process it?
The camera screen displays a internally corrected JPG image rather than the RAW image, even if you are shooting in RAW-only. It show what the final image 'might' look like if you know how to do the corrections in your external program (and if it's capable of doing them).
> No problem! Just a tip, when shooting snow, set the exposure compensation to over expose the image by some amount. ... I set it to overexpose by +1EV.
A lot of the time the shot is underexposed enough that there is just too much noise in the shadows, resulting in loss of detail after denoising.
I’m still on a learning curve, but one thing I’ve found useful is to set my camera to save in both RAW and JPEG. It allows me on downloading to see the in camera interpretation of the raw file as well as having all the available detail of the Raw file to alter if I wish later. Makes it easier for me to filter out duds at an early stage as well as giving one example of what is possible with the Raw data.
If your camera can save both, might be worth a try for a while. Of course you will use up memory card space, and longer term storage, quicker if you retain all the JPEG copies as well as Raw.
Thanfully, SD card capacity is absurd these days. The fact that you can get an 1TB microSD is mindblowing.
No idea about photography but that's Creag a' Mhadaidh. Looking very icy too.......ah well.
Yes; beyond my understanding it’s possible! I still remember floppy discs and having a PC with a 33 Mb HDD!!
However, one should not forget they could fail. I’ve had three SD cards fail in recent years, so much prefer numerous smaller capacity cards. Ideally, if organised, to change daily to minimise losses if card error occurs.
I had a look at your gallery after seeing this and wow.
Some fantastic photographs in there... That rope access one on the oil rig (I think?). Stunning.
I think it probably depends on the camera/lens. The profiles are reasonably good for (for example) my Canon 'L' lenses (on 6D), but rubbish for the more consumer oriented Samyang/Sigma ones (last I looked).
> I had a look at your gallery after seeing this and wow.
> Some fantastic photographs in there... That rope access one on the oil rig (I think?). Stunning.
We need to be careful. I said the same to him a few weeks ago. We don’t want him going all Diva on us!
I suspect that on lower quality lenses it isn't the profile itself that is necessarily bad, rather lens build quality being variable enough that each unit actually has different optical distortions. That would certainly explain why L-series lenses would have great profiles.
Bit of both, I think. Creating a good profile is a LOT of work, particularly for zoom lenses. You have accurately characterise them at all focal lengths, all apertures, all focus distances. And then you have to accommodate all the sensors that it might be used with. That's a lot of data to develop and the techniques to do this are also non-trivial (especially for wide-angle lenses). Consumer lenses also are more likely to need correction - certainly there are plenty in the compact camera range which have been designed on the principle that you don't correct distortion in the optics (expensive) when you can do it in firmware (cheap).
But you're right - cheaper lenses also have much more sample-to-sample variation, so the best profile is still probably of little use. Which is why if you're fussed about accurate correction (e.g., astrophotography mosaics) there's really no alternative to calibrating the lens YOU have, on the body YOU have and at the settings YOU use.
😂😂have you been speaking to my wife?
> Yes; beyond my understanding it’s possible! I still remember floppy discs and having a PC with a 33 Mb HDD!!
33MB! You were lucky then. When I started we made do with 32KB!
Martin
> No idea about photography but that's Creag a' Mhadaidh. Looking very icy too.......ah well.
We have a winner! If it's any consolation I didn't climb on it, just skied past to take some photos.
> I’m still on a learning curve, but one thing I’ve found useful is to set my camera to save in both RAW and JPEG.
This is what I arrived at after my first day out with the new camera. Cheers.