 |
|
|
Image post-processing: |
| Author | Message |
|
GemmaJF Admin Group Joined: 25 Jan 2003 No. of posts: 2090  View other posts by GemmaJF |
Posted: 18 Aug 2005 Hi all, Firstly I thought it was time we had a dedicated forum for photography as it appears most of us share an interest in the subject as well as native herps. I thought I would start it off with a question about image post-processing. I've begun to notice that images produced on My EOS 350D are rather soft. I've been saving them to the CF card as JPEG. Now it is says in my little booklet that the camera set to its default parameter of 1, will produce sharp piccies. It doesn't, well not sharp enough for close-ups. I've begun to realise having read a few articles that the more serious end of digital photography requires digital post processing. (something that happens in the camera with point and click digital cameras) Having hit the 'sharpen' button in photoshop on some of my JPEG files I've been pleasantly suprised to see my less then perfect piccies come to life. So to the point, would it be worth me saving in RAW format to give myself as many options as possible (I know it will reduce the capacity of the CF card) or do most peeps save as JPEG and find this good enough with a little post processing to get the results they want? Gemma Fairchild, Independent Ecological Consultant |
|
B Lewis Krag Committee Joined: 24 Aug 2004 No. of posts: 146  View other posts by B Lewis |
Posted: 18 Aug 2005 Hi Gemma, I posted a few adder pics on the thread yesterday eve. I currently save them as JPEG on the CF card. Have a look and tell me what you think. I haven't tried to sharpen them at all but am happy with the general result. I'd also like to hear peeps views on Gemma's question as I'd like to continue my improvement of photography. Need to find my instruction book as I don't have a RAW option in the menu on my 300D, has anyone else found it? Kind regards Brett. Lewis Ecology Brett Lewis Photography Kent Reptile & Amphibian Group DICE - University of Kent |
|
Vicar Senior Member Joined: 02 Sep 2004 No. of posts: 1181  View other posts by Vicar |
Posted: 18 Aug 2005 Yes yes yes...save as RAW ! RAW effectively records what the sensor sees. The JPEG conversion algorithm on the camera is optimised for speed, so you can take another pic without delay. The RAW -> JPEG conversion on your PC will be much slower, but also far better quality, especially re colour accuracy. Using RAW with post-processing is ideal, as you have not already thrown away some image data. Yes you get fewer pics per card, but quality is better. I switch to jpeg if I'm doing holiday snaps, but always revert to RAW when doing nature snaps. Using RAW images does require some discipline, else even the multi-gigabyte hard disks which are usual nowadays fill up, so view as RAW (using the excellent and free Microsoft RAW image viewer) and throw away unuseful pics before conversion. Steve Langham - Chairman 
Surrey Amphibian & Reptile Group (SARG). |
|
GemmaJF Admin Group Joined: 25 Jan 2003 No. of posts: 2090 View other posts by GemmaJF |
Posted: 18 Aug 2005 Hi Brett, if I had taken the shots you posted up yesterday I would have been very pleased with the way they had come out for sure. RAW option is under Quality on the 350D, perhaps the 300D is the same? I'll give an example though of the difference between one of my shots that I was quite pleased with, compared to after a bit of post image processing of the JPEG file as an illustration. What do people think, and do most apply this sort of processing to their digital images? From what I've gathered so far the softness of the first image is inherent to a digital SLR.
Steve, that sounds like a good option! Gemma Fairchild, Independent Ecological Consultant |
|
B Lewis Krag Committee Joined: 24 Aug 2004 No. of posts: 146 View other posts by B Lewis |
Posted: 18 Aug 2005 Wow, that really is quite a difference... I'll have to have a look through some of mine and check out that option. Do you use the default 'sharpen' or repeat until necessary?? Using Elements 2.0 which came with camera... But have Photoshop 7, any difference in its use. Brett Lewis Ecology Brett Lewis Photography Kent Reptile & Amphibian Group DICE - University of Kent |
|
-LAF Senior Member Joined: 03 Apr 2003 No. of posts: 317 View other posts by -LAF |
Posted: 18 Aug 2005 Basically, the sensor of your DSLR is laid out in a nice neat grid: 
LINK IN CASE IMAGE DOES NOT SHOW: http://www.sd9.org.uk/Beyer.gif This is known as a Bayer arrangement. As you can see, for every red or blue sensitive pixel, there are two green pixels (doesn't matter why for now). So, you effectively get red OR blue present on any given horizontal row of the chip. Now, if you are shooting something with a pattern that closely matches that of the bands on the CCD (imagine looking at a picket fence through barred windows) this chip design can lead to some very strange colour effects being produced. This effect is known as moire. The picture below is an example of this from a very expensive (and rubbish!) 14Mp Kodak DSLR: Full Image: 
LINK IN CASE IMAGE DOES NOT SHOW: http://www.dpreview.com/reviews/KodakSLRc/Samples/Issues/F6F M1107-001.jpg 100% Crop showing Moire: 
LINK IN CASE IMAGE DOES NOT SHOW: http://www.dpreview.com/reviews/KodakSLRc/Samples/Issues/F6F M1107-003.jpg To get around this problem, the camera manufacturers fit a filter in front of the CMOS (or CCD on some cameras) chip. This filter (sometimes called an anti-aliasing filter) serves two functions. Firstly, it filters out the invisible infra-red light that would otherwise show up on the image (camera sensors can see IR), and secondly IT BLURS THE IMAGE. It does this on purpose as, by scattering the light hitting the chip a little, any conflicting patterns are diffused and you don't get the nasty moire effect. Some cameras have stronger AA filters than others. Of course, a stronger filter means a blurier image. When you shoot a picture in JPeg, the camera applies a small to large (depending how you set it) amount of sharpening to the picture at the time of capture. The default setting is merely to compensate for the effect of the anti-aliasing filter. You would still need to re-sharpen to suit your output medium. If you shoot in RAW, then nothing is done to the information in camera. You will have to apply more sharpening manually. For 6&8 megapixel CANON dslrs (with the exception of the 1D Mk2), an unsharp mask as follows is a good starting point as it is just the right amount to correct for the AA filter: Amount: 300-320% Radius: .3 pixels Threshold: 0-10 Increasing the radius to .4 may be benneficial on 8mp canon cameras. So that's why pictures from digital slr's are soft. They're simply made that way to maximise image quality. As for RAW vs JPeg... I personally only shoot in RAW. We sometimes shoot in JPeg for clients if a rapid turn-around is required, but where picture quality is key, RAW is the only way to go. As is mentioned above, a RAW file is the information the camera sensor sees. Nothing added, nowt taken away. But you can't look at a RAW file, it's just data. It has to be previewed or processed as a viewable format (think of it like a digital negative). This means you get to do all kinds of ueful stuff AFTER you've taken the picture. For example, when you develop the RAW file on your computer, you can change the white balance to anything you want (with JPegs you are stuck with it as it's taken - to change it then loses image quality). You get to set the coluor saturation and the contrast too. You can even correect the exposure by upt to 2.5 stops either way! (and 1.5 stops either way with virtually no loss in quality at lower ISOs). There's more too, you can convert a RAW file into a 16bit Tiff image. Incredibly useful. I wrote a short piece on Tiff vs Jpeg which I'll post below. Lee. Lee Fairclough |
|
B Lewis Krag Committee Joined: 24 Aug 2004 No. of posts: 146 View other posts by B Lewis |
Posted: 18 Aug 2005 Cheers Lee, A very informative posting and one I will be printing off for future reference. The ranges for the sharpening are also very useful. Can you recommend a good book on this kinda stuff. (Not too techy mind) I don't know if its just my PC but can't see the pictures in your posting.. If its just me sorry about that... Brett. Lewis Ecology Brett Lewis Photography Kent Reptile & Amphibian Group DICE - University of Kent |
|
-LAF Senior Member Joined: 03 Apr 2003 No. of posts: 317 View other posts by -LAF |
Posted: 18 Aug 2005 The apparent quality differences between a Tiff and a max quality JPeg are very small (if you zoom a section to about 400% you will notice minor differences) so from a usability point of view there is little differences if you plan to use the image as-is. However, the fundamental differences between the two formats do have quite big implications in terms of what happens when you come to adjust or re-save the image. JPegs utilise a 'lossy' compression algorythm that saves the image in fundamental blocks of pixels. The higher the quality of the JPeg, the smaller (hence more detailed) the blocks of pixels. You can see this most clearly if you save as a really low quality JPeg, then these blocks become very visible indeed. Tiffs, on the other hand, do not use compression (well, they do offer a lossless compression option for greyscale images) and so the file sizes are much larger. Another fundamental difference is colour depth. JPegs (except for the new, and not very widely supported JPEG2000 format) are limited to 8 bit colour depth. Whereas Tiffs allow us to save in a 16bit colour depth. 8bit colour = 256 shades for each channel of red, green and blue. So, 256x256x256 = 16.7 Million possible colours within the image. 16bit colour = 65,536 shades for each channel, so with an RGB image, that gives a possible 281 TRILLION colours! So what do these differences mean in terms of use? Well, to start with, every time you re-save a JPeg (even if you re-save at the same quality setting) the image is re-compressed from how it is so you WILL lose more detail. Not much at high quality but the loss is there. So any image you fiddle with and re-save to send to a client will have effectively been compressed twice and not be of the same quality as the original. Furthermore, remember that the JPeg is made of lots of little blocks? If you need to interpolate an image (e.g. to upsize at 300dpi for posters etc) then you will also be upsizing these blocks (affectionately known as JPeg artifacts). So, what looked like a nice clean image at A4 300dpi may begin to look blocky, with visible artifacts, at say A2 at 300dpi. Tiff images don't use these blocks so will upsize much more cleanly. And the colour depth has issues too. A 16bit Tiff takes up twice the disk space of an 8bit tiff, but will give nearly 17million times the number of colour and shades within the image (a pretty good trade off!). This has great advantages in terms of shadow and highlight details, and especially when you come to adjust the levels, curves or exposure of the image. Most computers display 16.7 million colours, so obviously an image saved in 8bit or 16bit will look roughly identical on the screen. But when you adjust an 8bit image you can easily go beyond the limits of its colour range and cause colour blocking (ever noticed how if you try and adjust the exposure of a pale sky to darken it that it soons becomes patches of solid colour?). This can also cause loss of shadow or higlight detail depending on which way you go. With a 16bit image however, although still limited to that same 16.7 million colours on screen, you have a massively bigger colour space from which to work through, so running out of levels of colour becomes much less of an issue (hence, with a 16bit image, providing you have not completely blown or blacked any part in the original image, you can usually darken or lighten at least 2 stops either way without compromising quality). This extra detail can allow more accurate image interpolation too, and programs such as GenuineFractals can make awesome interpolations from 16bit files while retaining excellent quality. Lee. Lee Fairclough |
|
-LAF Senior Member Joined: 03 Apr 2003 No. of posts: 317 View other posts by -LAF |
Posted: 18 Aug 2005 Hi Brett, I'm affraid I can't really recommend any good digital photography books. Not that there aren't any, I just don't own any. I can recommend a couple of excellent photography sites with boundless amounts of information of them though: http://www.luminous-landscape.com/ http://www.fredmiranda.com/ Both of these site contain inumerable artiles written with great accuracy. The latter site is more layman friendly (the luminous landscape site does occasionally assume you have a degree in physics!), but both are very well written and explain some rather complex subjects in very clear terms. Lee. PS, will pass on message to Alison, cheers! Lee Fairclough |
|
Vicar Senior Member Joined: 02 Sep 2004 No. of posts: 1181 View other posts by Vicar |
Posted: 18 Aug 2005 In case people have not yet come across the MS RAW viewer, its pretty good (& Free !). It supports at least Canon & Nikon formats, and even allows you to view the images as thumbnails on windows explorer. Its good enough to wade through a folder of pics and delete the obviously dodgey pics, prior to JPEG (or whatever) conversion. Link here : MS RAW Viewer Steve Langham - Chairman
Surrey Amphibian & Reptile Group (SARG). |
|
GemmaJF Admin Group Joined: 25 Jan 2003 No. of posts: 2090 View other posts by GemmaJF |
Posted: 18 Aug 2005 Hi Brett, I just used the default 'sharpen' in Photoshop 7.0 and played with the unsharp mask, it was fairly obvious when I went too far with the perview in Photoshop, though Lee has given us all a bench mark (and a fantastic explanation of what is happening) so it won't be so hit and miss next time (I must admit I thought the blur was due to my wobbly hands or because I had done something awful to cameras focus at first Gemma Fairchild, Independent Ecological Consultant |
|
B Lewis Krag Committee Joined: 24 Aug 2004 No. of posts: 146 View other posts by B Lewis |
Posted: 18 Aug 2005 Just found why I couldn't get the 'RAW' mode on the 300D, it only works in the manual formats and is set to JPEG as standard in the automatic settings. Perfect sense - Thanks to Lee Brady for that one.. Brett. Lewis Ecology Brett Lewis Photography Kent Reptile & Amphibian Group DICE - University of Kent |
|
calumma Senior Member Joined: 27 Jun 2003 No. of posts: 351 View other posts by calumma |
Posted: 18 Aug 2005 I have to admit that I consider my 300D to be not too much better than an advanced point and click. I've been particiluarly disappointed with the focus and other elements of manual control. No doubt I'll upgrade again when prices come down somewhat... Since most of my shots are used in presentations (web or otherwisw), I tend to save files as (large) Jpgs, rather than RAW. Quick and easy to pull images out of the camera and drop them straight into a presentation. With regard to image post-processing, I have a rule of never changing the original file. SAVE AS... is there for a reason! That way if I decide to do multiple changes to a lossy JPG I can always go back to the original file and do all of the changes in a single save. There is still some loss, but it isn't really noticable for the uses to which I use the images. Being colour blind, I also tend not to notice some of the subtle differences in colour  Lee Brady Kent Herpetofauna Recorder | Independent Ecological Consultant |
|
Iowarth Admin Group Joined: 12 Apr 2004 No. of posts: 222 View other posts by Iowarth |
Posted: 18 Aug 2005 With my previous camera (Canon D30) I started to use RAW format largely to compensate for a high proportion of "fuzzy" pictures. With my current 10D I have reverted to JPEG simply because the result is as good as MY eyes need! On one or two occasions though I have regretted this as when printing in a large size (A3+) JPEG artefacts have become noticeable. Most of my pictures are optimised for the web or reduced to less than whole screen size though so the problem is rare. And, of course, I would never pretend that my photographic abilities to justice to either the camera or RAW format! Nonetheless the basic principle remains - RAW gives much better quality - so it does make sense to use it. Chris Davis, Site Administrator Co-ordinator, Sand Lizard Captive Breeding Programme |
|
calumma Senior Member Joined: 27 Jun 2003 No. of posts: 351 View other posts by calumma |
Posted: 18 Aug 2005 I've just bought an Epson R1800 A3 printer. It produces amazing output - well to my eyes anyway. I guess that I should also consider RAW if I intend to print more images at large sizes. Folks may be interested in the article below on processing images. Link. Lee Brady Kent Herpetofauna Recorder | Independent Ecological Consultant |
|
GemmaJF Admin Group Joined: 25 Jan 2003 No. of posts: 2090 View other posts by GemmaJF |
Posted: 18 Aug 2005 Righty, so my camera is now set to RAW (actually to both RAW and JPEG as I get the option to save as both) so I'll see how I get on with RAW files. (PS sorry Brett I didn't mention you had to be in the manual modes, I've not used the basic modes since the first couple of days because I was so fed-up that I couldn't choose the AF point when in basic - my only gripe about the 350D so far) Gemma Fairchild, Independent Ecological Consultant |
|
Vicar Senior Member Joined: 02 Sep 2004 No. of posts: 1181 View other posts by Vicar |
Posted: 18 Aug 2005 Just a word of warning with the Canon Viewer software. You cannot convert the RAWs unless you have first copied them from the camera onto hard disk. I too used both JPEG & RAW the first time (just to convince myself I wouldn't lose pics) but from then on have shot RAW only. I think you'll only get about 75 shots onto a gig card saving both formats :P - but it will be a good test: camera JPEG vs PC JPEG :P Steve Langham - Chairman
Surrey Amphibian & Reptile Group (SARG). |
|
Caleb Forum Coordinator Joined: 17 Feb 2003 No. of posts: 448 View other posts by Caleb |
Posted: 19 Aug 2005 Excellent stuff from Lee, but one small niggle: [QUOTE=-LAF] The higher the quality of the JPeg, the smaller (hence more detailed) the blocks of pixels. [/QUOTE] The blocks are the same size, no matter what the quality- they're always 8x8 pixels. The 'quality' settings determine how much information in each block is lost during compression. This isn't that important for most photos, but it does make even low compression jpeg format unsuitable for encoding graphics with high contrast and fine detail- like black and white line art or text. Personally, I always use jpeg format with my camera- it's good enough for my needs. |
|
Tony Phelps Forum Specialist Joined: 09 Mar 2003 No. of posts: 575 View other posts by Tony Phelps |
Posted: 19 Aug 2005 Bugger all this technology - I shoot on JPEG fine and do not do anything to the initial result - just 'as is' haven't even got photoshop. If my stuff can reproduce at 300dpi for books and mags then thats all I need. Still use the film camera reasonably often, the Nikon F5 is a nice piece of kit; but then again so is the D100. Perhaps its just cos I am a PC dummy 
T |
|
-LAF Senior Member Joined: 03 Apr 2003 No. of posts: 317 View other posts by -LAF |
Posted: 19 Aug 2005 Fair point Caleb, I'll get that ammended for future use. Cheers! Tony, I've always found that slide libraries and publishers always require the image unsharpened anyway, as they like to do that themselves. There's no problem sending high qaulity JPegs for publishing at A4 or less either. One issue though, most slide libraries (Alamy, NHPA, OSF .etc) will only accept Tiff files. They frequently require you to res them up to 50Mb (8bt Tiff) too. For this, shooting RAW will definately make life easier. You won't have this worry with 35mm 
Lee. Lee Fairclough |
- Image post-processing |
)