For those who are using Canon's Raw Image Task (RIT) and DPP, are you finding DPP (at least the latest version and maybe even earlier versions) to produce less details than RIT? I shot a number of white egrets over the weekend and pp them in DPP. By chance, I opened one of those RAW files in RIT and noticed alot more white feathery details than what was produced in DPP which basically blew out alot of that white detail. I lowered the exposure in DPP as much as 1 stop and only got some of the detail back and of course, everything else became underexposed. I am now going back to using RIT for quick pp. I know I should use other programs like PS but I like RIT/DPP's quick and simple interface. Anyone else seeing this disparity in detail between DPP and RIT and care to share your thoughts? Thanks.
Yes, I came to the same conclusions. I used to do all PP in DPP as a one stop service. Unfortunately RIT is limited to the camera settings and PP must be continued in another software such as PS.
My conclusion on the differences between DPP and RIT is how color falls off to white. If you want your whites to be WHITE, in my opinion RITs fall off from a color to white, transitions to a cleaner white (or maybe the right way to say it is, it has less color cast in the whites). If you want more color in your whites, my opinion is use DPP.
An example might be better here. I have Egret photos shot during the day that have very clean whites, I use RIT to keep the whites clean. I have Egret photos shot in the evening and I when I want to maintain the color cast from the late evening sun, I use DPP to process. The cleaner whites in RIT may translate to higher contrast.
thanks, cogitech, will have to try that next time.
designguy, it all kind of makes sense, i'll have to experiment with both (rit/dpp) a little more to see if I can live with only one or have to use both depending on situation.
maybe it's my misconception, but the few times i've used Photoshop to process a whole batch of raw files, i thought it to be cumbersome and slow. I now only use ps to refine those handfuls of photos I want to spend time on.
Here's my decision making process (or is this considered workflow?).
1. Process with RIT or DPP
2. Import 8 bit jpg into lightroom for management purposes (8 bit for smaller file sizes).
3. Minor tweaks in lightroom (LR).
4. If my changes in LR creates problems with the jpeg file, I import a 16 bit tiff and apply the same edits.
Here's my thinking on this methodology. I agree that Photoshop or Lightroom are slow and tedious. But RIT and DPP don't always give me the level of incremental adjustments I want. RIT and DPP give me better color, but sometimes adjustments are too big. Either to much contrast or not enough. So I do the large global color treatment of an image in canon software to give me better color. Using RIT or DPP based on how I want my whites to fall off. And then I tweak and manage files in Lightroom. If you manage your xmp correctly, you can use the same data in lightroom to open files in photoshop.
This may sound cumbersome, but it's very simple actually. Remember, I'm a graphic designer, and 15 years ago desktop publishing software didn't have all the bells and whistles where we wanted them in each piece of software. So designers used the strongest part of each software package for a specific purpose. As software vendors reviewed designer's workflow (workflow wasn't even a word back then) they slowly incorporated what we needed into really powerful software. I now use InDesign, but at one time Quark XPress was king and Adobe came up with a better way. If you think about it though, photoshop is kind of a hodge podge...
Ok, I'm done blathering! I hope it's useful!
Chris
wing tong wrote:
For those who are using Canon's Raw Image Task (RIT) and DPP, are you finding DPP (at least the latest version and maybe even earlier versions) to produce less details than RIT?
No.
I just replied to Edward's post, which made me curious about RIT vs. DPP, and based on a couple conversions of my 1DIII files, with settings as equal as possible (WB, picture style, contrast, saturation, sharpening), the DPP (v3.4.1) conversions hold better overall fine detail, though I don't have a file handy with highlight issues similar to your sample.
Were the sharpening settings different for each conversion, because the RIT version seems like it is much sharper in selective areas.
edwardkaraa wrote:
Yes, I came to the same conclusions. I used to do all PP in DPP as a one stop service. Unfortunately RIT is limited to the camera settings and PP must be continued in another software such as PS.
Not so sure about that because the version of RIT on my computer does allow me to change all the typical parameters, such as WB, picture style, sharpness, contrast, etc... It's version 2.6 in conjunction with Image Browser 5.8 (which I just noticed is no longer current).
Ron
Edit: Just downloaded Image Browser 6.1.1.17 (with RIT 3.2) and the RIT conversions look the same as the previous version I had installed.
Edited by rscheffler on Jun 02, 2008 at 06:23 PM GMT
In the above samples, the top one is RIT and the bottom one is DPP. To me, the RIT seems to have more detail in the white feather areas while the DPP has slightly more detail (or at least sharpness) in the eye and beak areas plus the DPP version looks a little more saturated. Both RIT and DPP settings had 0 value for contrast/colortone/colorsaturation with the RIT having 7 (max) for sharpness and the DPP at 8 - but I've tried a number of different settings as well and the RIT version always shows more detail or shadows in the white feather areas.
Note, edwardkaraa's note regarding RIT means RIT has the same controls as you can find in the camera even though you can still adjust them in RIT.
i guess what is bothering me with the RIT conversion is that the sharpening appears to be very uneven. The feather areas are sharpened a lot while the beak area looks softer than the DPP version. This is similar, though more dramatic, to Edward's pbase link above. If you compare those, the highlight areas in the RIT conversion are much sharper while at the same time the darker areas are slightly softer than the DPP conversion. Unfortunately Edward didn't indicate what, if any sharpening was applied to the DPP conversion compared to the RIT conversion (which was at +3) in his post here https://www.fredmiranda.com/forum/topic/652168 .
Just curious, but if you select the Linear option in DPP, just how much detail do you actually see in the bird's head? It will give you an idea of how much highlight detail you'll actually be able to recover with DPP. If it's blown out in Linear mode, then there's no hope of extracting more detail.
Ok, I've just read edwardkaraa's topic which I didn't see before posting my topic (edwardkaraa's topic name was not so obvious). One suggestion was to try Neutral picture style in DPP. I tried it, it brought some of the details back (maybe 25%) but RIT still has better details in the white areas at least.
Edited by wing tong on Jun 02, 2008 at 03:52 PM GMT
In my experience, DPP is sharper than RIT for most cameras (I've owned 300D, 1Ds, 5D), but doesn't seem the case for the 1DsII.
RIT seems to be sharper for in focus areas, while DPP seems to sharpen slightly OOF areas, thus compensating for poor Canon lens performance and the strong AA filter.
RIT seems to have a more sophisticated sharpening and demosaicing algorithms which seem to blur extremely fine detail in dark areas while emphasizing micro-contrast in lighter areas. As a result, less grain is visible in the darker areas and they appear smoother.
For the bird shots, if you have DPP set to "Standard" Picture Style, then the contrast within that style is usually boosted too much, and will blow highlights. Try "Neutral" Picture Style and see if the highlights are recovered.
Looking at the shrimp crops, I think DPP produces a sharp, solid, in-your-face image that could take even more sharpening in PP, while RIT produces some sort of digital-ish, oversharpened/mushy type of conversion, with more contrast, therefore, maybe it looks sharper at first glance. Hard to explain.
In these samples we can see the better colour and highlight tones rendition of RIT against DPP. I think that we have to further investigate the whole situation. For sure Lightroom and PS (which they use the same RAW conversion engine) have colour rendition problems and I think that in most cases are not so usefull software for raw processing). I suggest to make aletter to adobe about how useless RAW conversion they make or ask form Canon to make specialised Raw engine plugin for PS or Lightroom.
This is interesting. If the situation warranted, I'd be tempted to sandwich RIT and DPP conversions and selectively reveal areas of each. It's as if Canon designed RIT to give a better impression of highlight sharpness/detail vs. DPP.
It's just a guess, but if RIT is aimed more at "amateur" users who may use it as the final step and be more prone to clip highlights during exposure, then it might act as a safety net, to a certain degree. In contrast, DPP aimed at more experienced users may assume there will be fewer issues with highlight clipping and that images will go through further processing prior to sharpening. In the shrimp sample, I think one can see that the same information exists in the DPP conversion, just that it isn't accentuated to the same degree as the RIT conversion (it helps to look at the sample crops at his pbase gallery, with each in separate browser tabs, then quickly tab back and forth). From my experience primarily with DPP, I've found I can usually recover highlight details fairly well through careful combination of exposure compensation, picture style, contrast slider and tone curve manipulation, but a dedicated highlight recovery slider as found in other converters would be a great option to focus strictly on the highlights without severely affecting the rest of the image's characteristics.
But I do wonder how camera specific this is. Definitely seems to be the case for the 1DsII, as Edward's sample shows. I also seem to recall Rob Galbraith stating that his experience with RIT indicates it extracts a touch more detail, but unfortunately can't find that link at the moment.
