converting RAW to JPEG

[Toanaldino]

hi guys,

I use lightroom to edit photos, everytime I convert to jpeg. The quality of the images are somehow downgraded and skin tone is different.

I just like to know if there are techniques for coverting to JPEG and how improve skin tones in photos.

Cheers

 

[tirediron]

Check your options/properties (I rarely use Lightroom so I can't remember where, Help is your friend), it sounds to me like there's a default compression setting for the .jpg which is set to low or medium.

 

[JerryPH]

This came up somewhere in my reading before.

It did indeed have something to do with the default settings that LR uses. I am 90% sure that tirediron's suggestion is right on track.

 

[Garbz]

When you export all the options should be visible. Check the compression slider is at 100% and the colour space is set to sRGB and not AdobeRGB or ProPhoto. The final JPEG should look like the working product in lightroom.

 

[prodigy2k7]

When you export all the options should be visible. Check the compression slider is at 100% and the colour space is set to sRGB and not AdobeRGB or ProPhoto. The final JPEG should look like the working product in lightroom.

I second that, it asks you a few options when you click export, such as where to export the images to, the image type etc...

 

[Toanaldino]

Thanks guys,

just one more question, what resolution should I set the DPI too?

 

[pez]

When you export all the options should be visible. Check the compression slider is at 100% and the colour space is set to sRGB and not AdobeRGB or ProPhoto. The final JPEG should look like the working product in lightroom.

Hmm. Why use sRGB when AdobeRGB is a larger gamut? Display purposes?

 

[Garbz]

just one more question, what resolution should I set the DPI too?

Ignore it. This is an option that you set when you print only, or when you are giving someone a file to print in which case it's the horizontal resolution divided by the width of the photo in inches. It changes nothing in the photo other than embedding that number in the file somewhere.

Hmm. Why use sRGB when AdobeRGB is a larger gamut? Display purposes?

Colour management hell. AdobeRGB may have a larger gamut but ask yourself the following:
1. Can you display the wider gamut?
2. Can you print the wider gamut? (I highly doubt so unless you're paying $10+ for a print or have a $1000+ printer)
3. Hear's a big one: Can you reliably know what the rendering intent that your client or gallery visitor will use? Having an ICC profile does not mean someone you send the photo to will get the same results.
4. When you save the wider gamut files do you enjoy having to convert them everytime you upload them to the net or send them to a friend?
5. If you can't see the wider gamut can you reliably edit the wider gamut? How do you know what the colours will look like?
6. Do you work and save all your files in 16bit? A larger gamut means the spaces between colour bands are larger making posterisation worse. 16bit files fix this but aren't supported in JPEG.
7. Will you miss it? It may sound good on paper but take a photo of something with your camera set in RAW or jpeg with AdobeRGB support. In photoshop open the AdobeRGB file, setup softproofing for the sRGB profile, and the select the option for show gamut warning. That tiny tiny portion of the screen is what may look slightly more saturated if you were actually capable of displaying saturation.

All of this goes doubly so for ProPhoto which has a gamut so wide it's ful of imaginary colours which don't actually exist. Quite practical It all makes sense if you take a photo (most cameras are native AdobeRGB, but few scenes grander than a sunset will make use of the extra gamut), then edit that photo on a screen with a wide gamut, and then print it on a printer with native AdobeRGB support. Other than that, sRGB is the standard and the only way to make the process reliable.

Btw I have an AdobeRGB monitor. I rarely work in the larger gamut, unless I'm doing a HDR of a sunset or something else that colourful, even then often I check how much I would lose in sRGB, realise how little difference it makes and don't even bother saving in AdobeRGB.

 

[Bifurcator]

1. no. me thinks not.
2. I thought you could on a $300 inkjet... no? (I don't think I've tried yet)
3. n/a
4. save both?
5. but there is a difference in display colors. Isn't that enough? How much more is there?
6. Yes. I haven't noticed any banding in AdobeRGB. JPEG? Yuck!
7. Some red tones are completely different here. I don't notice or "miss" it otherwise tho - unless I'm purposely comparing.

 

[Garbz]

Pointing out a quick contradiction here:
If your red tones look different here (7), yet you can't display the wider gamut (1), could it be because of an incorrect rendering intent (3). How do you know when you're displaying the correct ones if you can't display the larger gamut, and more so if your screen displays the sRGB gamut and you can visibly see a difference when you convert from AdobeRGB TO sRGB then isn't this indicative of a loss of colour accuracy due to conversion since it implies that in the original wide gamut would display the original colour on a correct wide gamut display?

The printers some of the higher end CcMmYyK printers and the CMYRGBK printers can display the gamut. And the branding won't be instantly visible, this is something for the pixel peepers, but like 16bit editing it becomes relevant if you do quite a bit of editing on the files since when you extend the dynamic range without extending the bit depth your average processing error increases too.

Mind you this may sound like a moot point to some, and it is to a degree. There are no doubt quite a few people on this forum who will disagree with me, but I will bet a few lira that those are the people who either have nicer screens, printer their own photos either at a pro lab or with a nicer printer, use 16bit processing to get the most out of their pictures, and otherwise have something to gain.

But for the majority of this forum who will take a photo and store it on their computer and look at it on their screens, more data for data sake in this case is a step backwards since you're literally clipping the colour channels of the data you are displaying. It just looks more subtle than highlight clipping and would be visible as a shift in colour (e.g. from Green to Yellow when you shoot a high contrast plant).

If you're browser doesn't support color profiles you're not seeing what I'm seeing!

See this makes (3) very much applicable. More so to people with no rendering intent whatsoever

 

[Bifurcator]

Pointing out a quick contradiction here:
If your red tones look different here (7), yet you can't display the wider gamut (1), could it be because of an incorrect rendering intent (3). How do you know when you're displaying the correct ones if you can't display the larger gamut, and more so if your screen displays the sRGB gamut and you can visibly see a difference when you convert from AdobeRGB TO sRGB then isn't this indicative of a loss of colour accuracy due to conversion since it implies that in the original wide gamut would display the original colour on a correct wide gamut display?

Well, I only think I can't display the full spread of the gamut. I don't actually know. (My current monitor is a 6bpp screen with an 8bpp pallet) And it might be that with AdobeRGB or any other color-space that it will indeed select and display the pallet colors most populated in the file. So is there's "Red 100" in AdobeRGB but "Red 100" doesn't exist in sRGB and the image has "Red 100" in it then the app or whatever decides what to show (I assume it's the app) might display "Red 100" even though it can't actually display the entire gamut. When CYMK editing in PS printed and displayed colors are VERY different (especially reds!) then when printing and editing in RGB modes. And I can never get the same reds in RGB mode no matter how much I tweak. And visa versa. So, it may be selective with different weights and pallet selections available in different modes. Honestly, I dunno. Those are all just guesses.

The printers some of the higher end CcMmYyK printers and the CMYRGBK printers can display the gamut.

You mean 7-ink printers? Yeah, those are $350 now. Mine is 5-ink on purpose because I'm totally CHEAP! $0.15 a print "L" size in 5-color and $0.27 a print "L" sized in 7-color. The actual printers are the same price though. +/- $50.00

And the branding won't be instantly visible, this is something for the pixel peepers, but like 16bit editing it becomes relevant if you do quite a bit of editing on the files since when you extend the dynamic range without extending the bit depth your average processing error increases too.

Banding is very noticeable on my 5-ink printer anywhere there are closely spaced tones spread over a reasonable sized distance. I hear it's not so with the 7-ink models but that's the trade off I took for the cheeper prints. If I hate it terribly I can select it and fill with a solid color or a more diverse gradient to eliminate it. So far I've always been too lazy and just think "crappy epson/canon printers, bah!"

Mind you this may sound like a moot point to some, and it is to a degree. There are no doubt quite a few people on this forum who will disagree with me, but I will bet a few lira that those are the people who either have nicer screens, printer their own photos either at a pro lab or with a nicer printer, use 16bit processing to get the most out of their pictures, and otherwise have something to gain.

Let'em disagree. We'll have ourselves a nice discussion. For me it's not moot at all. I want to know what's going on under the hood and what my options are so if I need to skew something (write a program) or use that info to my advantage somehow in editing I'll know more about what I'm doing. But I of course use, edit, and store in 16 bit.

PS those who "have something to gain" are the funniest. :mrgreen:

But for the majority of this forum who will take a photo and store it on their computer and look at it on their screens, more data for data sake in this case is a step backwards since you're literally clipping the colour channels of the data you are displaying. It just looks more subtle than highlight clipping and would be visible as a shift in colour (e.g. from Green to Yellow when you shoot a high contrast plant).

Hmm. Yeah. I guess that's true (about the wasted space thing) but what about in 4 or 5 years if the new world order hasn't put us all in the ovens I mean.. Won't more computers, (display cards and monitors) be 16 bit savvy? What then when all you're coolest shots from long ago are in a dreaded 8-bit space.

Also, I don't see how anything is clipped differently. When you convert it into 8 bit from 16 bit (that was created from the 12 or 14 bit camera files) isn't it the same exact thing as converting it to 8 bit. Only it's done on the fly so to speak, and not actually saved that way? I thought so anyway.

See this makes (3) very much applicable. More so to people with no rendering intent whatsoever

No rendering intent??? At all? So just send them an empty file padded with all 0's to the same size. I'm not understanding this part at all. If they have 12, 14, or 16 bit files then it's overkill for rendering an 8-bit display (size-wise) - granted. But if you're keeping both the 8 and 16 bit formats on your storage device, just send them the 8-bit files and no problem. Tho in actuality, the only problem I can see is the file size. It's going to render to their screen the same or better than an 8-bit depending on their display subsystem and monitor's ability (or print if rendered to a printer). What am I missing here?

 

[manaheim]

^^^ There's one other thing which I don't think you guys mentioned, but I'm tired so I may have missed it.

sRGB also happens to be the most consistent color space for the web and for general JPG-ready viewers. There are a number of browsers that don't color manage properly, and overall sRGB is going to look the most consistent when viewed by other people.

What someone told me once... (and in fact, Garbz later paraphrased roughly the same thing)

• Capture in RAW.
• Convert to 16 bit ProPhoto (or at least AdobeRGB)
• Do modifictions in Photoshop (or whatever you have)
• Save final image for printing or re-editing as 16 bit TIFF.
• Save display image as 8 bit sRGB JPEG.

The one modifier being unless you wanna keep around tons of huge images, you're ultimately not going want to save lots of TIFFs. I usually only keep a few around for the more critical images, knowing I can easily re-convert a RAW if I need to later on.

Listen to Garbz and Bi... dudes know what they're talking about. :thumbup:

Hope this helps.

 

[Bifurcator]

Yup! right you are. On all counts except that I know what I'm talking about of course.

We didn't cover those I assume because they were covered in a previous thread and we all agreed. This so far, is more the whys and why nots of using sRGB and AdobeRGB and when which is appropriate or advantageous. I think I don't really know well - so I'm trying to get a finer point on it. And please don't send me to abu ghraib for whittling on the plane or not knowing where I am.

OK, that wasn't funny... Sorry.

 

[Garbz]

Indeed the future holds some great things for us. LEDs have a pure single peak spectrum and can display a colour completely on the edge of the CIE1931 chart. Already there are some cheap displays (Dell make them) which display the full AdobeRGB gamut, and well there are massive complaints about them because people don't know colour management and see everything as being way to saturated. This is a software problem. Windows still does not have full colour management support built in. The hardware is coming. OLED displays we'll see in the next few years will display all of the real colours in the ProPhoto gamut. So it may be worth archiving in AdobeRGB

I'm just pointing out for processing purposes the idea of converting to the largest colour space and playing and converting back, and saving in one format, and preparing a file to print in another and all that jazz is wrong simply because the conversion is a lossy process full of uncertainty. E.g. Photoshop for an identical set of colour profiles gives you all 4 possible rendering intents, in each case allows you to use Black Point Compensation if you so wish, and gives you the choice of using the Windows ICM engine or the Adobe ACE engine to the the conversion, each giving subtly different results. Couple that with editing a ProPhoto file where the few colours that are rich enough to make use of it are rich enough that you can't display them and thus have no idea what they are actually suppose to look like and :banghead:

/EDIT: may have been a more appropriate emoticon.

 

[Bifurcator]

I see. Yup. Makes perfect sense.

Informative, to the point, and made me smile (more than once). Thanks!

--
Now if I could just learn how to do that..