Does gimp have optical filter imitations?

[godek]

I am using gimp 2.9.8. Right after I use nufraw and open in gimp it is imported as 8 bit. My camera is a slr digital camera from 2005. Could it be possible it is not able to output more than 8 bits? It is a nikon d50 I plan on getting a newer one soon.

[Ofnuts]

(12-26-2017, 02:57 AM)godek Wrote: I am using gimp 2.9.8. Right after I use nufraw and open in gimp it is imported as 8 bit. My camera is a slr digital camera from 2005. Could it be possible it is not able to output more than 8 bits? It is a nikon d50 I plan on getting a newer one soon.

Most likely a limitation of nufraw. Digital cameras have 12 or 14-bit sensors (even if due to noise the actual dynamic is around 10-bit at high ISO settings)(your D50 is 12-bit).

But IMHO it would be better to use a more complete Raw application (RawTherapee, Darktable, Photozone) that can do a lot more things using the sensor data (like several different de-noising algorithms) and totally outperform Gimp in the color-processing department. For instance RawTherapee has a large collection of film emulation filters.

When you start using this kind of application to any significant extent (in other words, trying to get something better than the in-camera JPEG) you find that you spend more time with them than with Gimp. It not RT which is a pre-processor for Gimp, but Gimp which is a post-processor for RT (to remove unsightly elements, add signature, etc...).

All these free, open-source applications are discussed on pixls.us.

[rich2005]

Not a limitation of nufraw as such. 

Going back to post #7 If you have a nufraw that has the code for Gimp 2.9 included so the nufraw-gimp plugin works there will be an option for 16 bits.  same screenshot as before: https://i.imgur.com/xvXB720.jpg

Since you are using Manjaro linux, did you compile nufraw yourself or get from the manjaro repo? If using the stand-alone nufraw rather than the plugin you should still be able to export a 16 bit tiff to open in Gimp 2.9 as a separate action.

When it comes to the other applications, personally I find their interfaces cluttered and not to my liking. However the same applies, more recent versions are needed to send a 16 bit image directly to Gimp 2.9

If/when you try RawTherapee / Darktable / Photoflow you can check their availability in Gimp 2.9 from Edit -> Preferences -> Import Image as screenshot: https://i.imgur.com/CZcKjF6.jpg

[godek]

As you can see it says 8bits at the top but it allows me to change it to 16 bits in the menu though.


I selected 16 bit output as you see before I clicked on the gimp icon.
Nufraw is version 0.40-1


I get this error when I try exporting the image from nufraw to gimp but it doesn't keep the image from being loaded though.




I get this error when I import a 16bit tiff from nufraw into gimp.



It looks like this when loaded into gimp which is not what it looked like in nufraw.

[godek]

Is there any books, online courses and or tutorials you guys recommend for learning all these programs including gimp? I know there is gimp for photographers. But what about darktable, rawtherapee, rawstudio??? Can you just do a class room in a book with photoshop and try to apply what you learn there to open source apps?

[Ofnuts]

(12-27-2017, 12:23 AM)godek Wrote: As you can see it says 8bits at the top but it allows me to change it to 16 bits in the menu though.

Yes, but you have lost data... From the 14-bit sensor you image can have 16000 values per channel. When you load as 8-bit you truncate the small variations and only keep 256 values. Going back to 16 won't revive these data. To be more explicit, let's use decimal values: with you 14-bit sensor you have 16000 values (let's says 10000 to be decimal), it's a bit as if you used 4-digit decimal values from 0.0000 to 1.0000: 0.0000, 0.0001, 0.0002.... to 0.9999, 1.0000. With 8 bits, you have only 256 values (let's say 100 to be decimal) so you are using number with two digits; from 0.00, 0.01 to 0.99, 1.00. When you load your 14-bit image in 8-bit Gimp, a value such as 0.1234 is rounded to 0.12, as will 0.1200, 0,1175, 0.1245... And if you set Gimp to 16-bit mode all you get is empty decimals, you 0.12 becomes 0.1200 but the initial values 0.1200, 0,1175, 0.1245 have all become 0.1200 now. Of course, further processing can use these extra decimals to avoid round off errors, but the damage to the initial data is done.

(12-27-2017, 12:23 AM)godek Wrote: I selected 16 bit output as you see before I clicked on the gimp icon.
Nufraw is version 0.40-1

I get this error when I try exporting the image from nufraw to gimp but it doesn't keep the image from being loaded though.

I get this error when I import a 16bit tiff from nufraw into gimp.

It looks like this when loaded into gimp which is not what it looked like in nufraw.

Given the error message the output isn't unexpected. Two cases:

1) the Nufraw author explicitly states that Gimp 2.9 is explicitly supported (which would be kind of mandatory to load 16-bit images in Gimp): get in touch with the author to have the bug fixed in Nufraw
2) you have been told that Gimp will support 2.8 plugins for opening images: this is a Gimp bug: report the bug on http://bugzilla.gnome.org (but you'll still be limited to 8-bit when the image loads)

A Workaround is to use a separate demosaicing app (that can also be Nufraw) that exports a 16-bit TIFF, and then open that TIFF in Gimp. If the problem persists, the subject TIFF is a good clue to give to developers of both sides, to determine if the TIFF is actually invalid (Nufraw problem) or if Gimp has trouble loading TIFF files (Gimp problem).

[Ofnuts]

(12-27-2017, 01:25 AM)godek Wrote: Is there any books, online courses and or tutorials you guys recommend for learning all these programs including gimp? I know there is gimp for photographers. But what about darktable, rawtherapee, rawstudio??? Can you just do a class room in a book with photoshop and try to apply what you learn there to open source apps?

RawPedia (which is actually a expanded RawTherapy user's guide) has plenty of info.

The "theory" is common to all applications.

[rich2005]

@godek

Quote:I selected 16 bit output as you see before I clicked on the gimp icon. ...Nufraw is version 0.40-1..

Not quite the latest version although the update does not affect linux. The fact that you have a Gimp icon means that is not the plugin, it is the standalone version.

If you can not get it working, could be lots of things, older dependencies for example. I do not know Manjaro, so I can not comment.

Not worth pursuing. What you want is the photo into Gimp 2.9 for editing. Try either Rawtherapee or Darktable.

Without any sort of plugin installed you should get this message: Note the required versions.



Rawtherapee is probably the best bet (although replacing a 5 MB nufraw plugin with 80 MB application is annoying)

If the version is correct and Gimp 2.9 picks up the installation look in Edit -> Preferences -> Image Import for confirmation.



At a base level, without any tweaking, although heed Ofnuts advice on that, once Gimp 2.9 -> Open has the image up in Rawtherapee, closing sends it into Gimp 2.9



In Gimp 2.9 opens as 16 bit

[godek]

Thanks guys, now it works! I used rawtherapee to edit the image and then I saved it as a 16bit tiff and it works when I open it in the gimp!!! So it looks like it is a bug in nufraw cause it looks messed up in other programs to that I open the tiff created by nufraw.
edit: Maybe I did something wrong or it was just a fluke a one time thing or I didn't do it again what caused the bug in the first place cause I tried it again and the image looks fine loaded in gimp as 16bit.


In your guys opinion which is the best Raw photo editor that is open source and available for Linux?

Can you send 16bit images to Walmart for printing or do you need to go to a specialty store?
Also, can you have them use the embedded color profile that you used to calibrate your monitor with their printer?
Walmart or is this specialty store only also?

Can you print 16bit images at home with a consumer based printer?

Such as one of these?
https://www.thinkpenguin.com/catalog/printer-all-onee-gnulinux
 
Do you guys recommend this for color calibration? Going to use displaycal.
http://www.hughski.com/colorhug2.html

[rich2005]

Printing from Gimp in linux is almost as bad as in Windows. You are in the hands of the printer manufacturer to provide support and printer drivers. The ones you show HP are supported. Epson is not bad, I use a Brother inkjet and they provide drivers for it. Canon generally do not support linux very much. The printer might work but with limitations. For my home printing - some photographs - some DVD's - some work printed on art paper the 3 colour + black Brother is adequate.

Not too sure how up-to-date this is but it is a start https://wiki.linuxfoundation.org/openpri...abaseintro

If your linux distro provides gutenprint and the printer is supported, gutenprint provides an additional improved printing interface.



16 bit vs 8 bit I do not think it makes much difference for home printing, the printer is probably going to resize the image anyway. For a commercial printer, ask them. If they want a jpeg then it is going to be 8 bit - that is jpeg standards. Otherwise, they would need to accept a tiff. File size might come in as well. Walmart? look at their specs. There are some excellent printing companies using high grade inkjet printers for smaller sizes and color laser printers for large images. Not cheap. Example, http://www.digitalab.co.uk/photo-prints-...t-results/ - wrong country for you

Colour calibration, Ofnuts asked the question one time see: https://discuss.pixls.us/t/color-calibra...nux/2638/3