userid0
Member
Hi everyone,
i shot my last roll of HP5+ at nominal ISO400 and developed with Rodinal (1+50, 20°C, 11 min, Ilford agitation). Then a friend of mine, who owns a relatively simple negative scanner, scanned the negatives for me. What came out are always rather flat/greyish pictures as you can see by the one attached (which are not processed but directly from the scanner).
I had similar looks with other films (tri-x and apx100) at varying iso with different development.
Currently, I do not know what reason causes this look. I always have to give it a fair amount of tweaking in the computer to look reasonable.
Any ideas and/or suggestions on a valid combination of development and scan procedure to make the negative quality better?
Cheers and thanks,
Walter
i shot my last roll of HP5+ at nominal ISO400 and developed with Rodinal (1+50, 20°C, 11 min, Ilford agitation). Then a friend of mine, who owns a relatively simple negative scanner, scanned the negatives for me. What came out are always rather flat/greyish pictures as you can see by the one attached (which are not processed but directly from the scanner).
I had similar looks with other films (tri-x and apx100) at varying iso with different development.
Currently, I do not know what reason causes this look. I always have to give it a fair amount of tweaking in the computer to look reasonable.
Any ideas and/or suggestions on a valid combination of development and scan procedure to make the negative quality better?
Cheers and thanks,
Walter
Attachments
Brian Legge
Veteran
So flat negatives, straight from the scanner, are relatively typical. Contrast adjustments are typically required.
That said, nothing in these negatives is near black or white. As a result, you won't get much dynamic range from the scans. Typically when scanning, you either automatically or manually set the black and white points so that the resulting scan maximizes the useful range. I'm guessing your friend is either skipping or not familiar with that step.
That said, nothing in these negatives is near black or white. As a result, you won't get much dynamic range from the scans. Typically when scanning, you either automatically or manually set the black and white points so that the resulting scan maximizes the useful range. I'm guessing your friend is either skipping or not familiar with that step.
AndersG
Well-known
How do the actual negatives look? Looking at them will give a strong hint on whether the problem is in the film world or in the digital world..
userid0
Member
Hi Brian,
thanks for the fast reply. Well my friend actually has a rather simple negative scanner which is a Rollei DF-S 100 SE. According to what he said, he doesn't have any possibility to adjust the black and white points.
So from your opinion, it is not a development issue but more of a scanning issue?
I guess, i have to spend a little bit of money into my own film scanner.
Cheers,
Walter
thanks for the fast reply. Well my friend actually has a rather simple negative scanner which is a Rollei DF-S 100 SE. According to what he said, he doesn't have any possibility to adjust the black and white points.
So from your opinion, it is not a development issue but more of a scanning issue?
I guess, i have to spend a little bit of money into my own film scanner.
Cheers,
Walter
userid0
Member
Hi Anders,
well what shall I say...
The negatives look like they have good exposed blacks but there are no complete black spots on them (which would i turn mean to have not fully used whites in the negative).
Do you have other indicators to look for at the negatives?
Cheers,
Walter
well what shall I say...
The negatives look like they have good exposed blacks but there are no complete black spots on them (which would i turn mean to have not fully used whites in the negative).
Do you have other indicators to look for at the negatives?
Cheers,
Walter
Highway 61
Revisited
There is probably nothing wrong with the negative (45s of routine job in PhotoShop)... 
What does the scanner does ? It takes a digital photo of your negative. The goal is to have the scanner capture all the information which is on the negative. Then you have a digital file you must post-process with a capable imaging software so that it looks good.
Back in the 100% film photography days, did you frame your negatives as they came out of the developing tank ?
No you didn't.
At the end of the day, no scanner, even the most expensive ones, will give you a good looking photo straight away. There is nothing more stupid that this darn "directly out of the scanner with no post-processing" thing. You absolutely must post-process the files yourself, using the imaging editor curves and levels, a calibrated screen, and... your eyes.
What does the scanner does ? It takes a digital photo of your negative. The goal is to have the scanner capture all the information which is on the negative. Then you have a digital file you must post-process with a capable imaging software so that it looks good.
Back in the 100% film photography days, did you frame your negatives as they came out of the developing tank ?
No you didn't.
At the end of the day, no scanner, even the most expensive ones, will give you a good looking photo straight away. There is nothing more stupid that this darn "directly out of the scanner with no post-processing" thing. You absolutely must post-process the files yourself, using the imaging editor curves and levels, a calibrated screen, and... your eyes.
userid0
Member
Hi Highway61,
thanks for the nice work on the negative. This is what i normally also do in lightroom but i was wondering what the best process for this hybrid workflow is and if maybe something is wrong with my development or the scanning.
What is better: Adjust the black and white points with the scanner software while scanning or afterwards in PS or lightroom. I guess you loose dynamic range when doing it afterwards, right?
Thanks,
Walter
thanks for the nice work on the negative. This is what i normally also do in lightroom but i was wondering what the best process for this hybrid workflow is and if maybe something is wrong with my development or the scanning.
What is better: Adjust the black and white points with the scanner software while scanning or afterwards in PS or lightroom. I guess you loose dynamic range when doing it afterwards, right?
Thanks,
Walter
nongfuspring
Well-known
I don't know how your actual negs look, but the scan seems normal to me. If you can't set the black/white points in the scanning software you can do it in photoshop by using the black and white eyedropper in the curves tool (amongst other methods).
The PP of film negatives can actually be pretty fun and you can experiment with contrast etc more easily than in the darkroom.
The PP of film negatives can actually be pretty fun and you can experiment with contrast etc more easily than in the darkroom.
userid0
Member
Thanks for those clear words. That sets my world again on track and indeed, post-processing can be a lot of fun!
AndersG
Well-known
I'm certainly no expert, but to get some idea how the wet processing went I'd look at:
- The film-base - is it clear?
- The film leader that has been thoroughly exposed to light - is it really dense? (If not the film is probably under-developed.)
- The actual negatives -
--- How dense are the thinnest parts compared to the film-base? (As thin might mean under-exposure or under-development - but not if it is in parts you intended to be black).
--- How dense are the darkest parts? (Only very very bright spots, if any, should be as dense as the film leader).
nongfuspring
Well-known
Someone correct me if I'm wrong but I'm pretty sure that changing the B/W point in scanning software is the same as doing it in post.
Technically by making the whites whiter and the blacks blacker you're reducing dynamic range - the problem with "flat" negatives is that you have too much dynamic range.
Highway 61
Revisited
Theorically, yes, but most of the scanning softwares are poor imaging softwares and use very destructive algorithms to do so, while PhotoShop doesn't, especially if you scan as TIFF 16bits per layer, then work on such a file in PhotoShop.
So : drive your scanner with a scanner software and post-process your file with a photo software.
As always, I am advising everybody to look at Markus Hartel's website, especially the "Learn from Markus" pages (older entries).
Ranchu
Veteran
Just move the black and white clipping points in, I don't think it matters if you're outputting 16 bit tifs, but you may want to chop some off in the scanner software (if the software doesn't do it) if you're outputting jpegs. I don't remember seeing any 'destructive algorithms', I think that's just adobe's marketing efforts. Also, direct flash looks flat!
Attachments
userid0
Member
Guys, thanks for the hints. I guess, I will reassess the negatives again with post-processing.
Cheers,
Walter
Cheers,
Walter
Pherdinand
the snow must go on
scanner vs photoshop optimizing:
The problem is, there are 2 ways to do this in the scanner SW (and it's not clear what scanner sw uses what method).
You take a set of color values for each pixel from 0 to, say, 255. (just as example. Normally this goes to higher values )
Ideally, you have a nice gaussian distribution of the number of pixels vs the scale of 0-255.
However, depending on the negative material, usually the real situation is a distorted gaussian shifted to the middle of the scale: say, from 0 to 50 and from 200 to 255 there will be no pixels at all.
This means that, no deep blacks no bright whites appear on the scanned version.
You have 2 ways to correct this in the scanner:
You tell the scanner to move the "edges" closer to the light level where the pixels start appearing i.e. you tell the scanner to "map" the existing pixel distribution with intensities 51-249 to the full 256-levels capability of the scanner. This way the conversion from analogue lightness levels measured towards the digital (step-like) distribution will be using the full scale by "stretching" out the distribution and creating in-between levels with real distinct digital values correspondign to real analog measured values.
The other way is, you tell the scanner to throw away the 0-49 range and the 200-255 range and take the already scanned 50-199 values and map them on a scale of 0-255 leaving uniform distributed gaps between the values i.e. example every 3rd light level will have no pixels associated. So, example, there will be 2000 pixels with light level 120, 2001 pixels with light level 121, 0 pixels with ligth level 122, 2004 pixels with 123, 2005 with 124, 0 pixels with 125 etc etc.
You get a comb-like histogram.
This can show up ugly on large smooth color or lightness transitions such as a blue sky.
Photoshop and SOME scanner SW do this a bit more clever, it can interpolate and change some pixel levels so that the gaps get filled in.
What IDEALLY a scanner SW SHOULD do (but it doesnt - maybe some high end drum scanners do?) is to adjust the illumination of the frame before scanning, based on the pre-scan, so that even iwth a super dark or too light negative, the sensor will be used efficiently i.e. its sensitivity will be covered. To use the full analog scale of the sensor. This would translate to automatically stretching out the histogram to the full 0-255 scale in the correct way.
It's like adjusting the ISO rating on a digital camera sensor versus adjusting the light reaching the sensor by the aperture or shutter time.
I don't know why scanners don't do this.
The problem is, there are 2 ways to do this in the scanner SW (and it's not clear what scanner sw uses what method).
You take a set of color values for each pixel from 0 to, say, 255. (just as example. Normally this goes to higher values
)Ideally, you have a nice gaussian distribution of the number of pixels vs the scale of 0-255.
However, depending on the negative material, usually the real situation is a distorted gaussian shifted to the middle of the scale: say, from 0 to 50 and from 200 to 255 there will be no pixels at all.
This means that, no deep blacks no bright whites appear on the scanned version.
You have 2 ways to correct this in the scanner:
You tell the scanner to move the "edges" closer to the light level where the pixels start appearing i.e. you tell the scanner to "map" the existing pixel distribution with intensities 51-249 to the full 256-levels capability of the scanner. This way the conversion from analogue lightness levels measured towards the digital (step-like) distribution will be using the full scale by "stretching" out the distribution and creating in-between levels with real distinct digital values correspondign to real analog measured values.
The other way is, you tell the scanner to throw away the 0-49 range and the 200-255 range and take the already scanned 50-199 values and map them on a scale of 0-255 leaving uniform distributed gaps between the values i.e. example every 3rd light level will have no pixels associated. So, example, there will be 2000 pixels with light level 120, 2001 pixels with light level 121, 0 pixels with ligth level 122, 2004 pixels with 123, 2005 with 124, 0 pixels with 125 etc etc.
You get a comb-like histogram.
This can show up ugly on large smooth color or lightness transitions such as a blue sky.
Photoshop and SOME scanner SW do this a bit more clever, it can interpolate and change some pixel levels so that the gaps get filled in.
What IDEALLY a scanner SW SHOULD do (but it doesnt - maybe some high end drum scanners do?) is to adjust the illumination of the frame before scanning, based on the pre-scan, so that even iwth a super dark or too light negative, the sensor will be used efficiently i.e. its sensitivity will be covered. To use the full analog scale of the sensor. This would translate to automatically stretching out the histogram to the full 0-255 scale in the correct way.
It's like adjusting the ISO rating on a digital camera sensor versus adjusting the light reaching the sensor by the aperture or shutter time.
I don't know why scanners don't do this.
Highway 61
Revisited
+1 with what Pherdinand wrote.
The only way not to have a comb-like histogram is to generate a TIFF 16bits per channel file with the scanner, then post-process it carefully in PhotoShop.
The 16 bits to 8 bits mode shift (to save the file as Jpeg in case you don't want to store huge files you don't intend to have printed on your hard drive) must be done at the very end of the process.
The only scanner software which I found to be close to PhotoShop was Nikon Scan (unfortunately Nikon ceased to produce anything related to film scanning).
The Nikon Super Coolscan x000 scanners models could do some multi-sampling, which looks like to be several scans with different film frame illumination, and one file produced with all the date gathered during them. HDR scanning. It works pretty well for difficult negatives or slides. Also, there is an "analog gain" function in Nikon Scan which can be used after the preview, but not too sure if it modifies the way the film is illuminated during scanning.
Never used an Imacon so I don't know anything about them nor their softwares.
The only way not to have a comb-like histogram is to generate a TIFF 16bits per channel file with the scanner, then post-process it carefully in PhotoShop.
The 16 bits to 8 bits mode shift (to save the file as Jpeg in case you don't want to store huge files you don't intend to have printed on your hard drive) must be done at the very end of the process.
The only scanner software which I found to be close to PhotoShop was Nikon Scan (unfortunately Nikon ceased to produce anything related to film scanning).
The Nikon Super Coolscan x000 scanners models could do some multi-sampling, which looks like to be several scans with different film frame illumination, and one file produced with all the date gathered during them. HDR scanning. It works pretty well for difficult negatives or slides. Also, there is an "analog gain" function in Nikon Scan which can be used after the preview, but not too sure if it modifies the way the film is illuminated during scanning.
Never used an Imacon so I don't know anything about them nor their softwares.
Ronald M
Veteran
Darkroom is an art that is being lost fast. Without printing experience, how should you know what a negative is supped to look like.
Contrast is best adjusted in photoshop or LR or Gimp or other post processing. Levels and curves need to be applied to transform the scan.
Just as with a darkroom process, find a full tonal range subject, process the neg and scan. Aim to get detailed blacks and whited not blown out. Detailed blacks are put in by exposure only. Extending development will not add to blacks. Time in developer controls the whites. If grey, more time is called for. If you get a featureless white with no detail, less time in developer.
The scan is normally flat and you adjust later. Just get a histogram not falling off either end.
Contrast is best adjusted in photoshop or LR or Gimp or other post processing. Levels and curves need to be applied to transform the scan.
Just as with a darkroom process, find a full tonal range subject, process the neg and scan. Aim to get detailed blacks and whited not blown out. Detailed blacks are put in by exposure only. Extending development will not add to blacks. Time in developer controls the whites. If grey, more time is called for. If you get a featureless white with no detail, less time in developer.
The scan is normally flat and you adjust later. Just get a histogram not falling off either end.
Ranchu
Veteran
My epson v500 is scanning in 12 bit per channel when set to that, whether you output jpegs or whatever, so no combing of the histogram. I'd be surprised if a scanner didn't scan that way, but I guess it's possible some don't.
Chriscrawfordphoto
Real Men Shoot Film.
http://crawfordphotoschool.com/digital/scanning.php
This tutorial I wrote on film scanning explains it. You need to do the adjustments in Photoshop, Lightroom, or whatever editing software you prefer. Scan software really isn't suited to editing scans.
This tutorial I wrote on film scanning explains it. You need to do the adjustments in Photoshop, Lightroom, or whatever editing software you prefer. Scan software really isn't suited to editing scans.
Ranchu
Veteran
It's perfectly suited to setting preliminary endpoints, though. The idea that people need to post process and acrhive 16 bit files to get a picture that looks good out of a camera is incorrect. Tri X in Rodinal looks so good it's pretty hard to screw it up.