What is Portra look? My 6x7 scans look like they were shot on digital...

[MattKing]

I thought someone else might have pointed out already, but isn't Scanning and Scanners a better forum for this topic.

It is - Digital Negatives is the sub-forum relating to the preparation of large negatives which are (generally) used for large contact prints using traditional processes like Cyanotypes.
I will alert the moderators, who most likely will move the thread into the proper sub-forum.

 

[David A. Goldfarb]

Thread moved.

 

[trendland]

Thread moved.

Seams realy is a bit too much scanning issues. So it is right to move. ....

with regards

 

[Elix]

Not directly, no, feel free to do so and provide comparisons if you can.

I have tried your action and here's my take with it. I used an additional LAB saturation channel.
And here's a comparison of Adrian's action with ColorPerfect. In ColorPerfect I made sure there is no clipping in the highlights or shadows, then applied levels layer to remove the cast. The process is described in this book, which is free if you have Kindle subscription. Then I used curves to give more contrast. I might've went overboard on this.
I really like your action. 15 Euros well spent, I say

a factor of 2 is a multiple and 1 stop, 8 factor of 8 and is three stops etc.

Uh... Could you rephrase that? 4 is 2 stops, 8 is three stops more than 1? Then how can I calculate +1 stop if my base exposure is 2.5, for example?

The ideal exposure should bring the highlight of the scan (the shadow area) close to the right of the historygram you need to use the raw graph to see this. However keep in mind as long as the densitiest areas of you scan are well away from the left of the history gram you are less likely to notice any difference.

As I remember the raw histogram was far to the left... I'll check tomorrow.

You're fighting for the last 5%, which is cool to see. I'll keep watching and learning.

Thanks! I wouldn't estimate as optimistically as you. I'd say I still have 30% to go at least.

What gives a scan a "film look"? To some extent it is grain, which is not apparent in your image. In other cases it is saturated color such as is seen in Velvia scans. In some cases it dark shadows. In some cases, it is quirky color shifts. Looking at the three images linked in the original post, all three have dark shadows lacking in detail. I suspect that in all three cases, contrast has been enhanced digitally and then exposure has been decreased to preserve highlights and produce darker shadows. Although it is quite possible that the photographers carefully planned their lighting environments to get the results they wanted. The original scan on this thread looks a lot like the results I get with Portra, and the lack of real shadows probably explains why it looks different from the reference images.

So many great insights! I will be doing more testing with the processing and scanning to see if I can close the gap to reference images. Above I've tried to process with contrasty curves, but I can't say I like the result or find it closer to reference images.

I will alert the moderators, who most likely will move the thread into the proper sub-forum.

Thread moved.

Thanks!

I encourage anyone to give your processing workflow a try with this image to see if it could be made to look more like film. You can download linear tiff file here.

 

[Elix]

I'll make a bold claim and say that my method gives better results faster and with less fuss.

I did another comparison of your action with VueScan's output with a different photo. I was quite satisfied with this customized output from VueScan... until now: http://screenshotcomparison.com/comparison/131799

 

[Ted Baker]

Uh... Could you rephrase that? 4 is 2 stops, 8 is three stops more than 1? Then how can I calculate +1 stop if my base exposure is 2.5,

1 is just some normalised value you can imagine it as 1 seconds exposure for each pixel. 2.5 would be 2.5 seconds per pixel. Except you just don't know how long each second is...

Anyway a factor or 2.5 sound about right for color film, but you will need additional exposure for blue and green. Vuescan will do this automatically.

Something like this? I assume the light source is pretty cold?

 

[Adrian Gabor]

I did another comparison of your action with VueScan's output with a different photo. I was quite satisfied with this customized output from VueScan... until now: http://screenshotcomparison.com/comparison/131799

Yes! This example really highlights the strength of this method, it looks very nice!
As you see, what you do in the camera and the quality of the light is 90% of the work, the inversion then brings you the rest of 10% (if done right).

 

[calebarchie]

 

[Elix]

Something like this? I assume the light source is pretty cold?

It was a winter overcast day, 2 PM. Yeah, the color temperature was quite cold. I think I was quite close in my initial rendering. The walls appeared that blue.
Your's a little dark for my taste. Doesn't look like overexposed Porta 400 to me.

View attachment 195435

I like the last one (with least saturation) the best. What is your approach?

 

[calebarchie]

I guess all I am trying to say is: there's many ways to skin a cat..
adrians was on the top btw (tomato) beware of magic bullets..!

 

[trendland]

Interesting procedures ! I would like to state : the real advantage with film is the characteristic of different emulsions.
You may indeed speak about a look of special films in regards of such characteristic wich are different between some different films.
If you shot digital you may have some apps wich simulate special film looks via software.
I remember first apps from the very beginning with digital.
You may simulate a film look of agfa color films 1940. Or Kodachrome emulsions of the same period.
I've seen such "color looks" in different magazines - after a while it is a bit boring.
But if all parameters of a shot came together it definitivly make sence.
By the time you can't buy Kodakchrome films expired 1946...
But the real difference of this software simulated shots is indeed it is not 100% same characteristic as you get with the real film. But I noticed the apps become better and better (first software end of the 90th was nonsence).
To have a special look (in colors ) it is a real task of todays photographers to be different from others. Many tried their best - but nevertheless : Grandma Smith with her computer is also able to work via software in same direction - may be she reached phantastic results but never know about because she did not realize.
So - such manipulation became a mass phanomenia. .......just look at the instagram filters.
The task of a good photographer is to create shots never seen before and to avoid mass characteristics.
That isn't done with color looks - that's a complete conception in photography but color looks are sometimes a part of it.
An assistent (photography) is creating color filter manipulations.
It looks not so bad - so I asked him about the software he is using.
He stated it isn't via software. I was surpriced - and how did he REACH this filter look ?
He did it with color markers (edding type) but not directly onto the lens.
Cling film taped in front of his lenses he used. Well - thats an example of a real smart individual workflow.
And that is the advantage of using film today. It is a special individual workflow not from the mass - 99,2% didn't use film.
But therefore I (personaly) would not care so much about scanning software.
THE more digital manipulation you proceed the less individual (color) look you may create. The more within the mass characteristic your workflow will end. At last it is no wonder you didn't notice a different look of using film in comparison to digital.
with regards
PS : But notice - we just spoke about (color) look ! There is 85% left with photographical methods in general to create better shots. You may also proced in other directions.

 

[Ted Baker]

It was a winter overcast day, 2 PM. Yeah, the color temperature was quite cold. I think I was quite close in my initial rendering. The walls appeared that blue.
Your's a little dark for my taste. Doesn't look like overexposed Porta 400 to me.

I like the last one (with least saturation) the best. What is your approach?

The first one I posted was using the algorithm published by makers of colorperfect, It is easy enough to "reduce the exposure" digitally to make the print lighter.

I did it again using the algorithm used for DPX/cineon scanning (i.e. pretty much anything you saw at the cinema in the last twenty years that was shot on film). They are both similar in that they preserve the tonal relationships (the intensity ratios) that would normally result on a positive print. The DPX approach in particular works this way as it was originally designed create print film for traditional projector as its final product.

None of these have calibrated in any way to give colours, that would match a RA-4 print, a noritsu or fuji frontier. That's a later stage.

Do you have a shot that was taken at box speed of the same scene? and/or a one that was shot say 3 or 4 stops over?

 

[calebarchie]

Ted, you are a few steps behind - need to adjust film gamma curve (density etc) then of course color.

 

[MattKing]

Doesn't look like overexposed Porta 400 to me

Are you referring to optically printed over-exposed Portra 400? That is really the only "standard". Anything else merely reflects the post-processing preferences of the person doing the post processing.

 

[Elix]

I did it again using the algorithm used for DPX/cineon scanning

Can you give me more details on this process? Where can I find more info? I don't quite like this result, tbh, as it's too cold and dark for my taste, but still I'd like to know!

Do you have a shot that was taken at box speed of the same scene? and/or a one that was shot say 3 or 4 stops over?

No, I didn't bracket this shot. I did an extensive test of Portra 400 recently shooting the same scene with a flash and changing exposure. I shot 4 rolls and then processed them at a lab using normal process as well as push +1, +2 and +3. Now all is left is to scan the results and compare. Here's the test scene shot on digital and the workflow spreadsheet.

Should I scan using fixed or auto exposure for comparison to be meaningful? Does anyone have interest in seeing those results? Should I make a new topic for this?

Are you referring to optically printed over-exposed Portra 400? That is really the only "standard". Anything else merely reflects the post-processing preferences of the person doing the post processing.

But can't you also use different papers and filters during optical printing? But still I think you are correct, it all comes down to taste of the person doing the post-processing. I've stumbled upon Dead Link Removed. What is real Portra look among those? I think the following post by jnanian was spot-on regarding the first question of this topic.

people who post stuff on flickr don't post un-tweeked stuff. even if they say they don't they do ... its the way of the internet ...
negative conversion &c its all the same ... just do what you want with the film so you like how it looks, in the darkroom you would have
done the same .. and enjoy yourself ... the water isn't as pure as people suggest ...

It was the final nail in the coffin of my preconception regarding 'Portra look'. All doubts and questions have been cleared (for me) and I think there is nothing more to discuss here except why that photo looks digital (which has already been partly answered). Thanks everyone for your inputs!!!

 

[Cursor Major]

Just posing this to compare my neg invert method against others.

 

[calebarchie]

Can you give me more details on this process? Where can I find more info? I don't quite like this result, tbh, as it's too cold and dark for my taste, but still I'd like to know!

Ted is applying a standard kodak lin2log conversion see here
https://www.kodak.com/ph/en/motion/..._tools_and_techniques/digital_lad/default.htm

 

[Adrian Gabor]

Just posing this to compare my neg invert method against others.

It looks a bit green to me, but I like the way the highlights look.

 

[Elix]

Just posing this to compare my neg invert method against others.

Nice! I like it! It is quite bright and contrasty yet highlights are saved, and the skintone is natural. I can see slight green cast though. What's your processing workflow?

By the way, did anyone try to profile their scanner with Portra?

 

[Cursor Major]

Yes on reflection it did have a green cast. I just pulled a neutral off the shaded rear of the 'white' hat but that resulted in the cast, I've removed that with a slight change to the curves.

My method involves a photoshop action I've created that first runs the neg through colorperfect to do a basic inversion without any clipping to highs and lows, a one click colour balance in colorperfect which I preserve for each roll of film I convert. (I equally loathe and love colorperfect, it's layout and workflow is so obscure and unintuitive, but I've somehow managed to set the settings as I like, and I fear the day I have to try and figure out how I got there)

The action then creates a series of adjustment layers which I use to manually set the colour balance for black point, white point, then midtones. Set the black point, add a standard S curve and bump the saturation to a preset level which I feel compensates for the flat scan.

Sharpening is non destructive and adjustable Surface Blur method.

The process takes me less than a minute per image for a quick pass and fire it back over to my Lightroom catalog. I preserve the base layer as the original negative so I can alway reprocess or adjust if necessary.

 

[Adrian Gabor]

View attachment 195489
My method involves a photoshop action I've created that first runs the neg through colorperfect to do a basic inversion without any clipping to highs and lows, a one click colour balance in colorperfect which I preserve for each roll of film I convert. (I equally loathe and love colorperfect, it's layout and workflow is so obscure and unintuitive, but I've somehow managed to set the settings as I like, and I fear the day I have to try and figure out how I got there)
View attachment 195491

I suspected as much, CP does indeed do one thing very well, and that thing is highlight detail.

 

[Ted Baker]

Can you give me more details on this process? Where can I find more info? I don't quite like this result, tbh, as it's too cold and dark for my taste, but still I'd like to know!

I didn't make any colour corrections to any thing I posted, I just took a stab at the exposure. I just assumed it was lit by a cold blueish light. DPX is an industry standard for scanning film cinema film, it maintains the non-linear encoding of the negative so that the image can be printed using a film recorder on a normal print film which is also non-linear. In theory a print made this way should look identical to one made optically, with the additional advantage of being able to digitally manipulated. What you can also do is convert the non-linear film scan into a linear colorspace, then into colorspace of the print (for cinema it would be P3 for us on photorio it would sRGB), if you saw the latest star wars on a digital projector then I just described the process roughly.

What I haven't done, in my two examples and BTW none of the examples have done either is translated the colors into the colorspace of an actual print, i.e. make an adjustment for the actual dyes used in a print.

No, I didn't bracket this shot. I did an extensive test of Portra 400 recently shooting the same scene with a flash and changing exposure. I shot 4 rolls and then processed them at a lab using normal process as well as push +1, +2 and +3. Now all is left is to scan the results and compare. Here's the test scene shot on digital and the workflow spreadsheet.

I am doing something similar, but doing all the testing etc takes time... I am basically writing a set of tools open source scanning tools for photographers who want to keep using film. I am also keen to use different samples!

Should I scan using fixed or auto exposure for comparison to be meaningful? Does anyone have interest in seeing those results? Should I make a new topic for this?

I would recommend using the auto exposure for a negative, use the rebate to set your exposure and keep it constant for all scans. That will eliminate a few variables.

But can't you also use different papers and filters during optical printing? But still I think you are correct, it all comes down to taste of the person doing the post-processing. I've stumbled upon Dead Link Removed. What is real Portra look among those? I think the following post by jnanian was spot-on regarding the first question of this topic.

Even in the glory days, there wasn't that much choice, at least not in the way you process a film digitally. Each brand of paper would have it own set of dyes, and characteristic curve would be slightly different, so each would have its own look. But not quite the same options as digital unless you use other techniques.

 

[Elix]

The process takes me less than a minute per image for a quick pass and fire it back over to my Lightroom catalog. I preserve the base layer as the original negative so I can alway reprocess or adjust if necessary.

Thanks for sharing your workflow!

I didn't make any colour corrections to any thing I posted, I just took a stab at the exposure. I just assumed it was lit by a cold blueish light.

Was that a manual or automatic process? What software did you use?

I am doing something similar, but doing all the testing etc takes time... I am basically writing a set of tools open source scanning tools for photographers who want to keep using film. I am also keen to use different samples!

Wow, consider me interested! After I'm done scanning I will have 40+ scans of the same object but with different exposure and also 0/+1/+2/+3 push processing. I can share if you need it.

I would recommend using the auto exposure for a negative, use the rebate to set your exposure and keep it constant for all scans. That will eliminate a few variables.

I didn't quite understand that. What rebate? Should I lock exposure on a single frame and use that exposure for all photographs? Should I lock on a photo that was shot at base exposure? Then what about film rolls that were pulled? Then wouldn't many scans be grossly over- or underexposed? Doesn't that go against the workflow which most of us do (compensate for over- or underexposure when scanning)?

 

[Ted Baker]

Was that a manual or automatic process? What software did you use?

Automated, It's my own software for the inversion I have a couple of algorithms I am experimenting with, I use Rawtherepee do adjust the exposure, I did not make any curves or color corrections, that's your job...

Wow, consider me interested! After I'm done scanning I will have 40+ scans of the same object but with different exposure and also 0/+1/+2/+3 push processing. I can share if you need it.

Would be happy with copy of each, what you should consider doing is scan with Noritisu or frontier, then you could make a CLUT, to get the same/similar result... That's what I will eventually get around too.

I didn't quite understand that. What rebate? Should I lock exposure on a single frame and use that exposure for all photographs? Should I lock on a photo that was shot at base exposure? Then what about film rolls that were pulled? Then wouldn't many scans be grossly over- or underexposed? Doesn't that go against the workflow which most of us do (compensate for over- or underexposure when scanning)?

We are talking about raw scanning. When you do a preview in vuescan it uses the least dense area to determine exposure. So if your preview includes a section without film, you will find the RGB exposure is 1 or close to it, and you will see the orange mask in your scan. By the rebate I meant an area of unexposed film. If you do the preview on just that section it will adjust the exposure time of all three channels so that rebate is reasonable close to white. i.e. your histogram will be close enough to right, and the most dense sections of the negative will be far enough away from the other end of the scale. i,e, it will increase the exposure as much as possible without blowing out the least dense areas of the scan.

I would do it this way for each roll of film, the base will be more dense on the pushed film. To eliminate as many variables as possible. Most of this is hidden, i.e. it is either all done automatically, and not many are aware of it, or the software does not expose these features to the end user. e.g. Epson scan does not allow you manually adjust the exposure. If you negatives end up being very dense, you could consider multiple exposure, but I don't think that is necessary. I have never used your particular scanner BTW.

Hope this helps.

 

[nbagno]

Here's a video I stumbled upon just a few days ago. First of all I have to say that I'm impressed with her go getter attitude. She get's that light over exposed look with film and explains how she meters for different film stock. Take it as another data point...

Read a little more about this young lady, pretty impressive resume.