Lend me your eyes - working on a Tri-x simulation.

[Michael Floyd]

You have done quite a good job of achieving your objective. I have been down this road and know how difficult it is. And to automate this has got to be quite a stretch.

I decided to use a different path. Took a Sony A6300 and had the CFA scrapped off which turned it into a monochrome sensor. Because the camera is not aware that it is monochrome, I have to shoot RAW since JPEG would do a demosaicing of the JPEG file and it would lose detail I don't want to lose. I then convert the RAW into a DNG file with an app called Monochrome2DNG which gives me a DNG file to work with that has avoided demosaicing. And the mono file is so easy to work with. In fact, it looks good even if I don't do anything to it.

If I had to do it again, I would pick a full frame camera such as the Sony A7ii with can use non-lossy compressed RAW since I am finding that the A6300 only has the option for lossy compressed RAW.

Here is an example. Used a red filter. Camera is full spectrum, and without a UV_IR cut filter some IR feeds through the red filter turning leaves white. All the filters available to TriX can be used with this camera.

Looking at my camera usage for the past 2 years since having this camera modified, I find that it is now my most used camera. Having fun with it.

Interesting, thanks. Yeah, I am indeed curious about a dedicated mono camera. I haven't looked into it at all though really at this stage.

For your interest: RawTherapee gives a lot more control over the demosaicing than any other program that I have used, including a straight thorough option - ie no demosaicing.

 

[Michael Floyd]

Why Tri-X simulation toned? It's pure B&W

Also, there are some differences in Color response and Dynamic Range among TRI-X 320TXP, TRI-X Pan 400, TRI-X Ortho 4163, and TRI-X 400 Professional.

You may check for reference several Tri-X simulations with different filters for use in Lightroom / ACR Eastman Kodak Black and White Films profiles or LUTs.

Hi, thanks for the reply.

Toned because that is how the scans arrived to me. I'm aware that this is simply a choice made by whoever did the scans. Conveniently I quite like it and decided to leave it as is for the film scans and replicate it for the digital images. Of course it is just one click to go back to straight monochrome.

I would not pretend to be able to recognise or quantify any differences in the different kinds of Tri-x. Of course the 'look' of any film can be influenced significantly in the darkroom as well. I reckon a proper job would be to do a well done, old school, darkroom print - and then scan that for use as the baseline for a simulation. Maybe I'll do that one day, for now I'm comfortable with the hybrid approach that I'm taking.

I have looked at a number of Tri-x LUT's and profiles, there has been some good learning in them for me. Your tip is appreciated.

 

[Michael Floyd]

Close!

Thanks!

The film ones are a bit lighter in the highlights, so a curves adjustment that makes the highlights clip just a little bit an pull up the light mid-end should do the trick.

If only it were that simple. I do see what you are saying though and have indeed done exactly as you suggest. I can get one area right but another is then more wrong.

I feel like what it needs is something that works a bit like a 'shadows' adjustment, but targeting other luminosity ranges. That is, a tool that chooses areas by overall darkness or lightness and then raises or lowers all tones in the area on the basis of the whole area's tone. If that makes sense.

Also, there is a bit more grain in the film than in the digital version. If you use Photoshop to edit, you could add some grain, and using blend-if target that to just the midtones.

Yes, agreed. I don't use photoshop but I do use GIMP and can do as you suggest there. I'm aiming for a Capture One profile and at the moment want to keep it all in the one program.

In general, film is less sharp than digital, so go easy on the sharpening, or turn it off completely.

All sharpness settings are at minimum, I can knock it down more with a layer that pulls 'structure' down again. Layers don't save in profiles so I'm leaving it as is. I definitely take your point though, thanks.

If you are using photoshop, and your changes are adjustment layer based, you could turn the stack into a 3D-LUT, and export that to other programs like Lightroom to make it a one-click adjustment.

I can make LUT's in RawTherapee and have used them previously. Capture One doesn't use them at all and I actually prefer to use a profile than a LUT because it brings up all the settings visibly rather that importing the 'look' while maintaining default settings. I definitely want to keep all editing in the one program too.

Cheers!

 

[Michael Floyd]

Thank you for the raw samples and scan.

Wow that was hard but got pretty close i think. I'd need more raws and original photos of various colour scenes to get closer.



Capture one grain is not bad but it seems to apply uniform across the photos where as the scan seems to have less grain in the shadows and more in the highlights which i cant fix in capture one without layers and stuff. Maybe nik effects or some other tools would apply better grain.

Nice one! I'm glad you've had a go.

I see similar things in yours that I see in mine. Mostly too much contrast through the middle tones on an image wide scale (ie the opposite of locally).

I have to admit that I find the result subjectively preferable in some respects, this doesn't mean I'll slack up with the attempt though.

How do the other images look when you apply exactly the same settings?

 

[Michael Floyd]

The scanner used would make a difference in the quality Tri-X images.

Of course. I'm just working with what I have available.

Cheers.

 

[maltmoose]

Thank you for the raw samples and scan.

Wow that was hard but got pretty close i think. I'd need more raws and original photos of various colour scenes to get closer.

Capture one grain is not bad but it seems to apply uniform across the photos where as the scan seems to have less grain in the shadows and more in the highlights which i cant fix in capture one without layers and stuff. Maybe nik effects or some other tools would apply better grain.

Nice one! I'm glad you've had a go.

I see similar things in yours that I see in mine. Mostly too much contrast through the middle tones on an image wide scale (ie the opposite of locally).

I have to admit that I find the result subjectively preferable in some respects, this doesn't mean I'll slack up with the attempt though.

How do the other images look when you apply exactly the same settings?

This has been an interesting processing challenge, pushing how to control settings in different ways. I never thought reducing sharpness could be so difficult...I can see my toning is off but to be honest i would reduce that effect on my images.

Ive avoided the use of layers but some form of variable tonal contrast would help, ill have a look with the mid tones and some tweak ideas.

The other photos look ok with my preset but 1 of your cliff photos needs white balance fixing first. Then another just needs some exposure adjustment but suprisingly it seems to work ok on the images.

 

[Michael Floyd]

Ahh yes, I sometimes shoot with uniWB, so of course that can be brought into line with the others. And exposure adjustments I don't count as part of a profile either so tweak that as required. Sorry I should have mentioned.

And make them both pure greyscale if that is more to your taste, that would be interesting too I think.

 

[Windeguy]

Maybe try shooting an IT8 with Tri-X and the XPro. This will show you the difference in color response.

 

[Joachim B]

Doesn't C1 have a curves tool?

 

[JimH123]

You have done quite a good job of achieving your objective. I have been down this road and know how difficult it is. And to automate this has got to be quite a stretch.

I decided to use a different path. Took a Sony A6300 and had the CFA scrapped off which turned it into a monochrome sensor. Because the camera is not aware that it is monochrome, I have to shoot RAW since JPEG would do a demosaicing of the JPEG file and it would lose detail I don't want to lose. I then convert the RAW into a DNG file with an app called Monochrome2DNG which gives me a DNG file to work with that has avoided demosaicing. And the mono file is so easy to work with. In fact, it looks good even if I don't do anything to it.

If I had to do it again, I would pick a full frame camera such as the Sony A7ii with can use non-lossy compressed RAW since I am finding that the A6300 only has the option for lossy compressed RAW.

Here is an example. Used a red filter. Camera is full spectrum, and without a UV_IR cut filter some IR feeds through the red filter turning leaves white. All the filters available to TriX can be used with this camera.

Looking at my camera usage for the past 2 years since having this camera modified, I find that it is now my most used camera. Having fun with it.

Interesting, thanks. Yeah, I am indeed curious about a dedicated mono camera. I haven't looked into it at all though really at this stage.

For your interest: RawTherapee gives a lot more control over the demosaicing than any other program that I have used, including a straight thorough option - ie no demosaicing.

Yes, I have RawTherapee , and I have experimented with the different demosaicing options.

But the app Monochrome2DNG is such an easy route to go and can batch process all the files: Monochrome2DNG: the Missing Application for Сolor Сameras Сonverted to Black and White | FastRawViewer

Decided to get the app ON1 Nonoise also since it works well with the generated DNG files. Can't use DxO on the RAWs since it would Demosaic the image. And Topaz AI products won't handle the mono DNG files. They will it I convert them to RGB, and then back to grayscale afterwards.

I used Monochrome conversion of Sony mirrorless cameras (monochromeimaging.com)

to do the conversion. This guy specializes in Sony mirrorless cameras with APS-C for $900 and full frame for $1000. About half the price of maxmax.com. And as I mentioned previously, I would opt for full frame if I could do it over again. The A7ii with its non-compressed RAW would be perfect.

The images you had shown were not exactly Black & White. There was a subtle beige color to the images. I used Topaz AI and gave this image a somewhat similar look compared to the original harsher black & white. Maybe I didn't exactly match, but I moved into the general direction. There is a setting to add grain to, but I didn't add any.

 

[Vernon D Rainwater]

Normal Tri-X had an "ash type grain" that is unique. If I was experimenting with the idea of simulating Tri-X exposures and especially the grain, I believe I would use a roll of tri-X and expose a few exposures of a Gray Card for example at different exposures and use the scanned images from these exposures as overlays to hopefully simulate better Tri-X Grain appearances with other digital images.

The illustrated images have basically NO Grain results similar to that in Tri-X.

--
Vernon...

 

[Michael Floyd]

Maybe try shooting an IT8 with Tri-X and the XPro. This will show you the difference in color response.

Definitely a great idea and something I'd like to do sometime. Cheers.

 

[Michael Floyd]

Doesn't C1 have a curves tool?

For sure it does. It is simple to adjust the curve of course - however despite how it seems looking at the images, solving the mids proves to be more complicated than that, Unfortunately.

The raws are in the dropbox I linked if you're keen to have a go.

 

[Michael Floyd]

Yes, I have RawTherapee , and I have experimented with the different demosaicing options.

But the app Monochrome2DNG is such an easy route to go and can batch process all the files: Monochrome2DNG: the Missing Application for Сolor Сameras Сonverted to Black and White | FastRawViewer

Decided to get the app ON1 Nonoise also since it works well with the generated DNG files. Can't use DxO on the RAWs since it would Demosaic the image. And Topaz AI products won't handle the mono DNG files. They will it I convert them to RGB, and then back to grayscale afterwards.

Got it. Sounds like Monochrome2DNG is working well for you. I have no experience with it or ON1 or any of the AI Noise reduction programs.

 

[Michael Floyd]

Normal Tri-X had an "ash type grain" that is unique. If I was experimenting with the idea of simulating Tri-X exposures and especially the grain, I believe I would use a roll of tri-X and expose a few exposures of a Gray Card for example at different exposures and use the scanned images from these exposures as overlays to hopefully simulate better Tri-X Grain appearances with other digital images.

The illustrated images have basically NO Grain results similar to that in Tri-X.

Thanks Vernon, I appreciate your input. I do agree too, the grain character is way off.

I have actually tried as you suggest, basically sampling the grain with a neutral background - I've also managed to access a couple of other people's attempts at the same. It was a while ago though and my skills have moved on a bit since then so it's probably something that I should revisit at some stage.

Cheers.

 

[maltmoose]

Doesn't C1 have a curves tool?

For sure it does. It is simple to adjust the curve of course - however despite how it seems looking at the images, solving the mids proves to be more complicated than that, Unfortunately.

The raws are in the dropbox I linked if you're keen to have a go.

Yep, i used the linear mode then also added a luma and rgb curve to even get close. Along with other tweaks. No layers.

 

[Iliah Borg]

Yes, I know - nothing ever looks quite like film. I truly accept that.

Nevertheless, I'm giving it my best shot. I'm working on a Capture One profile for my X-pro3 images that hits the overall feeling and tone of Tri-x as well as I can. At this stage I've been looking for too long and could use some input. Thoughts, ideas, feedback, something that you're seeing stand out that I might be blind to, etc...

I shot a roll of Tri-x using a 50mm lens throughout, and for every frame I made a near identical image with a 35 on my X-pro3. ISO/aperture and SS are set the same on both cameras in each case so depth of field varies a little. The digital images have exif, the film ones don't.

Clearly there are differences in my image pairs. I can tweak one pair to be more similar however another pair then becomes less so, with the same settings. So again, I'm after overall feeling and ideally some sense of believability, or at least a minimum of falsity.

Here's the image pairs. What do you see?

-

Oh, I don't intend for this to be a test, or a commentary regarding which is best, or even a preference either way.

Shoot Q13, print it, measure the print, make a curve.

--

About LibRaw | LibRaw

www.libraw.org