Hi,
Is there a straighforward way to multiply width by 1.33 somewhere that will desqeeze the anamorphic image?
Is this a function of lens correction? I don’t think it would be as other processors just multiply the width.
Would it make sense to add this Sirui 35mm Lens to the noise profiles if I provided the images (I understand this is mainly a video lens)?
You can do this in gimp by just changing the image width.
I don’t think there is an easy way to do this in darktable. You might be able to do it in lensfun.
Unfortunately you can’t yet transform aspect in Darktable.
You may find a formula with Hugin and use that in the lensfun XML. It’s fiddly and takes a while. I went that way with a 15mm lens with heavy distortion and got a very good result.
Anamorphic lenses are mainly, if not only, video lenses. I wouldn’t use it for pix.
Thank you all for your responses.
So gimp seems to be the solution for now (use the scale , unlink width height and multiply width with the squeeze factor).
I think I could do it in Da Vinci too, import the image in and stretch it using the usual way and even edit it there 
Yes, anamorphic lenses are for video, simply trying it out to get some flares in the image.
Hugin, have not played around with it, good to know its possible there too.
Wouldn’t lensfun and specific lens be the kind of approach to go?
That’s basically what @st.raw and @chhil were talking about: use Hugin to calculate the parameters to use in Lensfun, for the specific anamorphic lens.
Lensfun is more or less just a database (in XML format) with data of various lenses. If your lens is not in it, you can try your luck with a similar lens. Or try to create some XML data yourself with Hugin. This said, I am not sure if Hugin has a model for ‘anamorphic’ lenses, they have a rather specific distortion.
Somebody mentioned DaVinci (they have a totally free version), they are more video oriented, but have some image tools. Maybe they can do something (DaVinci needs 16Mb RAM, or won’t start)
If the distortion is a simple linear multiplier, ImageMagick could do it:
magick in.tiff -resize %[fx:w*3/4]x%[fx:h]! out.tiff
If you (or the software you use) uses libraw, you can set the libraw_image_sizes_t structure element pixel_aspect to a floating-point number, and I expect that would work, but I haven’t tried it.