I understand how halftones of photographs work in terms of printing them on paper, your eye/brain understanding them, etc. But I can never seem to find any images/videos of what the actual etching process looked like. People will say, “they exposed an emulsion on a metal plate by shining light through a screen…” What kind of screen? Like a window screen? What thickness of wire are we talking about? What was the emulsion? Did the exposed/unexposed parts then protect the underlying metal while it was etched by an acid? Etc. etc. etc. How were the plates made? I’d love to see photos/videos of the whole setup.
I don’t know if this answers your question, but I thought it was interesting so I went and found this at britannica .com. If you scroll down to the section about the benday process, thats where the discussion about how metal plates are etched really starts going into detail.
The wiki on Half tone printing has some information.
Here is one you can buy:
https://www.gwjcompany.com/index.php?main_page=product_info&products_id=190
As the above links show, there were lots of ways of making a halftone plate. Photo resist was at the core of most of them.
What seems to be missing from all of the descriptions is a core question. How did the dots get to be of different sizes? Therein lies some of the magic of photography.
A photosensitive emulsion behaves in a somewhat counterintuitive manner. At least relative to a simple idea of how they work. The transfer function from light intensity to developed morphology isn’t linear. When we want to create a resist layer the resist needs to be either present at full depth or not at all. Photographically we would call this a very high contrast function, where the curve of light impinging on the resist to amount of resist has as near a step function as we can get. Too little light and nothing happens. Reach the threshold and it fully converts. More light than that and nothing else happens.
If you intersperse a screen with lots of tiny holes in it between a projection of the image and the photosensitive resist the image on the resist is a lot of dots. Choose the hole opening size and the distance from the resist along with the optical geometry, and each dot projected on the resist has a blurred shape with a smooth transition from bright in the middle to dim at the edge.
If the image is bright where a dot is, there is lots of light in the dot, and most of the extent of the dot will receive enough light to convert to retaining the resist. If the image is dark in the region covered, only the centre of the dot will receive enough light to convert. So the final dot only has a small diameter.
Thus the image brightness is directly converted to dot diameter. The above yields a negative image, but depending upon your plate production process, choice of resist chemistry and so on it is easy to get what you need.
The wiki page above also notes that a pre-exposed plate could have dots already present. This is a neat trick. If the pre-exposed dots have a tapering profile (like the dots emanating from the dot screen) they would serve the same purpose. They would be exposed so that if only a small amount of light hit the dot only the middle part of the dot would sum to enough exposure to be retained. More light and more of the dot would jump over the threshold.