Hacker News Reader

Comment by jl6 | Hacker News Reader

The headline before/after image is astonishing, almost in-credible. I can't see how the left image was restored into the right. It looks like there is substantial new detail on the right that I can't see anywhere on the left.

I can only assume that the image on the left is a low resolution scan produced for this web article, and that there must be a much better scan somewhere else.

View on HN

Well that photo must have come from a negative film, which can have an astounding amount of detail, even old film.

So, what improved is probably our digitalization tools and with some post, you can reveal a lot of detail.

It must be. The amount of detail is incredible, and even trying to extract data from the before picture, it doesn't come close to what you see in the newly processed image.

My attempt: https://i.imgur.com/QZDDEB5.png

Of course because you don't have access to the original data.

Imagine:

1. Film -> Method 1 -> Photo #1

2. Film -> Method 2 -> Photo #2

Instead you tried:

3. Photo #1 -> Method 3 -> Photo #2

Which instead gives you a badly edited Photo #1. You don't have the source code, so to speak.

Absolute least they did is to rescan original film with newer type of scanning process/device into higher resolution and bit depth „digital negative”. You cannot replicate that from low quality jpeg image.

I keep seeing a small dog in the helmet.

Looking at it "correctly" the man's image is obvious.

I enjoy image-art for this 'eye of the observer' opportunity.

This is my 30 second attempt with Photoshop: (original on left, mine on right)

https://imgur.com/a/YC2iBHX

It sounds like they took the original film negatives and rescanned them. You can scan them at varying levels of light intensity to get more detail in the highlights/lowlights. By comparison a single compressed digital file is going to have a limited dynamic range.

Reply on news.ycombinator.com