That IS amazing!

I know nothing about it, but if that’s what the result was from submitting a black-and-white photo, I’m impressed. The resolution is quite sharp.

Elmer, in addition to colorizing, it improves the images in other ways. That said, not all of the changes are improvements. I've deleted some of the results because the faces became unrecognizable. Most of my dad's photos were sharp, with good lighting, but the most dramatic example I have of Ai improvements is in this photo of a car. I think it's from the 1920s. I can't imagine what message the greyscale gave Ai about the color of the car though!

Maybe what happened is that the software outsmarted itself.

This appears to be a car from the 19-teens or 20s. Cars of that era were almost always dark. I don't know if this is a Ford or not but old Henry used to say "Customers can get cars in any color they want, so long as they want black".

The car in the image was a light shade of course and that's unusual, so maybe the software decided that the image tone was wrong.

Maybe my speculation is way off? No matter, what's done by the software is something to behold.

I hadn't thought of that, Elmer, but it makes sense. And I do like the blue.

I like when AI is put to good use. Thanks for sharing. .

These are amazing pictures. Maybe AI could identify the car itself and would know what colors would be possible? Any way, it is totally amazing and scary (at least to me) to see this.

I'm just happy to ”see” you, Alisande! It’s been quite a long time and I hope you're on the mend. your photos are beautiful and you’re def keeping up with the times! Much better than I.

It's cool technology but my personal taste runs to keeping old photos in black and white. It's like so much history is lost with the colorization and removal of percieved flaws. Like a remake of a classic movie.

I remember old cars of that era as sometimes being maize or yellowish, even a pinkish beige or tan. Knew people who restored them in my teens, and even had a a couple of elders in the neighbor still driving them when I was a child. Can't be certain but the curvature of the hood and setting of the headlights make it look like maybe a Ford Model A. I'm not car crazy like I used to be when I was in my teens.

Great idea, Murraysmom! The car, it turns out, is a 1915 Buick Roadster. It's most popular color was maroon, but it also came in dark blue and green. Sounds as though Grok knew this, and chose the blue.

My dad was only 12 in 1915, so either he didn't take the picture or he used one of the rather explosive, overkill flash bulbs of the day, and it obliterated the original color of the car when it picked up the shine.

This also gave me a hint about the identity of the woman driving the car. My dad's cousin Edith, who bears a resemblance, would have been around the right age. The car had 37 horsepower, by the way!

Holy cow!! That is great. It's really neat to have both the original and a lifelike copy of that photo. I, too, am happy to see you back here, sharing your wonderful photos. Maybe AI can be used for good after all. It's a very cool car and even better that you probably identified the driver as family. Thanks for the update.

It reminds me of my mom (born in 1910) telling me about how she and her friends would ride in the rumble seat of some boy's car and how they were always laughing so hard at the getting in and then getting back out. My mom laughed a lot in her life and I treasure that about her.

Wildchild, it's ironic that I'm having so much fun with these photos because before this I was not only uninterested in colorization but I was opposed to it. I still love B&W, and there are some images I would never colorize. Also, I want to stress that the new versions do not replace my old ones. I keep both. My dad was an avid photographer, and I have many albums filled with enlargements (8x10 and sometimes larger) of his B&W photographs. The colorized versions are smaller files, and I don't yet know how well they'll print. I have a program that adds pixels to solve this problem, so we'll see . I think prints are our best hope of saving family photos for future generations.

I took this cover photo--an example of an image that should remain in B&W.

https://www.amazon.com/Calculus-Imaginaries-Gerard-Grealish/dp/1630450766

Sjerin, thank you for the welcome! I tried for several weeks to get on the KT, but it didn't work. Houzz said they were sending me links to my email, but those never appeared. I was beginning to think I'd have to start from scratch with a new account, but then the door unexpectedly opened yesterday.

" It's cool technology but my personal taste runs to keeping old photos in black and white. "

Yes, but.

She still has the images in black and white, nothing is destroyed or discarded by colorizing them.

I too appreciate the impact that special black and white images can convey. Look no further than the work of Ansel Adams for examples. The drama of light and dark, texture and grain, can't be replicated in color. (Remember though that many of those effects are greatly enhanced by darkroom manipulation. So it's not like many of the scenes are 1 in a million situations. They were created in part by the primitive equivalents of today's Photoshop.)

Home photos, even by skilled amateurs, are typically not taken to compete with artistic works. Life is in color and a colorization does add more easily understood realism.

Acclaimed filmmaker Peter Jackson collaborated with the Imperial War Museum in the 20-teens to make a documentary about World War 1. His team took old archival film footage, subjected it to computer processing to colorize it and normalize frame rates to make motion more correct, and produced "They Shall Not Grow Old". It was a terrible war with terrible suffering - the processed and colorized scenes made the horrors come alive compared to the black and white jumpy films it was made from.

If you haven't seen it, I recommend it.

Taking an opposite approach, I removed the color from my photo on the book cover. There wasn't a lot of it, but a couple of blue streaks in the sky detracted from the sense of foreboding I was looking for.

Here's another I chose to colorize. When I saw it I said Ai must have decided to dress them in jewel tones. But I guess they must have decided that themselves. Incidentally, Google Lens identified the women as sisters from Australia. Um, wrong. The only one I recognize is the lady with the white-plumed hat, and she was an American opera singer.

And of course I couldn't leave myself out of my big project. 😍

Alisande - I went down the same path last night, trying to figure out the car and what colors it came in. Are you sure it’s a Buick, based on known family history, or is that what AI is telling you? I ask because my searches weren't conclusive. Chevy, Cadillac, Maxwell, Stoddard-Dayton, Ford, Buick, were among the guesses between Google and ChatGPT. But they were acknowledged as just guesses, due to image quality, lack of visible badge, and the fact that those cars of that error looked so similar.

Food, I'm as sure as I can be, given that the 195 Buick Roadster was the only possibility given to me. The other manufacturers you mentioned never came up. And when I questioned Google about the colors that model came in I was supplied with a bunch of pictures, all of which looked like the car in my photo. That was conclusive enough for me.

Thanks for trying to learn more. Those rabbit holes are always waiting for us . . .

Does it matter what make of car it is?

“Does it matter what make of car it is?”

Well, the subject line of this thread and some of the discussion of using grayscale to determine color, is about accuracy of the colors chosen. So yes, determining the make and model of the car would then lead to finding out the factory color options, which could then help determine the probability of having chosen a correct color, or why this one was chosen. As you pointed out, it’d be incorrect for many Fords of that era.

My searches, using different tools and searches, resulted in various options so I was curious.

Incidentally, I’ve seen AI be wrong about many things, including car and motorcycles specifically. Most recently a 1950’s Jeep from a family photo. They gave me year and model and I told them why I thought they were wrong - and they agreed. It’s not always easy to get AI to admit mistakes, it’s eerily human like that!

You asked why the model mattered, I answered. It was specific to comments Alisande made, I was continuing that line of discussion. That’s all!

I’ve been using AI for several types of projects recently, and I’d say my impression is that it can be powerful, extremely helpful, sometimes frustrating, and sometimes dead wrong. This interests me. In the case of this car, even prompting AI by saying ”it’s been identified as a 1915 Buick Roadster, do you agree,” they said plausible but would only assign a 50-60% confidence. But no, the Earth won’t stop spinning it it’s not a Buick!

Here's another car, just for fun. The faces are all wrong--if I hadn't seen them in the B&W version, I wouldn't have known who they are. My father's sister is behind the wheel, and my grandmother is in the backseat. Google Lens tells me the vehicle is a 1932 Buick Series 60 Sport Phaeton convertible. I further read that "There were approximately 103 units of the 1932 Buick Series 60 Sport Phaeton (often referred to as the 5-passenger Sport Phaeton, Model 65) built, including those produced for export. This body style is considered one of the rarest 1932 Buick models."

That was a depression year. If the car was relatively new when your relatives owned it, that speaks to their having been quite fortunate and much better off than so many others.

A 103 unit model run is small no matter what but I think US car production fell pretty steeply after the Crash and into the 1930s, compared to previous years.

Awesome ride, Alisande!!! It would be so fun to get to ride in one of those. I love the color!! They look like they are having a lot of fun!

I agree, Murraysmom--that's a great car! I don't k now why they didn't leave it to me. 😀

Elmer, yes, my aunt's husband (I don't think of him as my uncle since I never liked him) was quite prosperous.