The 2d Art Thread

[J. Wilhelm]

I made a thing.  Do you like my creation?

Snip

Very nice! Done with AI?  But "Twin Cáits" is 11 MB in size, and your server are knottycelt.com took over 5 minutes to serve the .png image with an average cable internet connection!

I took a screenshot of the picture with my cell phone and I'm embedding it here (Brassgoggles is the server, click on it to expand) for whoever has a hard time downloading it:

"Twin Cáits" by Sir Farthington-Smythe

[Sir Farthington-Smythe]

Very nice! Done with AI?  But "Twin Cáits" is 11 MB in size, and your server are knottycelt.com took over 5 minutes to serve the .png image with an average cable internet connection!

I took a screenshot of the picture with my cell phone and I'm embedding it here (Brassgoggles is the server, click on it to expand) for whoever has a hard time downloading it:

Thanks for that.  I forgot that I had the full print quality version there.  I have replaced it with a JPG version, which may take less time to load.  But I could resize it down a touch, too.

It was a combination of a couple of AIs, my own 3D models and textures, and some post processing touch ups.  I suppose it could be considered digital mixed media?

[J. Wilhelm]

Very nice! Done with AI?  But "Twin Cáits" is 11 MB in size, and your server are knottycelt.com took over 5 minutes to serve the .png image with an average cable internet connection!

I took a screenshot of the picture with my cell phone and I'm embedding it here (Brassgoggles is the server, click on it to expand) for whoever has a hard time downloading it:

Thanks for that.  I forgot that I had the full print quality version there.  I have replaced it with a JPG version, which may take less time to load.  But I could resize it down a touch, too.

It was a combination of a couple of AIs, my own 3D models and textures, and some post processing touch ups.  I suppose it could be considered digital mixed media?

Nomenclature-wise it makes sense. There's a whole world of digital art jargon that I ignore.

[von Corax]

"Modern Mordor"
Ominous overcast morning over Downtown Austin. Made me stop on my tracks to get the photo. The image is in full color. It hasn't been modified in any way save for a bit of sharpening.

"Modern Mordor"

This is exactly the type of photo I had in mind for the Photography board.

[SeVeNeVeS]

"Modern Mordor"
Ominous overcast morning over Downtown Austin. Made me stop on my tracks to get the photo. The image is in full color. It hasn't been modified in any way save for a bit of sharpening.

"Modern Mordor"

This is exactly the type of photo I had in mind for the Photography board.

maybe move it?

[J. Wilhelm]

"Modern Mordor"
Ominous overcast morning over Downtown Austin. Made me stop on my tracks to get the photo. The image is in full color. It hasn't been modified in any way save for a bit of sharpening.

"Modern Mordor"
Snip

This is exactly the type of photo I had in mind for the Photography board.

See? That's the thing. I started posting on this thread before the photography section was created. But we need an appropriate title for the thread over on the photography section. Should I move them?

Without trying to corner myself too much, I need a title. If I say "Urban Photography" I think that'd turn into a very lonely thread, unless others are contributing by running around their city taking snapshots (I can afford to do this because I spend many hours commuting every day). Perhaps call it "Contemporary Photography"?

[SeVeNeVeS]

"Modern Mordor"
Ominous overcast morning over Downtown Austin. Made me stop on my tracks to get the photo. The image is in full color. It hasn't been modified in any way save for a bit of sharpening.

"Modern Mordor"
Snip

This is exactly the type of photo I had in mind for the Photography board.

See? That's the thing. I started posting on this thread before the photography section was created. But we need an appropriate title for the thread over on the photography section. Should I move them?

Without trying to corner myself too much, I need a title. If I say "Urban Photography" I think that'd turn into a very lonely thread, unless others are contributing by running around their city taking snapshots (I can afford to do this because I spend many hours commuting every day). Perhaps call it "Contemporary Photography"?

May I suggest basic "Landscapes" others could include such titles as "portraits" "Animals and wildlife" and "other".......... just a thought

[J. Wilhelm]

"Modern Mordor"
Ominous overcast morning over Downtown Austin. Made me stop on my tracks to get the photo. The image is in full color. It hasn't been modified in any way save for a bit of sharpening.

"Modern Mordor"
Snip

This is exactly the type of photo I had in mind for the Photography board.

See? That's the thing. I started posting on this thread before the photography section was created. But we need an appropriate title for the thread over on the photography section. Should I move them?

Without trying to corner myself too much, I need a title. If I say "Urban Photography" I think that'd turn into a very lonely thread, unless others are contributing by running around their city taking snapshots (I can afford to do this because I spend many hours commuting every day). Perhaps call it "Contemporary Photography"?

May I suggest basic "Landscapes" others could include such titles as "portraits" "Animals and wildlife" and "other".......... just a thought

What I fear is that we don't yet have the density of photographers to do that.  I'd suggest instead to let the thread grow so we can see what kind of contributions we get. By setting a generic photo stream, we can let people get inspired on their own.

[MWBailey]

My first in-person music festival donation since the Initial Pandemic. Acrylics on stretched canvas, sealed with Krylon Kamar varnish (canned spray), entitled, "Over the Waterfall," in keeping with my personal tradition (if you will) of a musical theme for such paintings (Over the Waterfall is an Old Time tune much beloved by the Old Time community, especially hammered dulcimer players, or so it seems to me at least).

Over the Waterfall

[/size][/color]


In retrospect, I think I should have paid more attention to the coloratura, since one of the drawbacks of acrylics is that the color gets darker as it dries, which is probably why the brighter greens ran to "grass-like" hues in the end. It still garnered a lot of money in the form of "raffle" ticket sales, though, which contributed greatly to the festival breaking even. Which is the whole point of donating to the doorprize/raffle event.

[SeVeNeVeS]

AI art creation.

My Brother sent me a link to a site that creates AI art.

https://www.patience.ai/

I quickly played around with https://www.patience.ai/models/deliberate

Typed in Steampunk woman brown leather/ motorcycle and it ccreated these..........

Spoiler: ShowHide

There are 30 different sites, so if you need to waste some time, have a play.

[J. Wilhelm]

AI art creation.

My Brother sent me a link to a site that creates AI art.

https://www.patience.ai/

I quickly played around with https://www.patience.ai/models/deliberate

Typed in Steampunk woman brown leather/ motorcycle and it ccreated these..........

Spoiler: ShowHide

There are 30 different sites, so if you need to waste some time, have a play.

Interesting results.  I gave it a whirl using the default "Deliberate" model with  45 steps refinement. I used the prompt:

"Steampunk Male Elf, officer in United States Union Army during the Civil War"

Didn't get the period uniform correct, but it's more appropriate for the Mexican American War period. It's actually pretty good. I don't know that it's good enough to create book illustrations, but seems to me it can help create concept art.

In the Valkyrie and the Eagle, I have Officer Bahlmann separated from his^† love interest, who is left behind in Central Mexico, likely stationed in the City of Toluca, near the Nevado de Toluca volcano, whose caldera is being used as training grounds for Austrian and French invading forces. So the lovers are on opposite sides of the war unable to see each other*

*

Off-Topic: ShowHide

The Union, comprised of the US, Great Britain, and Prussia vs. the Confederate States with France and Austria, attacking from the 2nd Mexican Empire

Although Puebla is in the opposite side of the 80 mile diameter Valley of Mexico, I think it'd be fun to have the female^† Engelfolk wear a "China Poblana" dress, a typical Mexican style of dress actually brought by immigrants from India to the City of Puebla in the 19th. C. So I used the prompt:

"Blonde Elf Maiden wearing a Mexican dress"

†

Off-Topic: ShowHide

 Although I explicitly state that Engelfolk are Third Gendered people, binary pairings and gender roles are commonplace among the Engelfolk. There's no restriction on assuming any gender role, it's just that it's not settled during puberty at a given age. A genetic characteristic of the Engelfolk is a delayed puberty which only starts when a pairing between two individuals happens. This gives the Engelfolk their androgynous look and seeming endless elf-like youth.

Source image for dress:

Off-Topic: ShowHide

[J. Wilhelm]

Here's another one. I'm slowly getting the hang of refining the images if you can even call that a "skill"

"Hallucinations due to lack of oxygen aboard a high flying airship were not uncommon, however; these tended to occur particularly after reaching 30 kft of altitude. Improperly worn respirators would create certain illusions among the Engelfolk crews. One of the most common was having visions of Valkyries, from ancient Germanic lore."

Created in Stable Diffusion 2.0

Valkyrie Hallucination 3


Valkyrie Hallucination 4

[SeVeNeVeS]

Good stuff JW.

Its all a bit hit and miss, depending on the prompts used.

Here's another site to try for AI
https://dezgo.com/
I haven't had a chance to play yet.

Some tips from Big Bro' for beginners giving it a go for the first time.

Spoiler: ShowHide

Select "text to Image"

There are a few selections to make here:

1. "Prompt" - this is where you describe what you want to see
2. "Model" - that seems to be the AI model that renders the pics. There are several. Play with them, but I have found best results with "Epic Diffusion 1.1 (general)"
3. More options:
a. "Resolution" - shape of the picture (landscape / portrait etc.)
b. "Negative Prompt" - things you DON'T want to see in the picture
c. "Guidance" - How much freedom you want AI to have over your input. 7 seems about right, but experiment
d. "Sampler" - as far as I can tell this is the data set that the AI learned from. There are a few Anime options if that is your thing. For photo realism, I find "Euler Ancestral" works best, but again, play around with the settings
e. "Seed" - each generated picture has a unique seed number, so if you are trying to get the same or similar results you can enter a specific seed number to work from. Leave it blank for now. You can also use "-1" to keep making reiterations to the existing image.

When you have made your selection, just hit "Run"

Once you have a pic you like, the site has an "Upscale" selection to improve the quality of any pics you like by x2. This makes them sharper and more realistic. Its drag and drop.

One thing, if you find a prompt you like, copy it and save it off somewhere or you will lose it! The site resets itself if you reload the page or hit the back arrow in the browser!

Syntax is easy:

"A [type of picture], of a [main subject],[style cues]"

[SeVeNeVeS]

dezgo not as good I find, if you use it give us some settings.

Some prompts to make a little more realistic and natural on https://www.patience.ai/

Prompt:
full body shot, RAW photo, best quality, masterpiece, realistic, photo-realistic, ultra detailed, ultra high res, physically-based rendering, Natural, perfect fit body, detailed skin texture, without makeup, moles, freckles, body skin texture, body skin details, moles on body, freckles on body

Negative Prompt:
3d, render, cg, painting, drawing, cartoon, anime, comic, underexposed, low res, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, blurry, bad anatomy, liquid body, malformed, mutated, bad proportions, uncoordinated body, unnatural body, disfigured, b&W, asymmetrical

Download your piccy then run through an upscale........ drag n drop.

If you should find any more usefull prompts please share.

2 facial shots big bro has managed to produce that look almost real.

Spoiler: ShowHide

[Caledonian]

I drew a thing recently that I figured some of the good folk here would like too

[Synistor 303]

I like it! The detail on the wing edges is particularly nice as is the arm-to-wing mechanism. Big smiley face from me.

[Sorontar]

Nice! If we can have artificial limbs, so can birdfolk!

Sorontar

[J. Wilhelm]

I think I'm going to open another thread dedicated to illustrations of concepts introduced in "Valkyrie and the Eagle.". I'm slowly learning how to take seeds from previous illustrations to modify and get results I see in my mind first.

Example: Start with a photograph of the interior of an airship. In this case a Zeppelin styled dirigible. Create an explicit prompt, and don't expect the AI to give you exactly what you want. Then use the "seed" from that AI render and modify the caption to read something else that's closer to what you want, and use style qualifiers ("Steampunk") to change the style of say a 1930s airship to something that looks more 19th century.

Photograph of construction of a rigid airship.


First iteration specifying a photo of a factory like interior in the 19th C. Style.
You get 4 different pictures. One looks like an officer's lounge the other like some sort of war room (pictured below)


Second iteration using last seed and changing prompt asking for a photo of a Steampunk captain's quarters.
You get something very cartoony but interesting.


Third iteration adding negative prompt for "cartoon." It seems for every iteration the AI zooms into the picture.

If instead of "captain's quarters" one asks for "officer's lounge" you get


Adding the Steampunk instead of the "19th C" prompt colourized the image

The two images below are variations offered by the AI. Being a stratospheric capable airship, there wouldn't be too many threats from above, and the top of the envelope would be the safest place for the Officer's Lounge to be. The top deck would also house the US Office of Atmospheric Research. Civilian quarters, dignitary quarters and other amenities would also be on the upper decks.

Officer's Lounge. Just look at the details on those entrance doors!


This one matches Admiral Wilhelm's quarters. The blue color is more consistent with a European ship style,
So it could be the officer's lounge aboard USAS Orca, formerly known as the Austrian ship K.u.K Walkür

[J. Wilhelm]

Some of the things I get from AI aren't what I asked for, but somehow seem to hit the mark for the overall genre and may be useful for another project.

This one looks like a US Civil War submarine with an attached decoy!



This one almost looks like it was taken at Point Loma in San Diego Bay...

[J. Wilhelm]

Luftschiffengel Repair Crew on top of the KuK Walkür, flying over
the Himalayas, en Route to Mount Xinantécatl in Puebla, Mexico.

[J. Wilhelm]

And we have another illustration. This time, given the large volume of illustrations I can make, I'll be starting s new thread, probably in Textual just for The Valkyrie and the Eagle.

Here we have an Engelfolk woman, pictured near the Abbey of Aarmühle (today Interlaken), Canton of Bern, Switzerland in 1867. I started with a basic description of a female red haired elf with an Alpine forest, and the computer game me mountains, a valley and a lake in different configurations, and I picked the most attractive option out of 5. Then I took the "seed" and just refined it a bit. The rest is Photoshop.

With a little post processing...

[J. Wilhelm]

Two more illustrations

Head Portrait of an Engelfolk Woman, Late 19th. C.


Engelfolk Housekeeper, Bern, C.H., ca.1880

[SeVeNeVeS]

you are getting rather good at this my friend.

What prompts are you using and which site, if you don't mind me asking?

As I say I'm sure this AI art deserves a seperate thread somewhere.

[J. Wilhelm]

you are getting rather good at this my friend.

What prompts are you using and which site, if you don't mind me asking?

As I say I'm sure this AI art deserves a seperate thread somewhere.

It's the one you gave me. It's the Patience AI website, using the Deliberate model, mostly and sometimes using straight Stable Diffusion 2.0 (it's kind of picky though), Euler Ancestral sampling, 25 to 50 iterations with an image input, with a guidance scale of 7.5. Pretty much the default settings, because otherwise I run the risk of the servers rejecting my request because of high computer load.

As far as the prompt, I don't use code or hypertext (can you call it that?) to describe what I want, as I don't understand the sintax. The AI is pretty well trained by now to understand English sentences (Kind of scary isn't it?) I just describe what I want very explicitly, like explaining things to a foreigner, and insist on using the word "photograph" for the main subject and for a negative prompt always use "cartoon."

For the refinement technique, I always choose an image similar to the subject of the picture I'd like to see and verbally describe a background (the AI is pretty good at choosing suitable backgrounds). Then after I get the new AI images, then I use the seeds produced by the AI to generate new images, but I change the prompt to be more specific about what I want and iterate into a new image.

I don't always insist on getting a specific type of image if the AI refuses to get close to the picture in your mind; in that case it's best to let the AI "contribute" a bit as it'll be inspiration for a new prompt.

[J. Wilhelm]

And another couple of pictures. Believe it or not, the variations presented by the AI are now inspiring mr to expand the background story.

For example, below is what I imagine an Engelfolk priestess would look like in the Iron Age, prior to the Engelfolk coming into contact with a sprawling Rome (The Engelfolk weren't known by that name at the time).

The canon background story now goes like this:

The Engelfolk are an ancient race of nomadic people who first made their appearance during the 4th Century AD when mentioned by the Roman historian, Sulpicius Alexander, who chronicled the history of Germanic tribes before the death of Valentinian II in 392 A.D.

According to Sulpicius, the Engelfolk were thus nicknamed after angels by the Early Roman Church. They were said to be a permanently persecuted people due to a hereditary characteristic which gave them unnatural youthful and androgynous looks well into adulthood, as well as curiously pointed ears. According to Sulpicius, men and women began to differentiate only after pairing in arranged matrimony per the customs of their people. Another curious fact is that they always come in light skinned and swarthy varieties.

The geographic origin of the Engelfolk is unknown, but it's speculated they settled around Scandinavia after coming from Middle Asia, which might explain the apparent binary nature of their complexion. Some historians believe that they may have been the inspiration for the Scandinavian "Alfar," or Elves, as described in the Prose Edda, written by the Icelandic scholar, Snorri Sturluson, around the year 1220.