tlmfoundationcosmetics.com

# Comparing Eluna AI, Midjourney v4, and Midjourney v5

Written on

Chapter 1: Introduction to AI Image Generators

Recently, I had the opportunity to experiment with Eluna AI, and I decided to utilize my 20 credits for some creative fun. Before diving in, it's essential to mention an amusing tidbit: my girlfriend has a soft spot for both pasta and African porcupines. Back in the days of Midjourney v4, I attempted a rather unconventional prompt:

"Cartoon, African porcupine enjoying a large bowl of pasta."

Despite my less-than-stellar prompting skills, the result was surprisingly decent—minus a few oddities like the endlessly repeating fingers and an unusually shaped spoon.

Cartoon porcupine enjoying pasta

Midjourney v4 was impressive for its time, but now we have Midjourney v5.

What happens when we use the same prompt in this latest version?

The results have significantly improved, as evidenced by these four images generated from the same prompt.

Midjourney v5 output of porcupine and pasta

It's clear that Midjourney has advanced remarkably, though it still faces some challenges. It's essential to recognize the progress this technology has made in just a month.

Section 1.1: Addressing AI Anxiety

In one of my previous articles, I concluded with some reassurance for current and future college students regarding AI's evolving role in education.

Now, let's focus on that charming little character in the upper right corner. Let’s enhance the image—wait, what’s going on there?

Zoom and enhance.

Enhanced image of the porcupine

Midjourney v5: Pasta, porcupine, and… some peculiarities?

Rest assured, Midjourney does not misappropriate artists' works; it just closely mimics their styles, sometimes including unintended watermarks. Unfortunately, this raises ethical concerns regarding the use of materials without permission.

So how does Eluna stack up? There’s plenty to appreciate, but it also has its areas for improvement.

I adjusted the prompt slightly because Eluna interprets requests differently. The interface is considerably more user-friendly—no surprise, given that it's not hosted on Discord.

You can see details like the creation date, seed, pixel dimensions, and step count right on the screen, with sliders available for customizing creativity.

Screenshot of Eluna AI interface

However, the initial results left something to be desired. Here’s Eluna's first attempt:

Eluna's first attempt at the porcupine prompt

What on earth is that?! Let’s take a look at Eluna’s second try:

Eluna's second attempt at the porcupine prompt

While it doesn’t resemble a porcupine, it does have a unique aesthetic. Eluna's pixel model was next on my list. This model produced both a great success and a notable failure.

Output from Eluna's Pixel model

Back to Eluna, but with an increased step count. By raising the steps from 45 to 72 and doubling the credits from 2 to 4, I achieved a result that I found much more satisfactory, although it still resembled a mysterious creature with the head of a badger.

Enhanced output from Eluna

I genuinely liked this one. The image features some elements that resemble pasta, but the creature itself is quite believable.

Section 1.2: Observations on Eluna

A few notes about Eluna's performance:

  1. Eluna confuses porcupines with hedgehogs.
  2. It consistently depicts porcupines as either eight-eyed hedgehogs or as hedgehogs with badger heads.
  3. The model tends to show animals eating in a natural way rather than using utensils.
  4. It struggles with understanding how plate and bowl combinations work.

Overall, I wasn't overly impressed, but it's still early for this model, and asking for a porcupine eating pasta is quite a peculiar request.

Let’s try something more conventional:

"A portrait of Albert Einstein with multicolored, high grain, cinematic colored smoke and lighting."

Here's the 45-step version:

Portrait of Albert Einstein with colored smoke

Not bad! Now, let’s increase the steps to 65:

Enhanced portrait of Albert Einstein

While I preferred the composition of the first image, the quality noticeably improved with the extra steps. Both images are quite similar. I’m excited to see how Eluna progresses in the future. With its robust interface, I can envision it becoming a powerful tool as it continues to evolve.

If that seems ambitious, just look at the differences in results from the same prompt in Midjourney v4 and v5 within a single month. The advancements are astonishing.

Section 2: Further Exploration with Midjourney

Now, let's give Midjourney another shot with a different prompt.

Here are Midjourney's outputs:

Midjourney's output for the new prompt

Not too shabby! Let’s upscale the fourth image.

Upscaled image from Midjourney

Look at that—no watermark! I’m genuinely impressed with both Midjourney and Eluna's results this time, and even more so by the distinct differences between them.

At present, Eluna clearly demonstrates two main points:

  1. Midjourney currently holds an advantage over its competitors.
  2. The competition remains fierce, with plenty of potential for innovation.

As an artist, my initial excitement about AI has waned. What once felt like an exhilarating romance now seems somewhat routine.

To see a fun comparison of AI image generation, check out this video: "Blind AI Image Test! Is it MidJourney 5 or V4?"

For a deeper dive into the capabilities of various AI art generators, take a look at "AI Art Generator Comparison: Midjourney V5 (Paid) vs. Blue Willow (Free)."

Follow my journey on Twitch to engage with fellow AI enthusiasts and witness my ups and downs as a writer navigating new AI technologies. Don’t forget to subscribe for more insights!

Share the page:

Twitter Facebook Reddit LinkIn

-----------------------

Recent Post:

Understanding Sibling Dynamics in Gifted Families: A Deep Dive

This article explores sibling dynamics in gifted families, revealing insights from qualitative research on their unique challenges and experiences.

Integrating a Dendrogram into a React Application Using Visx

Learn how to incorporate a dendrogram into your React app using the Visx library.

Unlocking Unique Habits: Elevate Your Creativity and Output

Explore how unconventional habits can enhance your productivity and creativity, featuring practical tips and insightful videos.