Midjourney V5 vs V6: 10 Portrait Examples That Show the Differences (Before/After) : Amalie Mayer Altimira
by: Amalie Mayer Altimira
blow post content copied from Be on the Right Side of Change
click here to view original post
Example 1 – Taiwanese Girl
Prompt: "Hyperrealism. A taiwanese girl who looks about 20 years old. She has shoulderlong black hair with bangs, heartshaped lips and peach blossom-shaped, dark brown eyes. She looks over her shoulder at the camera, looking slightly surpised"
A great example of what Version 6.0 does better compared to Version 5.2. The second photo processes the prompt way better, as mostly visible by the fact that the second one is incredibly realistic, as asked for in the prompt, while the first photo has more of a cartoonish vibe.
Example 2 – Seventeen-Year-Old Boy
Prompt: "A fairskinned boy of about seventeen years. His brown hair is put up into a ponytail. He has defined cheekbones and a strong jawline. He has intense, forest green eyes."
The second photo makes a lot more sense. The hair of the boy is not only closer to what was requested in the prompt (Ponytail), but it also makes a lot more sense as the hair of the boy in the first photo is all over the place and, in some cases, not even connected to the head. As the prompt’s author, the second photo also shows what I had in mind as I prompted.
Example 3 – White Hair Boy
Prompt: "A boy of about seventeen years. His white hair is put up. He has defined cheekbones and a strong jawline. He has intense, forest green eyes."
The main difference between the first and second results is the color palette. While this may change depending on who looks at the picture, the blue background of the second prompt conveys the intensity of the picture a lot better. The first prompt also has some problems with the hair, while the second one does not have the prompted ‘forest green eyes’.
Example 4 – Burgundy Red-Uniformed Woman
Prompt: "A female with eyes that are burgundy and black and bright red, but still calm because the colors fade into each other. She wears a simple red and white ballet uniform. She has dark brown hair with a single red strand, put up into a strict bun. In the drawing she looks up into that camera that seems to be straight above her, around her there is darkness. On her face, there is soft dark red lighting. Realistic animated stile"
The first prompt simply disregards many of the requested features (red lighting, ballet uniform, stile, camera perspective). While the second prompt also disregards some of the above, it gets the general vibe of the photo better with the red color scheme and the more cartoonish stile.
Example 5 – Young Hunter with Ponytail
Prompt: "A young hunter standing atop a fallen tree of about 16 or 17 years. He is looking off into the sunny forest, looking concentrated. He has middle long hair that he has put up into a short ponytail. His hair has multiple different shades of brown with a single white strand. His face has a defined bone structure. He has green, sharp eyes with wrinkles around the eyes, as if he was smiling often."
Neither picture manages to convey all the prompted features. The second picture, though, looks a lot more realistic and bright, making the overall message easier to receive.
Example 6 – Golden Blond-Haired Girl with Glasses
Prompt: "A 17 year old girl. She has golden blonde, beautiful hair, golden glasses and wears light blue, flowy clothes. She has a middle parting in her hair and seems very modest, not really like someone who would be the center of attention. She stares off into the distance, obviously daydreaming"
The first prompt output (v5.2) has a more down-to-earth vibe, but not only doesn’t make a lot of sense regarding the hair, it also does not have the dreamy look aimed for in the prompt.
The second photo is so much more detailed and dreamy, showing beautiful golden lighting without even needing it to be prompted, but it also shows the specified clothing and glasses. The parted lips also give her a perfectly fitting, surprised look, playing with her facial expression much more.
Example 7 – Pretty-Eye-Shaped Girl with Short Nose
Prompt: "A 16 year old girl with fair skin, small, black eyes in a pretty eye shape. She has a small, short nose with a straight noseback and lips in a nude color. She has long, dark black hair and has it put up in a high ponytail on the lower top of her head"
Important to mention here is that the photos differ drastically in style, so there is no right or wrong here. Neither picture is perfect; the original one again struggles to show the hair properly, and the picture generated by Version 6 zooms in so much that you can’t even properly see the hair as well and neither shows the proper eye color.
Example 8 – Radiating Silver-Haired Girl
Prompt: "A 16 year old girl with whiteblong hair, mostly white and silver, that looks radiating in the sun as if it were transparent, but not see-through. Her hair is in an ponytail that goes a little farther than her shoulders. She has a pretty face with an sharp jawline, an straight, small nose, red lips and big, mint colored eyes. She has an serious expression and seems to be from the middle age."
In my opinion the only photo so far that is drastically better in Version 5.2. Also, the only picture on this list that was generated more realistically in Version 5.2 rather than Version 6. In this case, the second version is closer to the prompt, but the original one is more beautiful and seems more emotional, and the most important part of the prompt – the hair – is highlighted a lot more.
Example 9 – Tattooed Young Woman with High Cheekbones
Prompt: "A young woman with high cheekbones, a long, straight button nose, bow-shaped lips and a strong jawline. She has monolid eyes, in a golden color. Her chin is rather small, and her eyes are set low in her face. She has a her gold-white hair put up in a ponytail, and she wears and thin, light leather armor. On her chest, there is a simple tattoo of a plant, it looks like ivy. She is tall, and has elven ears. She looks distinguished and elegant, like a dancer. The drawing style of the picture is in the style of Baldurs Gate"
A few mentions needed here is that (1) I had a lot more promising pictures with the second prompt in V6, and (2) it took a lot less work prompting to get a result I was happy with. Also, I noticed that Version 5.2 had immense struggles with tattoos, while Version 6 got it right on the first try. Otherwise, the pictures don’t vary a lot in objective quality.
Example 10 – Green Siren Glowing Eyes
Prompt: "A female siren with forest green skin color, parakeet green glowing eyes and pine green, full lips. Her hair is green as well, it has the same color of her lips. Her eyes are almost closed, but her pupils glow so you can see that they are still open a little bit, her lips are slightly parted, her face leaned left slightly. She has her hands stretched, clasping each other in her lap. She wears an fern green dress. She looks evil, and has an evil smirk on her face"
The first prompt gets much of the posture better than the second result. Also, it fits the generally aimed-for vibe more. But the second one holds way more emotion and is way more realistic.
My Opinion on Which Midjourney Version is Better (V6 vs V5)
Midjourney Version 6 makes it way easier to get good results, and it gets realism a lot better. Some problems – hands, hair, teeth – are also seemingly fixed. Sometimes, Version 5.2 gets better results, depending on what you are looking for, but overall Version 6 holds many advantages.
Recommended: How Do You Check Your Midjourney Version and Change the Default Model (V6)?
December 28, 2023 at 08:56PM
Click here for more details...
=============================
The original post is available in Be on the Right Side of Change by Amalie Mayer Altimira
this post has been published as it is through automation. Automation script brings all the top bloggers post under a single umbrella.
The purpose of this blog, Follow the top Salesforce bloggers and collect all blogs in a single place through automation.
============================
Post a Comment