For the visualizations of The Perils of Bianca I used the same tools as I used for Sinderella: StudioNeo v2 and GIMP. StudioNeo is the scene creation toolkit for Honey Select 2. I noticed that this is a fairly unique choice for creating adult visual novels, for which developers mostly use Daz3D, generative AI, or self-drawn images (usually anime-style). I rejected these three approaches: Daz3D because I prefer the more cartoony look of characters created with StudioNeo, generative AI because its creations all look the same and because I have a distaste for genAI, and self-drawn images because they would simply take far too much time to create.

Ultimately I am ambivalent about StudioNeo. It is not well-documented. During my work on The Perils of Bianca I discovered many possibilities which I had not found before, which meant that (in my opinion) the visuals of The Perils of Bianca are much improved over those of Sinderella. However, they are definitely not at the level that I would like them to be. While I am sure that there is still much that I haven't been able to figure out concerning StudioNeo, I have been running into its limitations, in particular with respect to facial expressions and body manipulation.

Would I use StudioNeo again for another game? I probably would, simply because by now I know so much about the tool that learning another tool would mean a big time investment to get to the level of knowledge that I have of StudioNeo. But I am not completely happy with it, and I am definitely interested in finding a good replacement.

The second notable feature of the visualizations of The Perils of Bianca is related to the standard approach that Ren'Py has for dialogues (which are usually 80% of the content of a visual novel). This standard approach is the placing of "dolls" on a static background. With a "doll" I mean an image of a character, of which the facial expressions and sometimes the clothes and/or pose can be changed with the text. I used this approach for Sinderella, but for The Perils of Bianca I abandoned it, and went for full-screen images throughout the game (see Figure 1).

Figure 1: Dialogue between three characters in The Perils of Bianca (left) compared to a dialogue using the "dolls-on-background" approach in Sinderella (right).

The reason I went for only full-screen images is that, according to my experiences with Sinderella, they could better reflect the atmosphere of the scenes. I could have chosen to tweak the balance between full-screen and dolls-on-backgrounds a little, but I decided to go all the way and use full-screen images exclusively, partially to test how that would work out. 

A big advantage of using dolls-on-backgrounds is that it is relatively quick to create an illustrated dialogue. There are at least two disadvantages, however. The first is that with dolls-on-backgrounds not much is happening visually during a dialogue, which is a negative for a visual novel. The second is that the dolls will always be "standing" while in the actual scene they might be in a different position.

Full-screen images negate the disadvantages of dolls-on-backgrounds, but the developer pays for that by needing to invest much more time in image creation. And that doesn't always pay off. There are plenty of dialogues in The Perils of Bianca for which dolls-on-backgrounds would have sufficed, and would even have been preferable as more variation can be shown in facial expressions. So if I develop another visual novel, I will bring back the dolls-on-backgrounds, though I will consider for each scene whether it would benefit more from one style or the other.

A final element of visualizations which deserves some thought is how I had to construct images from multiple pieces. Take, for instances, images from the siege on the palace in Act IX. Bianca can be accompanied by any combination of the four fairy tale heroines, i.e., sixteen different combinations. I could have made sixteen different versions of each of these images, but that would increase the size of the game enormously (see Figure 2). 

Figure 2: Two of the sixteen possible combinations of characters in a scene during the late game of The Perils of Bianca.

Therefore most of the images of the game are pasted together from multiple smaller parts. Images consist of a basic background, to which elements are added conditionally (see Figure 3). 

Figure 3: Combining a background which contains Bianca on her own with two elements that are pasted in to display two fairy tale heroines, creating the image that is shown in Figure 2 (right).

I found that this approach is rather error-prone. Defining the conditions was often difficult; they are usually testing two or more flags to determine if a particular element should be added. I also had to make sure that each of the pasted-in elements was positioned correctly. It regularly happened that during a playthrough I noticed that an image was displaying something that did not belong in the scene for the present game state, due to a mistake I made in one of the thousands of conditions that I had to write. And occasionally one of the pasted-in elements was placed incorrectly. Using dolls-on-backgrounds would simplify the use of conditional elements considerably, which is one more reason to revert to dolls.

Size-wise, however, my approach was a good choice. The Perils of Bianca contains about 1200 basic images, with in total about 6000 variations. In comparison, Sinderella contains about 400 basic images without any variations, plus 160 dolls which are placed on top of about 25 backgrounds. Despite the enormous increase in number of images, The Perils of Bianca ended up being only about 50% bigger than Sinderella.