First of all, in the week just past I made a start towards adding music to the game.
A few of you reading may know this, but my brother (who is a musician) had previously offered to help me with music. To that end, I met up with him again in the week just past, and we further discussed and experimented with sounds.
It was difficult, I'll admit: I found it tricky to communicate what I wanted, what I thought might work, and he has little experience with game-music, specifically. On top of that, since he's doing this for free, I don't want to ask too much. In the end, it looks as though we may only produce a title theme from this collaboration--but that alone is exciting.
Right now I don't have the money to hire a composer, I fear. As a result, I've started looking through royalty-free music. I've pretty much only begun this process, however, so I don't yet have anything selected!
(As a side-note, I've decided that I want to avoid using Kevin MacLeod's music: my impression is that it's the royalty-free music most often used by indies, and I don't want my game's soundtrack to sound too familiar!)
(If I run a crowdfunding campaign, and if that goes well, I may add the hiring of a composer as a stretch goal.)
On the technical side, in my continued efforts to improve performance, I decided to remove the detail texture being applied to the first level's grass.
I had realised, you see, that the entire texture wasn't really called for: the nature of the grass geometry meant that I could easily just use the colours at the model's vertices. To that end, I used Blender to read the detail texture and store values derived from its colours in the vertex-colours of the grass, allowing me to have much the same effect without an additional texture-read!
(Alas, this doesn't seem to have helped much, which seems odd. I'm honestly a little mystified at some of the performance issues in the outdoors section of the first level! :/)
Another attempt at improving performance involved a change to my method of handling "sunlight"-shadows.
From what I've seen, (and as I've mentioned before, I think) part of the cost of those shadows is that they render into a texture much bigger than the one used for the "player-light". But simply reducing the size of that texture results in the shadows being of much lower quality. However, note that this effect is far more visible at close range than far: in the distance, perspective reduces the size of the "shadow-pixels", and thus the impact of their quality, I believe.
Thus, I'm attempting a two-layered system: Two shadow-cameras, both rendering to smaller textures. One handles nearby shadows; since it covers only a limited area, the lower texture-size shouldn't result in much reduction in quality. The second camera handles more-distant geometry, doing so at lower quality, but hopefully without that being too noticeable.
It's still a work-in-progress, however, so I don't yet know whether it will work!
Dealing with this, however, led to my discovering bugs in my handling of the "sun"-light and the sky-box, as I recall. For one, the "sun"-camera wasn't properly following the player; for another, the sun rendered in the sky-box and the "sun"-camera weren't properly aligned.
On top of this, I decided to change the way that I was handling the "placement" of the sun: instead of specifying a direction-vector, I felt that specifying an angle and elevation would be more intuitive. (And indeed, having implemented it, I think that it is.)
Getting all of this to work was a tricky business, as I recall, but I think that I have all of this working now!
And finally, as shown above, I decided that I wanted the end-of-demo "cutscene" to use my animated mist in its backdrop, instead of the static image that I had previously used. I was a little concerned that it would be too distracting, but with the appropriate speeds kept low, I think that it's fine, and that it looks rather better than the static backdrop!
That then is all for this week--stay well, and thank you for reading! ^_^