Drew Brucker’s Post

Think that AI generated imagery isn't any good yet? Think again. Firstly I'd argue you're probably not using it well enough yet. We're fooled by the interface into thinking it's easy to use, but there is a skill to using Midjourney and it takes time to master. I'll be sharing more on this over the next few weeks as we roll out our Midjourney for Creative Professionals course. You wouldn't expect to be able to use Photoshop without training, so why do you think you can master an amazingly complex tool like Midjourney without it? Secondly, it is getting better and better very quickly. If you haven't used Midjourney since early last year (when I started using it with v4), prepare to be blown away by v6. And just read below for the updates coming later this month with v6.5, and what's planned for the end of the year with v7. Thanks for the update Drew Brucker

Knowing VEX often lets us not know all the nodes. But knowing nodes lets us skip VEXing while there is a shortage of time during the hot phase of the project. For example, look at this tiny but precious tutorial by Emīls Geršinskis - Ješinskis https://lnkd.in/dmAV8fXM. He deletes the outer Voronoi fracture pieces in seconds just by using the GROUP_EXPAND node. When I approached this task before, I was actively VEXing: first calculating the intersections along the normals for every point and then removing those pieces that contained the points without any intersection (except themselves). And it was significantly more than just a few seconds :). While knowing 'one little checkbox' could save me some time. Or for example, usually when I needed to group ends of the curve, I used this VEX: if(neighbourcount(0, @ptnum)==1) {i@group_ends = 1;} Short and simple, but then I saw in the tutorial of Fabricio Chamon that it can be done in 4 clicks with the node GROUP_by_RANGE https://lnkd.in/d_R_mVZP. On the other hand, everything that is based on the number of items stored in our head is a very 'use it or lose it' kind of thing. When I was an active Adobe After Effects user, I was a big fan of gadgets like presets and snippets. My script panel had about 150 snippets for different kinds of things. I also had my system of hotkeys, optimized for the left hand. But when I stopped doing it every day, this knowledge started to dissolve very quickly. Much quicker than I expected. The ability to forget is a good thing, but sometimes it's so frustrating :). So I think the optimal way is always somewhere in the middle. Knowledge of items is good but not reliable from the long-term perspective. While it can be very advantageous in every-day-Houdini work requiring speed.

Ever find yourself in a design crunch without the luxury of eye-tracking or heat map tests? 🧐 Enter the 'Squint Test'—a quick, budget-friendly hack. Blur tool, anyone? How do you tackle design evaluation when time and budget are tight? Share your tips! #uxresearch #uxtesting #uidesign #designhacks #uxdesign

I was currently having problems on how to make more organized shader codes in my game engine. I found how to be more organized and it is hard to find how to handle these in SPIRV-Based Shader codes in modern apis such as Vulkan and DX12. You can read my gist and implement it according to your needs. Hope you find it useful. https://lnkd.in/dt_3e6rk #cpp #shader #vulkan #rendering #renderingengine #gameengine #glsl #hlsl

Midstone update: So, have you ever started a project and just used a bunch of ideas that work on their own and think that it should be easy to force them together? I learned that it doesn't usually work like that this week. Our midstone project is making a relatively simple maze game in SDL. The logic behind it is relatively simple, but for some reason, we used three separate coordinate systems for the pieces. The player was placed and rendered in the Physical Coordinate System, which is used when applying physics to objects in your scene. In projects moving forward, this is the system that would best be utilized. The 2D Maze Array is the second Coordinate System. We made an array to hold how the maze should be built. Ones would be tiles for walls while Zeros were for open space. For instance, it could look something like this (not to scale) : { 1, 0, 1, 1, 1, 1, 1, 1, 1, 1 } { 1, 0, 1, 0, 0, 0, 1, 0, 0, 0 } { 1, 0, 0, 0, 1, 0, 1, 0, 1, 1 } { 1, 1, 1, 1, 1, 0, 0, 0, 1, 1 } { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 } That system was used to store the locations of all the tiles. For the floors and walls and whatever else we are going to add after we fix this. The third system is the Pixel Coordinate System, which uses how many pixels are on the screen as a reference for locations in the scene. The Maze Tiles are using this system to render. As you may imagine, this over complicates things. We could be using 1 system, or maybe 2 at most, but using all three causes problems. We had some issues translating positions through the Coordinate Systems, and all of a sudden all of the tiles were hundreds of pixels away from the players starting location, but the collision detection still worked as intended, and so now we have a new goal. For this coming week, we are going to alter our program to utilize the Pixel Coordinate System above all else. This will basically make every object in our game talk the same language and will make it much easier going forward. Thanks for reading, - Aidan

What's new with the #Midjourneyv6 base model? • Much more accurate prompt following as well as longer prompts, • Improved coherence, and model knowledge, • Improved image prompting and remix mode, • Minor text drawing ability (you must write your text in "quotations" and --style raw or lower --stylize values may help) /imagine a photo of the text "Hello World!" written with a marker on a sticky note --ar 16:9 -v6 • Improved upscalers, with both 'subtle' and 'creative' modes (increases resolution by 2x) (you'll see buttons for these under your images after clicking U1/U2/U3/U4) The following features / arguments are supported at launch: --ar, --chaos, --weird, --tile, --stylize, --style raw , Vary (subtle), Vary (strong), Remix, /blend, /describe (just the v5 version) These features are not yet supported, but should come over the coming month: Pan, Zoom, Vary (region), /tune, /describe (a new v6 version) Picture by Midjourney

Random Walker algorithm can be employed for various purposes, contributing to the dynamic and unpredictable behavior of characters, objects, or elements within a game environment. Here's how the Random Walker algorithm might be applied: Map Generator: Green represents Grass/ground Blue represents Water By giving parameter you get result like this one . #unitydevelopment #unitydeveloper #unity3d

Good to know

"What do the --sw and the --stylize mean? Can you explain how they work?" --sw (style reference weight): This controls how much influence your style reference (a sref code or image) has on the output. The default value is 100, and the maximum is 1000. --s or --stylize: This adjusts how much the training data affects your output. The default value is 100, and the maximum is 1000. A higher value, like --s 1000, typically produces results that most Midjourney users prefer. Lower values help the bot understand your prompt better. Generally, higher --s values result in more photorealistic images, while lower values yield flatter images, depending on the prompt and style. For each prompt and style, you should adjust the balance between --sw and --s to achieve your desired result. #midjourney

The depth of SVG path commands is fascinating! You can draw any line and curve you wish at scale. So powerful! Here's a quick summary: https://lnkd.in/gQZE3_fa