In my project where I've been rendering Quake 2 maps (https://github.com/Vurich/goeld) I implemented tonemapping, by rendering everything to Rg11b10Float and then using an implementation of ACES tonemapping on the luminance value of each pixel (shader here). This makes the whole world look way better, and I can't recommend it enough - especially since you've already got a post-processing step. It makes everything in shadow a lot easier to see, and makes the image more cohesive overall. Check out some of the screenshots in the readme. You may need to play with the gamma and intensity amounts until you get something that looks pretty good, but it shouldn't be too tough to implement. Modern HDR rendering uses adaptive gamma, but old games designed their maps around having a single gamma value so you don't need to implement anything like that.
In my project where I've been rendering Quake 2 maps (https://github.com/Vurich/goeld) I implemented tonemapping, by rendering everything to
Rg11b10Floatand then using an implementation of ACES tonemapping on the luminance value of each pixel (shader here). This makes the whole world look way better, and I can't recommend it enough - especially since you've already got a post-processing step. It makes everything in shadow a lot easier to see, and makes the image more cohesive overall. Check out some of the screenshots in the readme. You may need to play with the gamma and intensity amounts until you get something that looks pretty good, but it shouldn't be too tough to implement. Modern HDR rendering uses adaptive gamma, but old games designed their maps around having a single gamma value so you don't need to implement anything like that.