Thanks total_newbie, this feedback is very helpful. The clipping error and misaligned brushwork were bugs that crept in during construction. I have uploaded a new version where those errors are fixed. I also added a couple of additional light sources in the last rooms of the level (though not too many, so as to maintain the intended atmosphere).

As before, the file is at

https://ubiquitousgame.wordpress.com/2016/09/18/quake-map-infernal-ascent/

Thanks again! 

It's not in the news, so here goes.

Yes, strong retro vibe. Combinations of enemies worked well for me. Best part of the map was the button sequence puzzle room. That shambler + ogre room could be improved, I easily worked it around. Dogs gave me a nice scare.

My criticisms are about lighting and signaling. Lighting felt kinda caotic to me. Spotlights not highlighting anything special; pitch black darkness at some points, worse being that section after you fall thru the funnel, where you navigate completely blind.

After beating the map I noclipped and "r_fullbright 1". That's when I could appreciate your brushwork. I had no idea that I fell thru a funnel, for example. Shadows are supposed to enhance the environment as much as light, right?

One last bitching: you hint attention points with textures as well, but without anything special around. If you choose a white light fixture set, a yellow fixture draws player attention, for example.

I think you managed to create that classic unsettling feeling of danger that iD maps used to give. Also, passing thru the same room from different levels/sights is something I love and value, you're putting the environment to use.

Hope I could help someway. Thanks for this release, it was good fun :) 

Skill 2, 2/5 secrets 64/65 kills (Glitched zombie!)

https://www.dropbox.com/s/1ggljkzaht0wh1t/mjb_infernals.zip?dl=0

Nice map to play on a Sunday afternoon! I like the theme of rising into different realms of Quake. The transition is done nicely through the use of broken bricks and the base textures. Game play was fun with lots of knights to ambush you in the shadows.

I would say a large thing I noticed is the lack of source lights being illuminated properly. So when I saw a light source, the actual lamp texture would blend into the background. Placing a light with a high wait value (7-8) can really bring that nice touch home! Also I agree with adib about lighting in general...bit all over the place.

This is a unique map and I am looking forward to seeing more! 

Thanks guys. The light surface textures not being properly illuminated has been bugging me too. I don't understand why id made light textures that don't use fullbright colours. Rather than do as you suggest, Bloughsburgh, and use a high wait value, I implemented the simpler (for me) solution of making custom versions of id's light textures that are fullbright and therefore appear to illuminate properly regardless of how much light falls on them. This fix is in a new version that i have just uploaded on the same page as before:

https://ubiquitousgame.wordpress.com/2016/09/18/quake-map-infernal-ascent/

Sorry to hear that you found the lighting to be chaotic. I'll be the first to admit that I am still learning the fine art of lighting, but there were, at least, a few elements of the lighting in this level I was happy with. Hopefully, what I learned through this experience will inform a more nuanced approach for my next creation.

Thanks a lot for the demo, it's valuable to watch how others get along with and interpret the level. 

I saw the map showed up on the front page news (http://celephais.net/board/view_thread.php?id=61352), so it might make sense for further discussion to be consolidated in that thread. 

Here :) 

At first, 8-bit color translucency was like this. Due to the limits of Quake's 8-bit color palette, straightforward blending of color values was very lossy.

Then I've figured out a way to apply color correction on both translucency and lighting. The colors improved a lot, although the results were very grainy (see how the shape of the slime texture became completely unrecognizable after being blended twice).

After that, I've decided to try implementing gamma-correct translucency. The blending became smoother, color hues became more faithful and the shape of the slime texture could be recognized again, but a few details were lost in the darker areas of the image.

Now, I've also implemented gamma correction on the lighting. This helped to recover those lost details back, and then some. Now the slime texture is fully recognizable, even through two passes of alpha blending, and details such as the pipe texture on the bottom left of the image are more faithful than ever.

For a final comparison, see this 24-bit truecolor mockup I've made in GIMP (which doesn't use gamma correction). If you convert this 24-bit mockup back to Quake's 8-bit color palette using popular error diffusion algorithms such as Floyd-Steinberg, my results are still smoother.

I have one more thing to change in Retroquad's color management, but I'm skeptical about further quality improvements. This "one more thing" will be just to simplify some calculations. 

Why does the glass look like a sieve in all your color/gamma corrected pictures? 

Because of dithering. 

OK. And I suppose there's no workaround unless you revert back to the uncorrected look? 

I'm so excited for this mankrip

quake in its intended 256 colour brilliance. 

this is pretty darn sweet. 

Exactly. The only workaround would be to convert the whole renderer to 24-bit color, which is what the Pixomatic software renderer in Unreal Tournament 2004 did. 

Which map are these pics taken from, BTW? I don't recognize it. 

Can't wait to start experiencing that by myself too* so this is your own engine project? 

Mugwump: It's Clean Cut, by MFX.

NewHouse: Yes. 

Haven't played that one yet. 

 

Qmaster: The name is Retroquad, but it may change before release. 

Just found this - might be useful for you:
http://bisqwit.iki.fi/story/howto/dither/jy/#Appendix%201GammaCorrection 

Take a look at the whole article. I pasted a link with redirection to "Gamma correction" by mistake, which is least interesting part of this page. 

Thanks, I've read all of it (minus the code, because I'm on a smartphone atm).

So far I've not used any iterative color search methods, only straightforward calculations. But that article may give me some ideas on how to improve TGA skyboxes and external textures. Skyboxes such as "Interstellar" and "Violent Days" still looks pretty bad in Retroquad.

Another idea on how to improve external textures is to perform topological analysis to scale the error diffusion in a non-linear way. This is in my mental list of things to try sometime.