I've helped many newbs and returned old players configure ezquake, and there are two bad malconfigurations that routinely occur:

Mismatched vid_conwidth/vid_conheight aspect ratio to resolution ratio which results in distorted geometry.

Uncapped cl_maxfps, which desynchronizes video/input fps with network fps, which results in varied delays, and also excess usage of resources which often adds delay to the processing of input. And unstable fps.

thanks bro

fuck you povo, mocking fuck

So, what is your point? What's the solution?

My solution was to make things known by doing this and this. I wouldn't bother ranting about here though, unless it was on HOWTOs section.

What's the point of this thread?

Your post says ezQuake, but does this also apply to the nQuake setup? (which i experienced was pretty flawless regarding the smoothness etc if i remember correctly)

My points are:

vid_conwidth and vid_conheight shouldn't be alterable, in increments of eight, seperate from each other in the menu, or better yet they should be restricted to the modes that won't distort the dimensions of the current resolution.

FPS limit (video fps) in the menu is restricted to either 0, 72, or 77, which is just wrong.

Renzo, in your widescreen guide the formula you propose for correcting vid_conwidth/vid_conheight only corrects the geometry, it'll never yield simple divisions of the resolution, the best values for displaying font and hud.

Ake Vader, It applies to nQuake package too.

I agree with these points.

This limit might be possible to remove. I looked into it a little after you submitted a bugreport on sourceforge and it seems that it might be possible to remove the restriction but I have not played around with it yet.


Submit a feature request on the ezQuake page please if you haven't already done so.


This is also dependent on the resolution of the font that is used so best values differ, does it not ?

/bio

Every affects everything, let's recap:

1) vid_conwidth/height affects the size of the font on the screen, it also affects world geometry and on-screen items for armors/health/etc
2) screen resolution, of course, affects the readability of the font. The lower the resolution the worse fonts look like.
3) console font image or "charset" affects the font directly. If you have 256*256 charset it will look like shit whatever the settings are, the higher the resolution of the charset the better it will scale with conwidth/height and resolution.

My advice:

- use vid_conwidth/height of 640*480 (768*480 on 16:10) or divide that by two to make text bigger (320*240 or 384*240 16:10). The higher the conwidth/height the smaller the text becomes.

- use as high screen resolution as possible, as long as that gives you 150Hz refreshrate. On TFTs use native or the one that gives you 75Hz if it works properly.

- use high resolution charset. Adjust contrast/brightness with photo editor and possibly add gaussian blur with value of 0.3 (photoshop specific value) to make it look better. Use gl_smoothfont 1 always.

I just looked through a single page of the configs section of qw.nu and three in twenty configs have distorted geometry due to mismatched vid_conwidth/vid_conheight to resolution ratios, 15 percent have fucked geometry!

Any particular reason for targetting 150hz? I think the best strategy is arguably to use the highest refresh rate your monitor supports, and then work backwards to find the max resolution it can handle (you can roughly calculate this as Width = Horizontal frequency/1.1/refreshrate). So for example lets say you have a 121khz monitor capable of 160hz max. We would find the resolution width as:

121000/1.1/160 = 687.5

So assuming an aspect ratio of 4:3, we'd want a resolution of something like 4/3*687x687 = 916x687. [You might want to tweak this slightly to give numbers more readily divisible for your console res, or more acceptable to your gfx driver]

Now obviously some video drivers are better than others for allowing custom resolutions; I've normally had pretty good results with Nvidia over the years.

I'm always thinking about full multiplications of fps and matching them with refreshrates, let it be mouse or monitor.

Basically 150Hz because it's close to physfps 77 * 2 = 154 maxfps. Running 150fps on 150Hz monitor causes visible tearing, running 154fps on 150Hz monitor causes less tearing. Running 300fps on 150Hz, again, causes some tearing while 308fps (4*77) causes less tearing. Read some of the links I gave above if you haven't already, there's explanations for everything (well, not that exact but anyway the theory behind it all).

I read that the other day and I'm still not totally convinced; I mean, why is 150hz any better than say 151hz or 157hz or some other mismatched value? It just seems a bit arbitrary to me. Most of what you have written is good stuff (particularly about the 'requirement' to have maxfps as a mulitple of physfps), I just think that things like 'X fps is close to Y*ZZZhz so is good' needs further explanation. I just don't see why 150hz is a better choice to use with 308fps than a higher value which is closer to 308/2 providing it doesn't give excessive tearing.

Don't get me wrong, I can see where you are going mostly and indeed refresh/res tweaking is something I've been searching for the holy grail in quite a while now (see some old, outdated and sometimes naive columns touching on the subject here!: http://web.archive.org/web/20040607101818/www.challenge-smackdown.com/div_neu/columns/hangtime/ ).

Some years ago (before networking and fps were unlinked) I used to play with 77fps and 154hz and that seemed fine to me while others said they got tearing (I found that 77fps/77hz gave me tearing, but 154hz was much better). I remember talking with Strider about it (at the time he seemed to be one of the few coders really looking deeply into hardcore client optimisations) and he said that without vsync, there isn't actually any technical reason why picking 154hz should be optimal because in reality the monitor isn't refreshing consistently at a discrete 154 times per second, it's actually maybe 154.002, or 153.99, or whatever, due to timing not being 100% accurate. So essentially the idea of having a precise number of frames per screen refresh isn't possible, although I did find it felt pretty good and I suppose you'd only get an 'anomoly' (screen refreshing 1 or 3 times instead of 2 per frame) very rarely as it'd probably only happen every few thousand frames or something.

It's very simple while it's not simple at all. 154Hz is "considerably" better than 150Hz if you think about the framerates (especially physfps 77 and maxfps 154/231/308/385/...) and will always give you the best results if you want the smoothest possible experience.

You can try it yourself, drop the fps-i off, set your mouse to 125Hz, cl_maxfps 125 and tune your monitor to 125Hz. It is probably smoother than say, physfps 77 maxfps 308, mouse 500Hz and monitor at 125Hz.

One good setting could be 125Hz monitor, cl_maxfps 500 and mouse 500Hz even if the physfps*n is 462 or 539 at closest. Like I said in the theory of smooth qw blog, after certain point it doesn't matter how much fps you are getting, it's not going to be any better, however running maxfps something else than physfps multiplied by integer causes slight annoyances at lower fps.

Besides, 150Hz isn't really matched to anything. It's just the value you can get out of NVIDIA drivers out of box as long as you have monitor inf installed and your monitor isn't from the 80s. If you can go higher then go ahead, but there isn't going to be much of a difference to the human eye, if any at all. The smoothness is very dependant on the all four settings, mouserate, physfps, maxfps and monitor refreshrate and failing on one can cause the smoothness to go away. 77/154/150 is the closest and the easiest one to get with modern hardware, and while it's not the best it's still better than 95% of the other setups.


Edit: heh, just checked my original post:

Looks like you didn't read it carefully enough.

That's what I'm getting at - if 154hz is considerably better than 150hz, why not go for it?


The thing is here, if we are doing pretty hardcore optimizations, why not go the whole hog and aim for something better than 150hz? Yes, it's a small difference, but my guess is most people who are doing serious tweaking will have the energy to make a one-time change to their gfx driver configuration.


Well, what you wrote in this thread was somewhat contradictory (implying that 150hz was optimal, not 150hz+), and the only reason I picked up on it is because this number of 150 still seems to crop up - it wasn't until now that you explained that the main reason is because it is 'easy' (due to being a default option in NV drivers).

I think it's great that you have taken the time to write the article and as you say it is probably better than 95% of other setups. But maybe it would worth be worth clarifying this in the article so that people who want to go the extra mile know what they should be aiming for. The chances are most monitors capable of 150hz will be capable of 154hz also (160hz seems to be a common limit) and as you have said yourself just now that will give the 'best results'.

The thing is, I didn't write that stuff to be the "ultimate setup" guide, but more like giving people information about what affects what, and what's the (minimum) setup they should try to go for.

I also gave example settings, and as you can see I also added "+" to the arguably "better" example too.

The finetuning should always be done by the player itself, as noone can tell you what's the best setting for you.