1911 replies to this topic

[TheCaptainJZ]

IraqiWalker, on 28 December 2014 - 09:37 PM, said:

One thing I should bring up is that terrain rendering sometimes causes part of this. I remember taking pot shots at a stalker that I had a clear line of sight to, and not getting a single hit to register. Moved two steps forward, one to the right, and suddenly there was a rock there that was blocking not only my shots, but also my line of sight. I should not have been able to see the Stalker, but because the terrain was there, despite not rendering, I did see it.

ah yes, I have seen this before on areas where the map is hilly and you are far away. The terrain "popping" effect. I've probably seen this happen on every map at some point but definitely remember seeing it on caustic and alpine, but don't have any specifics.

[IraqiWalker]

Hi Karl. First off let me say :Happy New Year!

PGI has been doing great work this past year, and I can safely say the community's overall confidence in you guys has been raised a lot.

On to my question:

The Plan page hasn't received an update in a while now. Has it been abandoned?

TheCaptainJZ, on 29 December 2014 - 08:17 AM, said:

ah yes, I have seen this before on areas where the map is hilly and you are far away. The terrain "popping" effect. I've probably seen this happen on every map at some point but definitely remember seeing it on caustic and alpine, but don't have any specifics.

Happened to me on Sulfurous as well. I don't know if it's a problem with my toaster of a machine, or not? Does it happen to high end machines as well? (For the record, I run on the lowest settings possible with 1024 resolution, and I get about 20-25 fps now)

[Wintersdark]

Mr. Berg,

I understand this is a mechanics question, but hopefully you can get the right fellow to answer.

In the case of crit bonuses (in particular, let's use a Targeting Computer as an example) how is the bonus applied?

Say I have +25% crit. Normally, the crit roll looks like:
1-3: 3 crits
4-17: 2 crits
18-42: 1 crit
43-100: No crit

Does the bonus apply once, overall, spread between the options? Such as 3+25*3/42, 14+25*14/42 and 25+25*25/42? Or is it purely additive? So 3+25, 14+25 and 25+25? In this last instance, "no crit" would be pushed right off the table.

Or some other way? Math me, baby!

Edited by Wintersdark, 29 December 2014 - 09:05 PM.

[Wintersdark]

The LBX/MG crit bonus is flatly added to each category, which suggests the latter, but I suspect that's just an XML change to the base crit rates.

[Reno Blade]

Karl Berg, on 27 December 2014 - 05:33 PM, said:

Elo per mech model makes a good amount of sense. It's one of those things we'd probably have done already, had the simulations backed it up. I suspect the reason it appears to perform worse is due to large amounts of players bouncing between various trial mechs, and not playing enough games on a particular mech to properly converge their Elo.

More investigation and simulation is needed on this topic.

As for settings and associated resource costs, funnily enough I suspect our QA/Dev support department would have the best insight into that, if anyone. I don't know of any engineers that have spent time examining the different load points for each setting across a variety of hardware. Sorry, I don't have a concise answer for you

I always thought that using some up/down modificator per variant plus the class-based Elo would be a better idea .

Lets say you have heavy mech elo of 1600 and played most of your games in a Cataphract 3D (old meta).
Then you buy a new Thunderbolt without any skills.
- Starting to play the mech with 0/8 basics will give you a mod of -20% (example) of your class elo (1600 - 20% = 1280).
- leveling your new mech will bring the 20% mod down the further leveled it is.
- you would "lose" more ELO per loss and "gain" less ELO points per win compared to the amount you would get from playing with this ELO rating, which would change your "Class-ELO" less than without
(Loss based on higher rating, Gain based on lower rating from either Class-ELO or Modded rating)

As you perform closer to your "main" elo contributing mech, your elo-mod will converge towards 0% and you will have normal ELO rating loss/gains.
(similar to cadett ELO)

[Void Angel]

Off-topic: Elo is not an acronym. =)

As for the critical bonuses, I'd expect that the targeting computer bonus to crit would stack multplicatively, not additively; thus a hypothetical 30/20/10% chance for 1/2/3 crits would turn into a 37.5/25/12.5% chance. Otherwise, as noted, Clanners would be able to push non-criticals off the hit tables - resulting not only in Clan guns simply shredding equipment when striking exposed components, but enjoying a constant 15% damage bonus against internal structure as well. If this was happening, we should have noticed.

[Circles End]

I don't know if inverse kinematics has been touched in the last 100 pages, if it has, I'd appreciate a page number.

I am curious about why it was removed. According to the survey papers I've read, it isn't supposed to be very computationally heavy, so to me it seems like an odd thing to shave off. Was it to reduce server load? That's about the only reason I can figure out, since the server would obviously have to incorporate this in its simulation.

[Heffay]

Circles End, on 05 January 2015 - 11:02 AM, said:

I don't know if inverse kinematics has been touched in the last 100 pages, if it has, I'd appreciate a page number. I am curious about why it was removed. According to the survey papers I've read, it isn't supposed to be very computationally heavy, so to me it seems like an odd thing to shave off. Was it to reduce server load? That's about the only reason I can figure out, since the server would obviously have to incorporate this in its simulation.

Good question, and it would be nice to have it turned back on. But it probably does create a lot of extra traffic. 24 mechs, each with foot, toe and heel bones (multiple ones in some mechs) that need to be calculated by the client and that info sent to the server, then resent out to each of the clients, multiple times a second. Or if you let the client figure it out based on the geometry of the terrain under the foot when it lands, it could be an issue on the client side.

Ideally it could be a client side setting where you can choose to have your computer calculate the additional info.

[Arnold J Rimmer]

Heffay, on 05 January 2015 - 11:40 AM, said:

Ideally it could be a client side setting where you can choose to have your computer calculate the additional info.

Perhaps, but then you'd run into server authority problems. You'd be seeing legs in different positions than the server would, so hit detection would appear wonky. Love to have it back, though, but it's still a cosmetic thing at the end of the day.

[Heffay]

Arnold J Rimmer, on 06 January 2015 - 03:01 AM, said:

Perhaps, but then you'd run into server authority problems. You'd be seeing legs in different positions than the server would, so hit detection would appear wonky. Love to have it back, though, but it's still a cosmetic thing at the end of the day.

Since it would only apply to the foot, and in reality only affect the parts furthest from the pivot points, it may be an acceptable compromise. Have the hitbox follow the foot rotation, take a small hit in performance, and let the toes do their own thing. If people are aiming for toes and upset that they can no longer hit them from 880 meters away, I can live with that.

[Arnold J Rimmer]

Heffay, on 06 January 2015 - 04:35 AM, said:

Since it would only apply to the foot, and in reality only affect the parts furthest from the pivot points, it may be an acceptable compromise. Have the hitbox follow the foot rotation, take a small hit in performance, and let the toes do their own thing. If people are aiming for toes and upset that they can no longer hit them from 880 meters away, I can live with that.

I remember seeing a screenie of a Hunchback on the rim of the Caustic caldera - damn thing looked as if it were kneeling. Any time a mech is not facing directly up or down a slope, its legs are going to be in different positions relative to both each other and their neutral pose. IK is going to do that for any mech, client-side, and thus will not match up to where the server knows the legs actually are. It will cause problems, because it's not 'just the toes'.

[Circles End]

Hence why I mentioned the survey papers indicating that the computations weren't that expensive and thus shouldn't give the server a stroke. But I don't know for sure, and thus i ask

[Lordred]

Sir Karl.

I ask these questions as a photographer, and my desire to bring some of that to the game for screenshots.

is there a way to increase the Depth of Field?

By that I mean, as it stands, we appear to have off, or on. But I am looking to narrow the range at which focus is found, I would like to even be able to fine tune it.

Also, there are several cvars for motion blur that I have been playing with, but none of them seem to do anything.

I can fake focal length with post editing.
I can fake motion with post editing.

But is it possible to do this in engine?

Example:

Spoiler

The focus of the image is the car, the car is in focus, but that picture was not created from 1 frame out of 60, that picture was likely taken from a 1/200th-1/60th of a second while panning to follow the car.

the closer I can get to creating that in engine (for the screenshot thread) the less work I would have to spend in post production to acheive that.

DoF/Motion Blur, I would like to use it if possible.

More examples of DoF | Focal Length (Focus spot) vs F-Stop (Focus Width)

74mm @ 'near' focus 2F-Stop


74mm @ 'far' focus 2F-Stop


74mm @ 'mid' focus 22F-Stop


Add on to this proper frame blending to give us better appearance of motion, I think I could do some fun stuff.


/camera rant.

I know this kind of thing would have no impact on game play, but if we could do this, it might give the content creators some fun stuff to play with

[Rebas Kradd]

Shameless bump.

[Void Angel]

No need to bump it - Karl has always gotten back to us after long absences due to actually having work to do (or doing silly things like talk to his family at Christmastime."

[Lordred]

Second thing @ Karl.

Will there ever be a pass made on Forest colony?

I might just be me, but I think the map has been WAY to dark and muted compared to the beauty it had in closed beta.


Forest Colony now.



Forest Colony ideal? (for some, might be speaking for a minority, I do not honestly know)

[Cimarb]

I am on the fence about that. You technically could just turn up your gamma, could you not?

I honestly would prefer if the lighting was dynamic, though. Your first screenshot looks like a dusk battle, while the second one looks like a midday one. I like both in game, actually.

[Void Angel]

I dislike the look of Forest Colony as well. It's not the level of light so much as the dim, grainy appearance that makes it so hard to identify targets on the move - yet it looks... grimy, too, and not in a "wow, this neat 'Mech garage even has oil stains!" sort of way.

[Cimarb]

Void Angel, on 16 January 2015 - 06:33 PM, said:

I dislike the look of Forest Colony as well. It's not the level of light so much as the dim, grainy appearance that makes it so hard to identify targets on the move - yet it looks... grimy, too, and not in a "wow, this neat 'Mech garage even has oil stains!" sort of way.

It is actually one of my favorite maps.

[Rebas Kradd]

Cimarb, on 16 January 2015 - 06:49 PM, said:

It is actually one of my favorite maps.

Agreed. A lot of people want Battletech immersion, but when they're given realistically dank or grungy maps to play, their desire for immersion conveniently goes out the window, replaced by typical gamer desires.

Terra, for example. I hate it. It's just so damn hard to maneuver, and it's dark to boot. Not fun on a gamer level. But it's realistic. I can see a volcanically challenged planet looking just that way. And it's also the only map designed to actually require scouting by denying easy lines of sight. The rest of these maps take the need for scouting and absolutely bury it.