PrBoom on Speed

Interesting stuff. I can see times when those parameters would come in handy...

Nice work on that episode run as well. I'm not familiar with how cheated runs are made (you don't have to avoid calling it that, btw ;) but it looks pretty optimized. Well done!

Calling it "cheated" suggests something dishonest/improper/underhand has been done, and leads to confusion. To those "in the know" it is obvious what it really means, but to newcomers to demos I'm sure it sounds very odd indeed.

"Tool-assisted" is the right term, IMO.

just a little tip for Andrey. I think it should be possible for you to improve that e1m8 time. If you kill the second baron while the first one is in the death animation it will open immediately but otherwise it will open after the death animation of the second baron.

I want demos for the other three episodes now.

And for Final Doom.

And every decent PWAD ever made.

So are those demos TAS? 14:41 Doom 2 speedrun would be hard enough just by splicing single-level demos together, but the page and the text file makes no mention of slow motion or demo resuming.

They are indeed TAS. Check out this thread (Scroll down a bit for the relevant posts covering this).

EDIT: The first thread where these runs were discussed might also be of interest.

Perhaps Andrey doesn't sufficiently stress these runs are TAS... Of course, if you've seen a lot of unassisted runs it's very obvious they're either built, or TAS, or both. If not, then, as Grazza so elegantly put it, it must be "confusing". There's a passage on the very download page that says "re-recorded after 24-th level", which should remove all doubts if the .txt's aren't clear! :-)

Kristian Ronge said:

Perhaps Andrey doesn't sufficiently stress these runs are TAS...

No, he doesn't. If a demo is tool assisted the text file should explicitly say so, I just watched the 14:41 run the other day and I figured it was TAS but I couldn't say for sure since nothing seemed to specify it was.

Bloodshedder said:

allows you to specify a point at which to skip forward when viewing demos.

Neat!

The best Andrey run yet.

It's a pity he uses that strafe50 hack though, it's like a regular run with turbo 110 or something. Not comparable with what's gone before.

Fredrik said:

I want demos for the other three episodes now.

Soon I'll try to record a demo of 4-th episode. I think that the result should be less than 3 minutes. But you should know - I do not like Doom1, as against doom2, Plutonia, Alien Vendetta, Eternall, HR2 and others fine megawads

And for Final Doom.

I tried, but have received desync at 3-th level (first level in 10 seconds)

Linguica said:

So are those demos TAS?
but the page and the text file makes no mention of slow motion or demo resuming.

Certainly it TAS. I thought that it clearly without mentions :)

As you see here, for outsider people it is certainly not obvious.

Ryback said:

The best Andrey run yet.
It's a pity he uses that strafe50 hack though, it's like a regular run with turbo 110 or something. Not comparable with what's gone before.

I can not use Strafe50 at turns, but I do not think, that it will change time. I well see and I feel a difference between competn demos (such as 30uv2609 by Radek Pecka, 30nm3413 by Henning Skogsto, a plenty of examples from Drew DeVore "stx-Vile") and my runs. But I wanted to record:
1) Faster than existing by the current moment
2) Completely compatible with original doom
and without the message "green is turbo!" at viewing

If I did not use permanent strafe50, be sure - time would be same, but the demos would appear much later (as it more hardly) or would not appear never.

I used very low sensitivity and tried to turn very seldom (run by back is a style of my demos :)). It seems to me that looks much more pleasantly than 30uv1441.lmp

Vincent Catalaа should confirm it after viewing :))

If you want the *fastest* time, you should consider that E1M1 has been done in 8 seconds in a demo built manually frame by frame... that's without grabbing a shotgun though, but there might be other maps where one could optimize similarly. For example, your 1:01 on E1M7 simply must be compressible down to 0:59 :)

Adam Hegyi said:

As you see here, for outsider people it is certainly not obvious.

But in 30uv1441.txt and in e1uv-405.txt it is written that
demo recorded using PRBoom 2.2.x with changes of source codes and Strafe50 is always ON.
I do not think that people from http://www.gamershell.com/download_5369 know that such TAS and slowmo.

Fredrik said:

If you want the *fastest* time, you should consider that E1M1 has been done in 8 seconds in a demo built manually frame by frame... that's without grabbing a shotgun though, but there might be other maps where one could optimize similarly. For example, your 1:01 on E1M7 simply must be compressible down to 0:59 :)

Quite probably. I badly know Ultimate Doom. Write to me please times for each level of the fourth episode which you would like to see.

P.S.
I knew that 8 seconds is possible, as since my time is 9.20 instead of 9.97

Andrey, I did not watch your demo yet, but the time looks great.

By the way, you should insist more on saying its TAS in your text files. And also say it's not possible to do that without heavy software assitance and manipulations.

Another point, a question especially for AdamH, I hope this new TAS prBoom makes demos that can be detected for the Compet-N... or it might mean the end of the Compet-N...

E1M4 is probably possible in 11 seconds with the strafe50 trick where you simply run to the exit and jump over instead of raising the bridge.

actually, I think 12 has been proved to be optimum... besides, that lift trick is probably the coolest trick in UD.
:)

I strongly suspect that 11 is possible on e1m4 because Adam's 12.82 doesn't use strafe50, except for the jump of course. However, this 12-second run of Budko's on an alternative route is cool.

The whole movie itself is very neat too, I enjoyed it a lot. Surely some of those times could be squeezed even lower, but this is the first TAS speed movie of the episode that I know of, and a very welcome one.

EDIT: Btw I think the e2m7 rocket jump is the coolest trick in udoom. =)

This isn't the first TAS speed movie of the episode, there's ep1x428 too. As for E1M4, I think someone built a demo and got 12 seconds, but I could be wrong about that.

By the way, the 10 maps max run is extremely impressive too. I can't wait to see that continued.

Just got through watching that E1 4:05 speedrun, damned impressive I must say. :)

well, I've been proven wrong many times before, and if you and Xit say so, I'm more of a believer.
http://www.doomworld.com/sda/tas/c1m4x012.zip is pretty conclusive to me though. Machata coop: that's damn fast. and 12 1/2 seconds... a long way from 11.

adamH nomonsters time is 12.34 and I bet he does not use strafe50 except for the jump.

So I think 11 is possible in nomo... monsters might trickier... but with some luck you can prolly avoid getting stuck. But strafe50 handling makes monsters dodging harder.

Fredrik said:

By the way, the 10 maps max run is extremely impressive too. I can't wait to see that continued.

Thank. I believed that it has interested nobody.
To continue this record becomes heavier and heavier (with my PRBoom modification). After each restart (thousand restarts) it is necessary to wait about 4-5 seconds. And if me so abuse for "very slow" e1uv-405.lmp it is difficult to present that would be with max run ;))
And about my .txt
I have not written all that it is possible (it is necessary?) because I don`t know the English. Who is dissatisfied with contents my e1uv-405.txt or other - let sends to me your variant (Because I awfully write in English) I even allow to correct grammatic mistakes. If someone has nothing to do :))

But strafe50 handling makes monsters dodging harder

But you can try with my prboom. (have Strafe50 setting)

By your comments my time for the first episode should be 4:00 :) I very much would want to receive less than 3 minutes in the fourth episode

xit-vono said:

E1M4 is probably possible in 11 seconds with the strafe50 trick where you simply run to the exit and jump over instead of raising the bridge.

I think that I am ready will share with you the full version of my prboom (with rerecord feature). You are ready to help me with second episode? (I love your demos. Especially maxruns of long and very long levels)

about strafe50 monster dodging entryway said:

But you can try with my prboom. (have Strafe50 setting)

By your comments my time for the first episode should be 4:00 :) I very much would want to receive less than 3 minutes in the fourth episode

I was speaking about regular hand-made Compet-N run, with regular old DOS engine and without special mouse driver :)

With your modified engine it's rather easy to do :)

Have you seen Sedlo's TAS run in plutonia ? (from map01 to 15)
Sedlo is active again, so he might finish it on his own now :)
I was also thinking about doing some TAS AV movie. I dont have too much time... but I guesss I could help.

about Xit-Vono entryway said:

I think that I am ready will share with you the full version of my prboom (with rerecord feature). You are ready to help me with second episode? (I love your demos. Especially maxruns of long and very long levels)

Xit-Vono does not play max, he plays mostly Tyson :) and some bits of speedrunning.

Actually he's getting used to doom2 max since we are trying to record a full cooperative Doom2 movie with all monsters killed and all secrets found. It's of course for the Compet-N, recorded with good old doom2.exe and obviously no deaths allowed.
We are working on it for a month now (something like 6h a week on sundays), we know map01 to map20 pretty well now... other maps are done a tad slower.
Most of times one of us miss a secret, leave a monster alive or just dies :) So we keep retrying...

Xit : maybe next week we will get this sucker ! ;)