Sam Woodman's Hacx demos

Sam Woodman did a bunch of HMP demos for Hacx back in 1996/2000. table zip

I think I watched most of these without any trouble back when they were released, though my recollection may be hazy. However, I'm now finding that many of them desync. For example, hx13-927 and hx15-802 both desync early on. This is using dhacx.exe (i.e. Doom2.exe with hacx.deh applied) or any of the ports that ought to be able to play them back (Chocolate-Doom, Eternity and Prboom-plus 2.2.6.22). These desyncs occur at the same point and in the same way in all four engines on the two demos I mentioned.

I'm using the full registered version (i.e. the one made available for free by the authors) of Hacx 1.1 (hacx.wad = 22,102,300 bytes, date-stamp 09:40 16-Sept-1997). I have no problems playing back the TAS movie demo hacx3215, so I'm a bit puzzled by this. Do these demos work for anyone else? Any suggestions what the problem might be? (And no, I haven't rebuilt the nodes or screwed about with the wad in some other manner.)

does the TAS movie play back with the original dhacx.exe too?

if not that exploits a bug introduced in the ports ? or another obscure doom2.exe unknown incompatibility ?

The TAS movie demo was recorded with TASMBF, so it can't play back with anything that doesn't have MBF demo compatibility. I mentioned it in this context to demonstrate that my copy of hacx.wad probably isn't screwed up.

My problem is that demos apparently recorded with dhacx.exe (doom2.exe with hacx.deh applied) aren't playing back with dhacx.exe (or any of the ports that ought to support its demo format). The ports I mentioned are matching dhacx.exe's behaviour in these cases. They are not reporting any problems with the REJECT either (something that can cause same-exe desyncs, or ones that are OS-dependent).

They desync for me too with the 'real' game.
I seem to recall an email from Sam when he sent these to me stating that fact... although it is odd that I didn't annotate that somewhere.
I recall watching all of the demos, so I must have known about it then.
He also indicated that it would be nice if the project was finished -- UV demos and the other maps filled in, etc... :)

Opulent said:

They desync for me too with the 'real' game.
I seem to recall an email from Sam when he sent these to me stating that fact... although it is odd that I didn't annotate that somewhere.
I recall watching all of the demos, so I must have known about it then.

Odd indeed. Maybe I didn't watch too many of them back then (several more of the Maxes desync for me in addition to the two I mentioned - I tested them on my old Win98 notebook, just to check it wasn't an OS issue with XP), or maybe I didn't feel the need to investigate desyncs back then. I wonder what he did record them with.

Anyway, thanks for confirming that I'm not going mad or doing something stupid.

Opulent said:

He also indicated that it would be nice if the project was finished -- UV demos and the other maps filled in, etc... :)

With Prboom-plus now (hopefully) fully supporting the recording of Hacx demos, maybe there will be some progress in that respect. I might try some when I have (or rather, if I ever get :( ) some spare time.

Grazza said:
For example, hx13-927 and hx15-802 both desync early on. This is using dhacx.exe (i.e. Doom2.exe with hacx.deh applied) or any of the ports that ought to be able to play them back (Chocolate-Doom, Eternity and Prboom-plus 2.2.6.22). These desyncs occur at the same point and in the same way in all four engines on the two demos I mentioned.

It appears the cause of the desynchs is the DeHackEd patch; to watch those demos extract a patch from the copy of DHacx contained in the v1.1 public release package and use it instead of the included patch.

To be specific: the hit points for things 11, 16, 18 and 21 aren't the same in the patch and in the engine.

Yes, well spotted: they're not identical. The differences that have an impact on gameplay are:

By extracting from the dhacx.exe included in the files-only package:

Thing 11 (Chaingun Sargeant)
Hit points = 440

Thing 16 (Baron of Hell)
Hit points = 540

Thing 18 (Hell Knight)
Hit points = 1280

Thing 21 (Arachnotron)
Hit points = 840

In the separate deh file, included in the standard package and the dehacked-only package:

Thing 11 (Chaingun Sargeant)
Hit points = 400

Thing 16 (Baron of Hell)
Hit points = 450

Thing 18 (Hell Knight)
Hit points = 800

Thing 21 (Arachnotron)
Hit points = 600

So, the files-only version beefs up four of the monsters in terms of health, but otherwise is the same. I wonder which should be regarded as the "correct" version.

In Hacx, Thing 11 is the Monstruct (big fat thing with the uzi), Thing 16 the Terminatrix ("Arch-Viless"), Thing 18 is the Mecha Maniac (rocket-launching mechanized bastard) and Thing 21 is the Thorn Thing (big spider).

Edit: heh, a bit of simultaneous posting/editing.

As for which patch to use... Metabolist did kind of set a precedent with the demo set.

The EXE was built in '98, while the patch is apparently from '00.

I checked the shareware (DH-SW.EXE) and it coincides with the built EXE.

myk said:

the patch is apparently from '00.

The file date is probably due to the comment that was added at the start of it:

# This is the DEH file to be used with the registered version
# of HacX 1.1, available only at: http://www.sonic.net/nostromo

Perhaps only Nostromo can say which was the intended version. Maybe these monsters were deliberately toned down in version 1.1, but he accidentally included a different exe in the files-only pack.

I've noticed that three of the Max demos (on maps 07, 11 and 17) still desync in Prboom-plus and Eternity. They do work with Chocolate-Doom though. At first I thought this might be to do with the Monsters Infight thing (since it's something ChocDoom supports that other ports don't), but this isn't used in the Hacx dehacked patch (either version), so there must be some other explanation.

To save other people the effort, I've put a copy of the deh extracted from the exe in the files-only package here.

Grazza said:
Maybe these monsters were deliberately toned down in version 1.1, but he accidentally included a different exe in the files-only pack.

Hmm, it seems like something like that is quite the case; the multi-file distribution package builds an engine (with a new time stamp) using the DeHackEd patch with the lower hit point values.

The multi-file package was compressed in '97.

Hacx's demo1 and demo3 will desynch using the pre-built engine in the files only package.