TAS demos: Slow-Motion or Built?

[Andy Olivera]

Greetings all. I was just curious as to peoples preferences are regarding TAS demos? I have a couple old TAS collections that I never finished and was thinking about resuming them, but with the advent of DRE I'm torn between whether to start over and go full built, or just continue with Slow-Motion.

My take on the issue is, while built demos are very impressive, even when limiting the jerkiness by avoiding massive turn values, the split second accuracy makes them hard to follow and, hence, less entertaining.

Is there a general consensus?

[Grazza]

I would view built demos as an entirely separate category from TAS.

TAS should (and does!) mean assisted - i.e. the player is still controlling the player in the usual way, but assisted by slow-motion and the ability to re-record, but without directly manipulating the play at a tic-by-tic level in a text editor (or equivalent system).

As for a preference, both are good and worth seeing if done well.

[tatsurd-cacocaco]

For example,

====================================================
GF50 SL40 TR26          #   677 (19.34s) 
GF50 SL50 U0x08         #   678 (19.37s) Unknown Use
GF50 SL50 U0x08         #   679 (19.39s) Unknown Use
GF50 SL50               #   680 (19.42s) 
GF50 SL50               #   681 (19.45s) 
GF50 SL50               #   682 (19.48s) 
GF50 SL50 U0x08         #   683 (19.51s) Unknown Use
GF50 SL50               #   684 (19.54s) 
GF50 SL50               #   685 (19.57s) 
GF50 SL50               #   686 (19.60s) 
GF50 SL50               #   687 (19.62s) 
GF50 SL50 U0x10         #   688 (19.65s) Unknown Use
GF50 SL50               #   689 (19.68s) 
GF50 SL50               #   690 (19.71s) 
GF50 SL50               #   691 (19.74s) 
GF50 SL50 U0x08         #   692 (19.77s) Unknown Use
GF50 SL50 U0x10         #   693 (19.80s) Unknown Use
GF50 SL50               #   694 (19.82s) 
GF50 SL50               #   695 (19.85s) 
GF50 SL50               #   696 (19.88s) 
GF50 SL50 U0x08         #   697 (19.91s) Unknown Use
GF50 SL50               #   698 (19.94s) 
GF50 SL50 U0x10         #   699 (19.97s) Unknown Use
=====================================================

I don't like such TAS demos very much, but the attitude to challenge the limit to the time is very great, yeah.

I wouldn't make such TAS demos, so my TAS demos might not be always too hard work. Actually, I spent only 4 or 5 hours to make my lv01x030 (Not editing frame-by-frame), but jongo spent 16 hours to built his lv01x (Editing frame-by-frame). I think 29 sec is possible and perfect. However, it is nearly impossible by my way, and I don't intend to make it.

Though some people might criticize my TAS demos as half-baked TAS, I don't want to make the TAS which I don't enjoy making, and there are a lot of undone magawads.

[ultdoomer]

I think it's best if you only build the starts of levels, to start out at the perfect angle, and then use TAS (slow-motion / re-recording) -- like Andrey's demos.

[Andy Olivera]

ultdoomer: That was my thought, as well: use mostly Slow-Motion, but build the tricky parts(line skipping, glides, etc.).

[abyrvalg]

Yeh, this is quite a good decision. Actually, that's how i recorded most of my TAS things. And building isn't worth your time unless you already know the map very good (i.e. know all tricky spots where the time could be lost, and have an idea on how not only to avoid this loss with building, but to maximise the profit). But it's way to go when you want to make a perfect run (like the one RamboBones is building now. I don't know if it is possible to reach these times with a simple slowmotioned running).