Seems like the problem is the enemy rng has completely changed, despite every action being exactly the same save for the actual boss fight. Unless there is some way to copy paste the rng values, I don't think this method will work.
I would like to hear some suggestions but I have currently a plan if all fails. It's not what I'd like, as it will not be the proper way of doing things but I think it might work for the intended purpose.
I have already registered the frame count difference between the newly recorded fight and the previous one. I will do the same with one I still have to record. I can then add the values of the both frame counts so I can find out how many more additional (or most likely subtractive) frames it has. Afterwards, I will make another recording where I clear all the enemies and make another count. Let's say that there a total less 2000 frames from the newly recorded boss fights and the part where I clear all the enemies and return is 3000 frames. I can then, in the original movie file, skip clearing all the enemies and return in exactly 1000 frames (gaining the 2000 frames from the boss fights by not clearing all the enemies) so that the timer aligns to the expected time from the new boss fights. Unfortunately this means that I need 4 different files (the original file, the file with new boss fight 1 & 2 and the file where I clear all the enemies) and will need to break apart the whole thing into 7 different appends before creating the finished video (I fear keyframes in VirtualDub might make my work harder but if it's anything like FRAPs encoding it should be fine since every frame is a keyframe - I imagine this is where the codec chosen comes into play).
What do you guys think?
This is the first frame when the door animation shows. However, the previous frame is highlighted in red. Should I choose that frame instead?
Since the header is 2 lines does this mean that frame 1 is on row 3?Yeah, remember to check at the top how many lines the header is (the stuff before the first frame of input). If the header is 20 lines, then frame 1 is on row 21 instead of 1.
It's better to copy everything after frame 110120 in the original over to the new file. That way you have a movie guaranteed to play back in-case the edit would desync. Otherwise you got it correct. Obviously use names for the files so you can tell them apart.
Regarding optimization, if you deem it optimized enough you could always try to submit it, but it may end up rejected if it's considered too in-optimal. Submission is no requirement though, it's entirely up to you to choose if you want to or not.
Let's see if I understand the process correctly. First I need to create a backup file of my PSXjin movie which contains all of my progress. Then, go to the part with the boss fight which I wish to remake and record and overwrite the original file.
Now I have two files. The starting point of the part I want to replace should be the same for both of them since they are identical - and I'll be using the door loading animations before the boss fight as a reference. Let's call it the first frame and pretend it's frame 101102 for both files.
The second frame (the part which ends the remade segment) that I'll pick up from the original file and the new file will obviously be different because the segments will have different lengths. This means that I need to find the exact frame (using frame advance) where both files can be matched (such as the door loading animation on the way back I previously used for reference point). If, say, it's frame 110120 in the original file but frame 110000 in the new file, I'll copy everything from 101102 to 110000 from the new file over 101102 to 110120 to the original file. Is this correct?
Unfortunately I could not do full optimization for three reasons:
1. Because this will be a lengthy challenge. I have looked at the final timer (from an incomplete play) and my guess is somewhere near the 2:2X:XX mark. Assuming there's 60 frames per second (which I think is the case here) that's roughly 504000 frames.
2. Because Resident Evil is a 3D game that relies on pre-rendered backgrounds making it very difficult to establish what the best frame alinement to navigate through sections in as little time as possible. Plus I'd have to account for different enemy RNGs in combat for the least time possible. It's possible, but it would take months or maybe even years (depending on time available and interest) to get it done and I'm not willing to invest that much time in it.
3. Because time is not the core factor of this playthrough. I did my best to navigate in the least time consuming way possible (several attempts until I was satisfied with it) and changed the speed to drastic lows for inventory navigation, item storage and puzzles but this is all about the challenge itself and the combat against the enemies. The purpose is to see how the enemy behaves to different situations, what patterns the A.I. is coded and programmed with and what the player can do without compromising its safety. It's a design study, like previously mentioned.
I find it an interesting idea since no one has actually attempted anything like this and I did have some similar requests after I presented my rudimentary TAS of Extreme Battle LV.3 in RE2DS of No Weapon & No Damage (that one done only with savestates & loadstates during door loading animations - although with the clear purpose of showing that the challenge was actually possible despite how strikingly difficult it is). I'll upload it to YouTube nonetheless, but I'd be more than happy to see it on TASVideos too.
I assume each different part is a segment of the video. How do I determine which part corresponds to what I want to edit and how do I change it?
And thank you for that information regarding the codecs. I'll definitely check out that section to find out what suits me more. I could submit my TAS here but it will not be a speed-run so I'm not sure if it will get accepted. It's a Knife Only All Clear No Damage - a study of the mechanics and design of Resident Evil's Director Cut. It's mostly informative and educational for players of the series who have probably never seen many things depicted in it. Some things are within the reach of human ability (and that was one of the reasons why I choose speed 50% for navigation and combat, speed 36% for inventory and puzzles and speed 25%/16% for storage and item checking/combination) but some things require an insane amount of precision and luck so they are unlikely to be performed by a human playing on normal conditions.