Back

on the genesis of the project:

Konami's last Gradius game was the 2010 browser game Gradius Arc, and while they've reissued games via Arcade Archives etc, they've been thinking about releasing something new since 2017 or so

the 40th anniversary presented an opportunity to make something happen—ideally with something brand-new & not just legacy—so Ueno brought it up to M2 and ultimately settled on Origins Collection

on M2's side, they've been pitching Gradius-related projects for quite a while but never got the go-ahead for anything, so they were surprised & happy to get the pitch from Konami

they know Gradius is a premier title & that the fans (not to mention folk within M2) wouldn't tolerate something substandard, so they've been working to make sure they meet expectations

on the all-new Salamander III: M2 made the ReBirth series for Konami back in '08 and have regularly maintained those games & dev environment since, so they proposed that a new game built to that spec would be very cost-efficient, and Ueno qas quick to approve

as mentioned previously, Ueno really wanted to include a new game, and it personally bugged him that Gradius Arc, a non-STG, was the last official Gradius game, so he really wanted to rectify that

why, then, is the new game Salamander III and not Gradius?

one issue was the naming convention: what do they call it? for reasons Horii feels hardcore fans are aware of, "Gradius VI" wasn't an option

Horii also cites checkpoints as a hurdle: a game titled Gradius would need to be designed around checkpoints & recovery patterns, but their schedule didn't allow for the amount of balancing required to do justice to that style of game

in Ueno's words, this simple wasn't the time: a brand-new mainline Gradius presents many hurdles and they'd really have to get their ducks in a row before taking on that challenge

Horii figured a new Salamander would present them with fewer hurdles, but development wasn't smooth... he promises to talk about the development of Salamander III, but that talk's being held back until release day (August 7)

on Origins Collection:

M2's outline to Konami was actually a collection of console ports—most of the arcade games had just been reissued via Anniversary Arcade Collection, after all—but eventually people within Konami came to realise they could produce a more definitive version of the arcade games with M2, and so pushed in that direction

Horii's like, well, technology's always improving, so if we can make people happy with better ports, that's what matters

on the feature set:

it has the typical stuff (savestates, rewind, etc) but the big emphasis is on the training mode, which lets you try specific sections with granular state options (loops, powerups, etc)

using this, you should be able to clear anything that might have tripped you up before... except, as Kubota's quick to point out, Gradius III—if you practice the cube rush every day for a year then maaaaaybe...

on the sheer volume of the collection:

Ueno knew there were multiple revisions of certain games & somewhat selfishly requested they include every version possible, but didn't really grasp what that meant in totality

M2's policy has always been to include everything they can—that's particularly true of the ShotTriggers line, which is made with the directive of "no matter what, leave nothing on the table"

on why the collection only goes up to III:

they initially intended to include Gradius IV & even Solar Assault—once they realistically reviewed the schedule, they knew Solar Assault wasn't going to happen (not only is it a 3D game but also a game that used a large moving cabinet, which presents its own hurdles re: reproduction), but they did keep trying to find a way to include Gradius IV til the end, but regretfully had to let it go

the issues with Gradius IV were surpassing the not-great PS2/PSP version, but also raw technical hurdles

Gradius IV used PowerPC/Voodoo-based hardware that was tough to accurately reproduce (which includes accurate slowdown) & given the limited schedule, making it work on Switch in particular (the biggest projected platform) wasn't feasible, so they had to shelve it

they're not happy about it, and Horii does mention that it'll be back on the table if and when they do something else, so be nice

Ueno gives props to M2 for giving it their all, and says that if they hadn't been on top of things, even Salamander 2 might not have made it in time

they sort of figured that the Salamander games would be fine, given they'd already emulated them for PSP (& still had the same programmer who handled them back then, somehow managing 60FPS on PSP) but Salamander 2 proved to be troublesome even now on contemporary hardware, to an extent that Horii wasn't aware of until this press run

on M2's staff culture:

Horii says the work culture's fairly permissive, and people who are on top of their work can walk out during the day to catch a movie or whatever (which Kawachi quickly points out is very position/situation-dependent)

M2 attracts people who have the dreams of one day working on classic titles, Gradius included—or as Horii puts it, there are lots of staff who own the arcade boards & think, "if we port these games, I'll be able to play them without pulling out my PCBs"

on their particular focus when porting the games:

priority one was accurate reproduction, and from there they focused on ease-of-play additions

Kawachi points out that, in the ~20yrs since working on the PSP collection, they've made big strides both in technology but also ease of verifying accuracy and A/B-ing with original hardware—they have internal methods of simultaneous input-matching on original hardware vs. target platform, per-frame analysis of lengthy attract mode recordings, etc etc

not only have they developed technical processes to ensure accuracy but they're also staffed with people who intimately know and understand the games and can perform intuitive checks for accurate slowdown and so on—seasoned Gradius players know they can induce slowdown with the laser, for example, so they're able to go in and manually test and verify all the fringe cases for accurate performance

in porting these games, Ueno had to search Konami's archives & elsewhere for not just the ROMs/hardware but any official dev docs & other material that might exist, and then send that through to M2 for analysis/verification

they uncovered a massive amount of material; rather than sending it all to M2, they had M2 come to them and made copies of whatever seemed relevant, as M2 couldn't possibly house it all themselves

the interview asks about Konami's corporate culture re: preservation/archival: Ueno says it differs from game to game & not everything's in great condition, but he thinks they have things relatively well-stored (not that he knows how other companies handle things, so "relative" to what, he can't say)

sourcing documentation was hard because everything was split between different depts & nobody knew precisely what they had, so corralling everything required a lot of coordination

Ueno cites his experience working in the soundtrack dept. as a big help: around 2020, he started helping with CD production & selfishly took over at the retro producer—he wasn't able to buy the big STG/Castlevania box sets from years prior, so he made new versions himself

he particularly loves the Dracula XX (SFC) version of the st.1 theme & was happy to re-record that—incidentally, M2 made him the SE-less ROM he used to re-record those tunes

it was through the process of putting together those box sets that Ueno developed not only the contacts but the know-how that'd later allow him to gather everything for Gradius Origins Collection—without that foundation, no matter how enthusiastic he would've been, he thinks he might've been more likely to compromise & not go the extra mile if he didn't come in knowing precisely what might plausibly be uncovered and how to get it

the biggest discovery was the legendary AM Show pre-release build of Gradius III—Horii actually played this back in the day, and recalls that he spent an entire day in line to try it and only got to play it twice

funnily enough, it was M2's youngest member—a rookie programmer in their 20s named Kaito Higashi—who was most gung-ho about possibly including the AM Show version, and insisted to Ueno that they look for it, even if the possibility of finding it was tiny

Ueno checked their internal database & all their recorded PCBs and didn't turn up anything of note, but going through the archived ROMs for each game turned up something interesting...

most of the ROMS were easy to immediately identify, but Higashi excitedly proclaimed that they couldn't figure out one particular set—the graphics were all corrupted, and by mentally troubleshooting the issue, they realised the stage itself must be different to that of the final game...

Hiroshi was thrilled—according to Horii, one of the reasons he'd joined the company was in the hopes of someday unearthing that specific version he'd read about in Gamest

Kubota says that even once work on emulating the build was in full swing, Higashi would walk over to his desk and prattle to him for 30~40mins at a time about all the differences: "the sand in this section's animated in this version", "you can see the stars in the background at x", etc etc

Konami's games of that era were built using UNIX workstations that couldn't be recreated, but Higashi worked to reproduce as much of the old dev environment as possible & produced the rest from scratch

from there, they A/B'd their version with old production/dev milestone docs, as well as the source code that was unearthed during the making of Gradius III & IV for PS2, and got it to the point where they were confident in presenting it to the world as the bona fide AM Show version

that said, it seems that there may have been minor fixes between the day-1 and day-2 AM Show build, and Ueno posits that the version put on location immediately before AM Show may have been different again, so while this is assuredly authentic & warrants being presented as the AM Show version, it might not be the absolute one and only version produced for/around the event

on the content of the AM Show version:

it has 7 stages vs. the final game's 10; some are very similar but many differ significantly—Horii explains that the AM Show build was made before any degree of balancing had taken place, so the stages start off reasonable but tend to devolve to the point where they clearly weren't intended to be release-worthy

even so, Higashi happily cleared em, just to prove they're theoretically beatable—at its worst, it might even be harder than the released version

Ueno talks about how much of the internet, and even his own memories, claimed the AM Show version only had 3 stages, so he was happily surprised to find out it had more—that one Gamest article everyone read said the build had 7 stages, even, but he supposes that people just mentally "corrected" to 3 stages due to Gamest only providing images of the first three stages

ofc, according to the internet, the AM Show build no longer existed, so the moment they determined they had it, Ueno OK'd it

when they found this build, Ueno thought the collection would've been justified even without the all-new Salamander III

Horii hypes people up by saying they need to see the cube section of the AM Show version, which is apparently completely unhinged—there were serious discussions about whether they should present it as is vs. making changes to make it less insane (especially considering it clearly wasn't intended for public consumption), but in the end it's presented in all its raw glory

the new invincibility assist feature means anyone can make it there, but clearing it legit is a whole other story

that said, Higashi did record a no-miss video and sent it through to Ueno for feedback—Ueno was like, what could I possibly offer in response to this, but he also had Horii's words in his mind: 'you cannot use this guy as any sort of benchmark"

on other non-AM Show-related discoveries:

when Kawachi went with M2's design leader to inspect Konami's archives, they remember being struck by an unused design for a man-faced rock (not moai) from the original Gradius; this enemy ended up appearing in Parodius, so finding that design demonstrated just how far back some of their ideas went

Kawachi was also intrigued by their design docs for displaying scrolling stars in the BGs; some PCBs used dedicated chips for this, but they did it manually