The Bubble Bobble Fiasco

[5MinuteRetro]

The biggest embarrassment of all, though, came when New Computer Express had some abject moron...

Good hair though.

[jpnz]

The biggest embarrassment of all, though, came when New Computer Express had some abject moron cover the budget re-release and somehow manage to assert that the 128K version didn't have any music. (Having presumably loaded the 48K version into a Speccy 128 by mistake.)

Perhaps the issue with the music is not your fault

There's a noticeable delay after you select the player controls, and what seems to be happening there (for those interested - starting at 0xfac3) is a timing check - the end result being if you've loaded the tape on a +3 then the music is explicitly switched off. Odd that there's no problem with the disk version though

Anyway, a message in the code

Spoiler

Won't mention anything about the haircut... epic

[TMD2003]

I was putting together a Bubble Bobble compilation on my Retropie t'other day, which throws up many things of interest, but just for funz I thought I'd see how the rather fine Speccy version - surely one of the most accurate, colourful and playable coin-op ports ever on the machine, other than only getting 80 levels out of 100 in - was received by magazines back in the day. And lordy, what a shambles.

Crash was the least embarrassing, with a reasonably informed Crash Smash review. Sinclair User had some total custard-brain apparently dragged out of remedial English at the local primary school who described it as "a bit like Pac-Man" , which is true in the same sense that Doom is a bit like Pac-Man because you run around a maze collecting stuff.

Call me a total custard-brain from the local primary school if you must - though bear in mind I may retaliate with an interesting variety of sharp objects - but I've always found Bubble Bobble to be one of the most gratingly annoying games I ever owned back in the day. And of course, here we have a problem - because the game was the work of the Follin brothers, who are revered as some kind of gods, along with the Stampers and Sir Clive himself (I suppose Harry S. Price would be the equivalent of Satan). So I stand to be excommunicated from Spectrumland forever for saying this... but I'm going to anyway. This is a game that I would like to like so much more than I can, put it that way.

I don't particularly care for the ugly outline representation of Bub and Bob - surely there must have been a better way, and I would assume this was Mike Follin's work. The Spectrum ROM font looks out of place for a game programmed in 1987, especially the double-height version on the title screen, but that might have been because they ran right out of memory, so it'd have been an acceptable compromise in the 48K version. But still, why not include an arcade-machine-style font in the 128K version, which, as far as I can tell, only added the soundtrack?

Oh, yes, the soundtrack, and that's where I fall foul of Tim Follin. Straight up I don't like the jumping sound effects, which I would compare to having a dentist's drill inserted in my ears. But what really sets me reaching for the mute button is the 128K music. In a way, I can see what he was trying to do - a valiant attempt at the jangly-bells sound effects in the arcade original. But for some reason, every single note of the jangly-bells is off-key and if there's one thing I'm going to notice and which will drive me mad, it's that. Think of it - somewhat incongruously - as if the main soundtrack was represented by BEEP statements in which the notes are the expected whole number values (i.e. 24, 23, 21, 19, 23, 21, 19, 17, etc.). The jangly-bells should have been an octave above those (i.e. 36, 35, 33, 31...) but every note missed the mark by the same small amount (i.e. 35.9, 34.9, 32.9, 30.9...). What might have been...

So I went on YouTube to look for other versions of Bubble Bobble and what do you know, there's a helpful video that shows far more versions than I will ever need:

[media]https://www.youtube.com/watch?v=UgynBsgwSyI[/media]

The arcade original is clearly what they were all aiming at, and my first port of call was the Amstrad CPC version - it was a machine with a Z80 and an AY-3-8910, so it stands to reason that it should be the closest comparison. I wasn't surprised to find it used the Commie-style double-pixel graphics, but I was surprised to find it had no AY soundtrack. Was it only available on the CPC464 and, again, the Follins ran out of memory for it? Was there never a specific CPC6128 version? I'm not an Amstrad expert.

At least I thought it was worth trying the Atari ST version, seeing as that had an upgraded AY rather than the Amiga's SID chip, but for some reason the ST version sounds all wrong on that video (I think it's played at the wrong speed and framerate and comes out sounding more Amiga-esque). This video shows off the ST version, and though it's absolutely nailed the jangly-bells effect on the soundtrack without being off-key, it suffers from the usual ST problem in complicated soundtrack whereby the SID on the Amiga would be playing a chord, but the ST has to compromise and play the invididual notes rapidly like an arpeggio, or maybe an Irish diddly-diddly-diddly jig played at twice the speed it's supposed to be. So this wasn't the holy grail either.

Of course I looked at the Commie and Amiga for comparison, and to nobody's surprise it uses the ubiquitous SID effect that sounds, to my ears at least, like an unoiled garden strimmer. (Having seen what the SID effects look like on an oscilloscope I now know that this is created by a square wave that's "more up than down" or "more down than up" - it was quite capable of making regular square waves, but maybe the programmers of the day thought it sounded too much like the AY and they wanted to swank off the extra toys they'd been given.)

The unreleased BBC micro version is pretty good, though I assume it wouldn't run on an Acorn Electron. The Apple II version is utterly horrible, and I'd take the 48K Spectrum version ahead of it any day - and the Atari 8-bits didn't get a version at all, that I can find.

But I think the clue as to the holy grail was the NES and Sega Master System versions, both of which I'd assume were programmed in-house by Taito - after all, I'd expect Japanese programmers to be familiar with Japanese equipment. And this leads us to the MSX version, also presumably Taito's own work. Yes, yes, I know, even the MSX1 was more powerful than our favourite machine, and to compromise the comparison further, this is an MSX2 game... but it had a Z80 and an AY-3-8910, so is there any chance that the Spectrum soundtrack could have sounded like this, or am I hoping for too much?

Then there's Rainbow Islands... and I have absolutely no problems with that game, nauseating cuteness aside. That one nailed what a Spectrum coin-op conversion could look and sound like, even if it took three years from the release of the arcade original. Bubble Bobble only took one year and seemed rather rushed in comparison. But I suppose that's the benefit of hindsight.

I wonder... is anyone up for a remake? We had a thread on that a while ago, "what games would you like to see remade". There was clearly someone out there who was sorely disappointed at the Amstrad CPC version, and made Bubble Bobble 4 CPC in 2014, which blows the 1987 version into the weeds.

[akeley]

Fun write up, though I wouldn't say it's as bad as you paint it. Most of your gripes are just some small niggles, most of them actually praise the game after all. YS absence is indeed inexcusable, though there must be some reason for it too.

More importantly the mags back in the day were really brilliant, despite their occasional failings. I love their half-serious, half-DIY-punk-zine ethos and the fact not everybody was some super serious journo, so, yeah, the custard brains brought a bit of colour to this scene as well.

And Slaine is a Megagame, through & through. Fite me!

[mention]TMD2003[/mention] I don't particularly agree with your nitpicking either, I think ZX version is really good and of course it's really not possible to do a 1:1 arcade port to inferior hardware. More important by the way is playability, something which is hard to experience from watching vids on the tubes.

In any case, your wish seems to have been answered recently in the form of the allegedly excellent Tiny Bobble for Amiga.

[DouglasReynholm]

The biggest embarrassment of all, though, came when New Computer Express had some abject moron cover the budget re-release and somehow manage to assert that the 128K version didn't have any music. (Having presumably loaded the 48K version into a Speccy 128 by mistake.)

Perhaps the issue with the music is not your fault

There's a noticeable delay after you select the player controls, and what seems to be happening there (for those interested - starting at 0xfac3) is a timing check - the end result being if you've loaded the tape on a +3 then the music is explicitly switched off. Odd that there's no problem with the disk version though

Anyway, a message in the code

Spoiler

Won't mention anything about the haircut... epic

I may be biased, as it is one of my favourite all time games (still play it at least once a week), but how can you 'hate' it? It's such a fun game. Oh well, each to their own.

I would love someone to disassemble the original arcade code and work out its state machine logic for how bonus items are awarded, as I think that's what a lot of the conversions didn't get right. As I say, I still play it regularly, and you can see patterns set up across multiple different games. Hard reset the code, and the patterns go away. But even after 34 odd years it's very hard to come up with an 'if this then that' set of rules to play by, I think even the player position on the screen can influence matters..

Edit: I should add I'm talking about playing it on MAME with the original ROM. Haven't played the Speccy version since 87 as while it wasn't a bad conversion at all, nothing holds up to the original version, in my humble opinion.

[DouglasReynholm]

[mention]TMD2003[/mention]

Having seen what the SID effects look like on an oscilloscope I now know that this is created by a square wave that's "more up than down" or "more down than up"

The technical term for this is PWM, or Pulse Width Modulation, a common technique used on most synthesizers, alongside triangle, ramp (sawtooth), and plain 50/50 duty cycle square waves. Pure sine waves are generally less common as an audio oscillator, as they have no harmonics, but are used to modulate other aspects of a sound, for example PWM!

[TMD2003]

The technical term for this is PWM, or Pulse Width Modulation, a common technique used on most synthesizers, alongside triangle, ramp (sawtooth), and plain 50/50 duty cycle square waves. Pure sine waves are generally less common as an audio oscillator, as they have no harmonics, but are used to modulate other aspects of a sound, for example PWM!

Now you mention it, I have heard of that term, most likely on the same video where I saw what the SID waveform looked like.

Is it impossible on the AY, or was it just a lot easier on the SID so that all the Commie/Amiga programmers used it all the time?

[AndyC]

The biggest embarrassment of all, though, came when New Computer Express had some abject moron cover the budget re-release and somehow manage to assert that the 128K version didn't have any music. (Having presumably loaded the 48K version into a Speccy 128 by mistake.)

Perhaps the issue with the music is not your fault

There's a noticeable delay after you select the player controls, and what seems to be happening there (for those interested - starting at 0xfac3) is a timing check - the end result being if you've loaded the tape on a +3 then the music is explicitly switched off. Odd that there's no problem with the disk version though

Anyway, a message in the code

Spoiler

Won't mention anything about the haircut... epic

Interesting. I wonder if this was a deliberate fudge to counter for the terrible sound on +3 machines. I guess if it is there might be differences elsewhere in the +3 version to attempt to fix the sound in some other way (given that a disk version would certainly be loaded on a +3)

[ZXDunny]

In any case, your wish seems to have been answered recently in the form of the allegedly excellent Tiny Bobble for Amiga.

Tiny Bobble really is bloody excellent. It's the closest to the arcade I've yet seen in a home computer conversion. That it runs on an OCS A500 is gobsmacking.

[StooB]

Interesting. I wonder if this was a deliberate fudge to counter for the terrible sound on +3 machines. I guess if it is there might be differences elsewhere in the +3 version to attempt to fix the sound in some other way (given that a disk version would certainly be loaded on a +3)

Bubble Bobble was released around September 1987, only a couple of months after the +3. The adverts only mention a cassette version for the Spectrum, with disk versions for the C64 and Amstrad. It seems likely the disk version came out some time later.

[1024MAK]

@TMD2003

Having seen what the SID effects look like on an oscilloscope I now know that this is created by a square wave that's "more up than down" or "more down than up"

The technical term for this is PWM, or Pulse Width Modulation, a common technique used on most synthesizers, alongside triangle, ramp (sawtooth), and plain 50/50 duty cycle square waves. Pure sine waves are generally less common as an audio oscillator, as they have no harmonics, but are used to modulate other aspects of a sound, for example PWM!

Ahh, but the frequency of the level switching has to high (fast) enough so that a human ear only hears the average. Otherwise it sounds horrible.

The technical term for this is PWM...

Now you mention it, I have heard of that term, most likely on the same video where I saw what the SID waveform looked like.

Is it impossible on the AY, or was it just a lot easier on the SID so that all the Commie/Amiga programmers used it all the time?

On sound chips like the General Instrument AY-3-8910, AY-3-8912, AY-3-8913, the Yamaha YM2149F and the Texas Instruments SN76489, (a slightly crude form of) PWM can be achieved by altering the volume from minimum to maximum and vice versa. That’s similar to how samples are played back on these sound generator chips. It is rather processor intensive though.

[media]https://youtu.be/4xVfuuSty48[/media]
( link )

Mark

[RR7]

With regards to the music, the copy of Bubble bobble i had was on the taito coin-op hits compilation, it had the loader with the countdown timer on the loading screen, and on my grey +2 in 128K mode there was no music. Which i found disappointing.

however, i discovered if i held the button down to start the game, then the music played, but it resulted in graphics corruption on the bubbles.

I don't know if such a thing can be tested under emulation.

EDIT: confirmed! here is how to re-produce, at least in Spectaculator

load the game from the coin-op hits version (tape 2 side A)

if you select controls, such as pressing 2 for sinclair then 3 for kempston, the press 1 or 2 page appears with no music, the game plays normal.
if you select controls, by pressing 2 but then keep pressing 3 repeatedly, the music starts and then the 1 or 2 page appears, but the bubble graphic appear corrupted.

it doesn't seem to care if i choose +2 or +2A

EDIT2: the Original version with the firebird bleed loader thing works properly, as I'm sure everyone already knows.

EDIT:3 the hit squad one is the same as the compilation one, (typed before I read Stu's reply!)

[RR7]

maybe someone can have a look and see if it was done on purpose due to the change in loading scheme, or if it's just a bad side effect no-one noticed or cared about.

EDIT: sorry just re-read and understood this:

There's a noticeable delay after you select the player controls, and what seems to be happening there (for those interested - starting at 0xfac3) is a timing check - the end result being if you've loaded the tape on a +3 then the music is explicitly switched off. Odd that there's no problem with the disk version though

if you load it in pentagon mode, the music is enabled again (different timing?), with the graphics corruption, so i guess they discovered this error when they changed the loading scheme and hacked in the disabling of the music if it detects that specific timing. i don't see any re-release or compilation disk version.

perhaps mashing a button on that screen disturbs the timing measurement somehow and fools the code into not disabling the music, revealing the corrupted graphics.