MSX1 vs C64

Página 8/11
1 | 2 | 3 | 4 | 5 | 6 | 7 | | 9 | 10 | 11

Por PingPong

Prophet (3793)

Imagen del PingPong

31-01-2018, 18:45

syn wrote:

Uuuh i said opll aka msx-music aka ym2413 aka fmpac. Your example shows opl2 thats a different beast.

And as i said three times now, it depends on situation. For example when you need real time control over the cutoff filter or want a pwm type of lead id say SID over opll. I suppose both effects could be faked on a opll using basic tracker techniques but results may vary.

people that want control use the pcm way.
and mangling SID is as mangling PSG. It require CPU effort.

Por PingPong

Prophet (3793)

Imagen del PingPong

31-01-2018, 18:48

Metalion wrote:

AFAIK, nobody mentioned the biggest technical difference between SID and PSG : the waveforms.
- PSG has only ne waveform available : square.
- SID has three : sawtooth, triangle, and rectangle pulse.
That's already there a clear advantage over the PSG.

And furthermore, from my point of view, the main advantage of the SID are the cut-off filters. That's what gives it its warmer sound, with more bass.

I must say I was disappointed by the PSG when I got my MSX in 1984. My neighbour had a C64. The graphics were better on the MSX, but there was no match for the C64 sound.

The others waveforms are the noisy and chaisaw ones, that are abused and charaterize almost the entire music on c64.
the sid sound better only with the square wave .
And effecively under the tech point of view this make sID superior to PSG. But only on PSG. OPx is far from SID or PSG.
About the warmer sound i would call it more chainsaw instead of warm.... ;-)

Por wolf_

Ambassador_ (9906)

Imagen del wolf_

31-01-2018, 20:20

PingPong wrote:

Maybe, but in the old 80 days there were no synth that used the SID, but a lot used Yamaha OPx chips.
Should mean something i think.

There are two reasons for that;

1) Many keyboards were made by Yamaha, and obviously they were making FM-chips with the same frequency as LEGO made their bricks.
2) It was a new sound, yet affordable. Keep in mind, back in those days it was all analogue synthesis. People were desperate for something new, compared than sawtooths, blocks and triangles, yet samplers (Fairlight, Synclavier) cost a fortune, and not much of an option for consumer electronics. So, yea, FM was the thing to go for if you wanted something new and affordable.

When sample RAM got cheaper and processors more powerful, the market gradually moved towards sample playback, and FM quickly ended up in the hands of fans and sound designers (however few they were), while the mainstream moved towards sampling. The Roland MT32-family (which includes the early E-keyboards) is a nice example of this. Those keyboards were an Italian invention btw (the SIEL company), not Japanese.

Por PingPong

Prophet (3793)

Imagen del PingPong

31-01-2018, 20:34

wolf_ wrote:
PingPong wrote:

Maybe, but in the old 80 days there were no synth that used the SID, but a lot used Yamaha OPx chips.
Should mean something i think.

There are two reasons for that;

1) Many keyboards were made by Yamaha, and obviously they were making FM-chips with the same frequency as LEGO made their bricks.
2) It was a new sound, yet affordable. Keep in mind, back in those days it was all analogue synthesis. People were desperate for something new, compared than sawtooths, blocks and triangles, yet samplers (Fairlight, Synclavier) cost a fortune, and not much of an option for consumer electronics. So, yea, FM was the thing to go for if you wanted something new and affordable.

When sample RAM got cheaper and processors more powerful, the market gradually moved towards sample playback, and FM quickly ended up in the hands of fans and sound designers (however few they were), while the mainstream moved towards sampling. The Roland MT32-family (which includes the early E-keyboards) is a nice example of this. Those keyboards were an Italian invention btw (the SIEL company), not Japanese.

if SID alike chips were more hi-fi yamaha could produced those chips.
About sampling way was affordable as technology improved. And of course a lot flexible than previous ones.

Por PingPong

Prophet (3793)

Imagen del PingPong

31-01-2018, 20:38

NYYRIKKI wrote:

Did you know that PSG can emulate SID as well? Hear this: https://www.youtube.com/watch?v=Zr1r6FSrsz4

this proves the fact that the SID is a more sophisticated AY. and without the overestimated low/band/high pass filter ;-)
By constrast the AY could never be able to emulate a FM OPxxxx.

Por TomH

Champion (327)

Imagen del TomH

31-01-2018, 21:34

My every-hasty reading on how the OPL works internally is that each channel is two lookups into a 10-bit 256-entry logarithmic sine table, summed and used as the index to an 8-bit 256-entry exponent table, using the formula:

out = exp(logsin(phase2 + exp(logsin(phase1) + gain1)) + gain2)

So for the instantaneous output that's four table lookups and three adds. Then presumably two adds to update the two phases. Per channel, so do that nine times.

So I think you're right that emulation would be a whole-CPU effort, and even then produce an unacceptable output. Call it 200 cycles per channel at a very rough guess (doing it all in 16-bit arithmetic rather than shifting back and forth: 12 cycles per 16-bit add, 8 for a lookup, is probably 140 just for tables and phase updates, even before worrying about register and memory juggling), and you've topped out a little below 2000Hz. For less than 1/20th of the audible range covered.

Por Grauw

Ascended (10181)

Imagen del Grauw

01-02-2018, 01:19

TomH wrote:

So for the instantaneous output that's four table lookups and three adds. Then presumably two adds to update the two phases. Per channel, so do that nine times.

So I think you're right that emulation would be a whole-CPU effort

Assuming you were replying to my “but really, the SCC is the master of all” comment…

It’s much cheaper than that… Because the modulator is multiplier-locked to the carrier the waveforms are periodic. So for each step of the envelope generator you can just generate the waveform once with the fairly trivial math you described above (thanks logarithmic space).

Similarly, you can implement a filter quite easily on the SCC; if you implement an FFT and IFFT (granted, expensive), and then just recompute the waveform while moving a mask on the real values in the frequency domain (like a 32-band equalizer with key tracking), you have your filter effects.

These FM or filter effects don’t need to update for every cycle of the waveform, but rather are driven by an envelope generator (or MIDI CC) which can run as slowly as you want it to, e.g. at 60 Hz, and will still sound fairly good.

What’s more, you can precompute these waveforms into wave tables and just scan through them during playback and get real fast performance.

Both of these ideas are on my to-do list to try sometime, it would be neat. Probably first as some PC or web tools (I already have a test somewhere for the filter), and later maybe try a realtime implementation in an MSX synthesizer.

Wavetable synthesis is ridiculously powerful, and easy to implement in hardware. All it needs really is to have the right tools to generate interesting wave tables, and on MSX typically the only tools used are “let’s draw a shape with my mouse or cursor keys and see what it sounds like”, and “let’s linearly interpolate (morph) between two waveforms”. Which is nice, but just the tip of the iceberg. I wonder if Konami’s audio engineers used some more advanced tools internally to generate wave tables for their later titles, or they just stuck to the above…

The biggest weakness of the SCC as a wavetable synth is the limited number of samples per wave (just 32) which causes stepping and the stepping introduces high frequency overtones. So it e.g. can’t really play a pure sine. But unlike imaging in a fixed-frequency DAC, the overtones are harmonically related to the pitch, so it isn’t really a bad thing. I would say that’s what the rich sound character of the SCC attributes to. If it wasn’t a bit low-fi 8-bit sounding, it wouldn’t be retro.

Por msxlover

Supporter (1)

Imagen del msxlover

01-02-2018, 10:36

This is a fantastic debate, but I'm afraid there's no contest! The MSX computers, even MSX1, are far superior to the C64. This is because their MSX BASIC language makes it easy for users to do the fundamental things, such as drawing lines, and playing simple little tunes, even in three channel polyphonic sound, but the C64 makes it so difficult to do these things that most C64 owners just gave up trying. Why should anyone have to learn Assembly Language just to draw a line across the screen? Why would they want to? For more information about this, please read the blog "Commodore 64 Crap", which started in August 2012 to put down the 30th Anniversary of that crap computer. You can read it on https://commodore64crap.wordpress.com/ , but it's best to start by clicking on the archives for August 2012, then read through it all in chronological order to appreciate it fully. The blog contains lots of mentions of MSX, Atari, Sinclair, Acorn, and other computers, explaining why they're much better than the Commodore 64.

Por ARTRAG

Enlighted (6584)

Imagen del ARTRAG

01-02-2018, 09:25

https://youtu.be/oBegD7k2wvo

Some effects in this demo can be done only on msx2+

Por AnsiStar

Expert (101)

Imagen del AnsiStar

27-07-2020, 23:28

I don`t know, if somebody said it before, but another good reason for an MSX were the manuals (Handbooks). My first Computer was an MSX I. A Sony Hit Bit. As you know there were two really good books to inform you. Very comfortable for an beginner. The C 64 had only one booklet with some peeks and pokes if my memory is correct. Another reason: there is no Penguin Adventure for the Commodore... Wink

Página 8/11
1 | 2 | 3 | 4 | 5 | 6 | 7 | | 9 | 10 | 11