| |
Conrad
Registered: Nov 2006
Posts: 839 |
Playing digi samples with a tune running at 12x or over.....?
Hi.
Is is possible to play samples using a tune playing at say 12x speed or maybe over.... rather than using a standard NMI interrupt routine?
After all, samples are played by basically altering $d418's low nybble (volume) ... just curious if a music player could achieve this by playing at 12x or such.
At the same time also, you could have cool ringtone effects playing in 1 to 3 voices AND play digis at the same time (and also have clear sharp filters possibly). |
|
| |
Stryyker
Registered: Dec 2001
Posts: 465 |
Samples are typically played anything from every third to every tenth rather line. Why not do the experiment? You can play samples using IRQ too or not using any interrupts. 12x would be less than half of the typical sample play rate. |
| |
ChristopherJam
Registered: Aug 2004
Posts: 1391 |
12x would still only give you a 50*12=600Hz sample rate - any frequency components over 300Hz (around middle C) would be completely lost.
As a rough guide, you need at least a 2000Hz sample rate for barely intelligible speech, probably more like four or five thousand for any kind of worthwhile instrument.
I think you'd need more like a 100x player to do digis this way! |
| |
Stryyker
Registered: Dec 2001
Posts: 465 |
And the player routine would probably use too many cycles if it does things other than just play the sample. |
| |
Dane
Account closed
Registered: May 2002
Posts: 421 |
The digis in Digital Magic are replayed every second line and do not use NMI interrupt. They still sound like crap though. :/ |
| |
algorithm
Registered: May 2002
Posts: 702 |
Then there is the 'hammerfist' part in Wonderland 6 which removes all borders, topbottomleftright full screen rotating (sort of) sprites and digi. good for its time |
| |
Style
Registered: Jun 2004
Posts: 498 |
Quote: The digis in Digital Magic are replayed every second line and do not use NMI interrupt. They still sound like crap though. :/
Thats the magic :) |
| |
Dane
Account closed
Registered: May 2002
Posts: 421 |
Nah, the magic part was more about sideborder opening and 8 sprites on FLI while playing those (crap) digis.
I really should do a sequel some day. |
| |
Style
Registered: Jun 2004
Posts: 498 |
I started to do a demo part with open sideborders, rasters (every 2 lines), digi track and a fullscreen sideborder scroller......
I really should get back to it
|
| |
Mace
Registered: May 2002
Posts: 1799 |
There's already a demo doing full-screen rasters with scroller in all borders, playing digi music.
But damn me if I can think of its name... |
| |
SIDWAVE
Account closed
Registered: Apr 2002
Posts: 2238 |
New Limits |
| |
chatGPZ
Registered: Dec 2001
Posts: 11165 |
XAKK did it aswell in "microsleep"
<3 |
| |
Conrad
Registered: Nov 2006
Posts: 839 |
@ChristopherJam
Yea, I just realised about the minimum Hz for a clear digi sample, which would be at least 2000Hz.
This would be:
2000Hz / 50Hz = 40x player.... OMG!
Hmmmmm... coming to think of it, a player could be made which would do a different job in each call (i.e. one call would work on just voice 1, then the next call would work on voice 2, them voice 3, then filter, blocks, etc) Using optimisation, each call could use about say 4 rasterlines, where in each call you could alter $d418 at the same time using, sample data. Then taking 40x as 28 in hex, that would be:
4 * 28 = $a0 rasterlines, which would fit nicely within max number of rastertime on a c64...except u cant have any cool visual effects in the background i.e. sideborder, or even rasterbars :(
But anyway... as it seems this would be too complicated (yet phenominal) to do, so i would think sticking to NMI interrupts would be better. Besides, with NMI you can also alter the frequency of the sample, allowing pitch shifting etc, where doing that in an IRQ would be IMPOSSIBLE! |
| |
Mace
Registered: May 2002
Posts: 1799 |
That's the one, Rambones.
Thanks ;) |
| |
chatGPZ
Registered: Dec 2001
Posts: 11165 |
Quote:Besides, with NMI you can also alter the frequency of the sample, allowing pitch shifting etc, where doing that in an IRQ would be IMPOSSIBLE!
hu? explain |
| |
Dane
Account closed
Registered: May 2002
Posts: 421 |
Depends on how you do the IRQ of course. |
| |
JackAsser
Registered: Jun 2002
Posts: 1997 |
Also for the record:
CIA1 timers produces IRQ
CIA2 timers produces NMI
VIC produce IRQ.
RESTORE produce NMI.
BRK produce IRQ.
etc..
So, when you speak of NMI I guess you mean "using timer interrupts (CIA2)"... and when you speak of IRQ I guess you mean "using VIC raster interrupts".
|
| |
Cruzer
Registered: Dec 2001
Posts: 1048 |
Sample/raster stuff in this classic as well... Mixer
|
| |
algorithm
Registered: May 2002
Posts: 702 |
And another demo in the wonderland series (Displayed a FLI pic of a dragon) with VSP and digi's |
| |
trident
Registered: May 2002
Posts: 85 |
Allright if we should boast about our digi + border effect code, I'll do that too :). Hellfork, all-border playable Wolfenstein clone with sampled sound effects over SID music. |
| |
Oswald
Registered: Apr 2002
Posts: 5034 |
*yawn* hmm anyone around that can do something more complex than digis together with rasters ? how about some hot filledvector talk, who can come up with the fastest 3d rot, linedraw, and fill routine ? |
| |
trident
Registered: May 2002
Posts: 85 |
Quote: *yawn* hmm anyone around that can do something more complex than digis together with rasters ? how about some hot filledvector talk, who can come up with the fastest 3d rot, linedraw, and fill routine ?
Hmmm, didn't someone say something about digis with all-border filled 3d? ;) |
| |
Conrad
Registered: Nov 2006
Posts: 839 |
Quote: Also for the record:
CIA1 timers produces IRQ
CIA2 timers produces NMI
VIC produce IRQ.
RESTORE produce NMI.
BRK produce IRQ.
etc..
So, when you speak of NMI I guess you mean "using timer interrupts (CIA2)"... and when you speak of IRQ I guess you mean "using VIC raster interrupts".
yes. Sorry about that, must be the way I said it :p |
| |
Oswald
Registered: Apr 2002
Posts: 5034 |
trident, well anything is possible, how about a 50 fps phong torus in 64x64 pixels on the sideborder with digis ? :) |
| |
Style
Registered: Jun 2004
Posts: 498 |
Oswald has done it all :)
|
| |
Stryyker
Registered: Dec 2001
Posts: 465 |
IRQ via CIA 1 works fine for samples with some care. I remember experimenting with using two timers to mix two samples in real time. Wasn't good design but worked :) |
| |
Dane
Account closed
Registered: May 2002
Posts: 421 |
Quote: trident, well anything is possible, how about a 50 fps phong torus in 64x64 pixels on the sideborder with digis ? :)
Great idea Oswald. I'll write a sideborder tutorial someday. |
