Noise Approximation

[thorfdbg]

 

Some changes on the giana tune. Remember that some changes take 3 VBI cycles due to the given abilities of RMT.

Some sounds "open" some sound "close" and variations.... and still experimental stuff.

 

 

 

Sounds almost unlike "Giana Sisters", and I really had to use some phantasy to identify what this was supposed to be. Plus, it is - as often - completely out of tune. "My ears bleed" to use your words. Sorry, but if that is "drastically different sound", I'd rather say that I prefer the well-tempered scale....

 

Sorry, but that's not even a close catch. It's just awful... but thanks for trying.

 

Actually, this may be used as a good example of my point: Maybe Pokey can do some effects, but if you try, you'll find that these effects are not suitable to play music. If you just stick to the standard square wave it can generate, it may be ok if the notes are not too high.

[emkay]

If I wanted to have it in tune , I'd check for a remix at RKO. If you put the "better" sounding of all created sounds aside, as SID is simply better in playing deep notes, you realize the "out of tune" on the SID-Synth also. Just the 8580 plays it better...

 

Could be even more interesting to have a similar sound playing routine as with "Fanta in Space" , as it seems C64/SID is also not able to play "fitting" synth sweeps to the sampled part. POKEY offers much more variations there...

 

Could also be interesting to have a modulation feature, using two channels for digis. Or to have GTIA playing a digi channel and POKEY uses all 4 channels for waveshaping....

 

Could also be interesting to see what happens if the audio channels of POKEY and GTIA had been timed together, using additive or substractive waves, manipulating the analog circuitry....

 

a.s.o.

 

Actually , there is much more to explore than already has been explored.

 

Perhaps only 99% of what A SID could do in a stock C64 has been explored, not 100% .... but music on the Atari is at 30%-50% of what's possible, if you compare the "eventuallities"

Edited August 13, 2015 by emkay

[emkay]

Back to "Noise Approximation"

 

When you try to find a "new wave combination". It's still the problem that the sounds will jump over the 8 bit resolution. So you have to take care of creating "frequencies" inbetween the 8 bit resolution.

As RMT isn't supporting this feature natively, you have to find a solution of waveform-manipulations with the least hearable sidenoise.

Also, if you have a listen, the "2 operator" sound also produce lower frequency based sounds as one might think.

To make it more obvious, the 1.79MHz isn't used, the "main voice" is played on channel 2 and 4.

Also, this time gen. "E" used for a more stable bass sound...

 

 

[emkay]

More Noise approximation, as when editing the basses got distorted... the result sounds interresting somehow So much more noise is to expect

 

The cause for this demonstration is the high pitch sound (lead). From a calculation point, POKEY gets more unsharp in the correct pitch , the higher the sounds were. The " 2 Operator modulation" helps to find the right pitch. Additionally, you get interresting waveshapes and "waveshaping" is still allowed for one instrument...

 

To mention: still no 1.79MHz used.

 

[emkay]

Seems , more cannot be reached with RMT there...

 

 

At least some nice sounds in there

[analmux]

Indeed the '2-operator modulations' are interesting extra possibilities, and nearly no one uses this technique. The sawtooth wave approximation for example is a nice surprise for (very) low notes.

[Bryan]

1st: SID was designed in a straightforward way to create specific sounds. It's easy to program and predictable (well, until the filter revisions started happening).

 

Pokey was designed based on Atari's research into digital circuits which could create arcade sound effects so it contains various ways to generate abstract bit patterns. However, no one involved in its development ever knew what sounds it was really capable of because the configurations can get pretty complex and the bit patterns will create interesting harmonics if run at a high speed or combined with another channel. This has resulted in a lot of new sounds, but it's taken decades to find them and they're still rather difficult to program around (that is, you can't just write a couple lines of code that say, "give me sound #26 at middle C"). Pokey has turned out to be a great tinkerer's chip though and a good example of how a simple circuit can exceed expectations. It's far more flexible than something like an AY-3-8910.

 

If someone had developed a good low-level Pokey experimentation tool in the '80s, we'd be a lot further along than we are. I think my favorite classic example of Pokey programming is the Bill Williams stuff (Alley Cat and Necromancer). The guy was a master of sound effects.

[analmux]

... so it contains various ways to generate abstract bit patterns.

Do you mean PolyCounter bit patterns? Or what else?

 

 

... and the bit patterns will create interesting harmonics if run at a high speed or combined with another channel.

In what specific cases are you talking about harmonics?

 

 

If someone had developed a good low-level Pokey experimentation tool in the '80s, we'd be a lot further along than we are.

To my opinion RMT is a great tool to do some imports and use the song basics. I think it's the easiest way to play any .rmt files on a real A8 machine, and use a still not existing expanded tool to enable:

 

(1) AUDF synchronization RESET & OFFSET

(2) Polynomial Counter (LFSR) RESET & OFFSET

(3) 2-tone filter

(4) Add more 8-bit or 16-bit NOTE2PITCH tables as desired

[emkay]

1st: SID was designed in a straightforward way to create specific sounds. It's easy to program and predictable (well, until the filter revisions started happening).

Even the base waveforms vary, not just the filter alone...

Pokey was designed based on Atari's research into digital circuits which could create arcade sound effects so it contains various ways to generate abstract bit patterns. However, no one involved in its development ever knew what sounds it was really capable of because the configurations can get pretty complex and the bit patterns will create interesting harmonics if run at a high speed or combined with another channel. This has resulted in a lot of new sounds, but it's taken decades to find them and they're still rather difficult to program around (that is, you can't just write a couple lines of code that say, "give me sound #26 at middle C").

SID had been created to be easily understood by musicians and has the three 16 bit channels and the external filters. Thus the chip is with no doubt the better chip for music... with 3 channels... ltd.

 

 

What do we have with POKEY ?

 

POKEY has clean different noises, directly usable , always in the "8 bit" reolution and 4 channels

 

POKEY also has 2 independent 16 Bit channels, that could play several different waveforms. Depending on the lesser frequency-range, the pitches in that mode were closer to "real Music" than played by SID.

 

 

POKEY has the "ringmodulation" feature AND the "2 tone (ring-)modulation" feature, with the limits of 8 Bits.

Wich results in "endless" variations of resulting waveforms.

 

 

What does that mean?

 

A tune sometimes needs a good replay in melodic parts.

That's where 16 Bit is recommended

 

When playing a chord , 8 Bit resolution with "Brass" sounds does well.

 

When morphing sounds were good to enhance the tune... just use "modulation"...

 

and so on.

 

But, to actually do that, there should be some software available. RMT is the best tool now. But , according to the usability for "optimal" POKEY- sounds, the Tool is like a Wrecking Ball for building a house, if you're carefull enough, it works

[emkay]

Some experimental stuff from an recent game development.

Referring to soundstyles...

 

 

 

Edited October 28, 2015 by emkay