I'm tired with showing RMTs. I DID show what it is about, released rmt files... did teaching videos.
But it isn't the solution to "have a look how it is done". Just using the setting doesn't help for a new tune. The musician knows it better, if you understand. I offered cooperation a thousand times, but no reaction. I also found SOME tunes that used my setting , but there is no mentioning about me. As long as people don't understand the physical part of music creation, everything is wasted time.
If at least a tracker existed that allowed to set those instruments "physically" correct, I'd do a library for sounds. Releasing RMTs of what I'm doing is actually 180 degrees wrong to the direction of proper POKEY development.
Things are really weird , when it comes to POKEY music development.
Particular two of my edit, really should have made people nosy.
This one , sack/c0s did in a complete version for SID, then he did a small loop in RMT. I grabbed it and this is the result:
Also the next one. It's a very dark theme. The variantions in sound should have woke up people's interest. It sounds a bit like 15kHz, but it's clearly more in tune.
To make it clear once again:
It is real POKEY sound. The result is only limited by RMT and the emulations. CPU usage is very low. On other platforms, for sure, people would ask for coopoeration, as the music taste is different, it could be a result of "fitting taste plus fitting POKEY sounds in one" . But it doesn't happen.
I was surprised myself, that people liked this one that much:
It's a "simulation" of a 2 channel beeper engine. As it might be clear, it's also just POKEY programming in RMT. Very low CPU needed though. It's also originally a 4 channel Mod, and I put the channels together to play it at 2 channels.
I know that chip-tunes often have a little problem with proper tuning. Word "de-tuned" in this situation means that many notes don't match proper frequencies to sound nice. There is a frequency margin for every note, but in this case this margin is too big for my ears. Don't take it personally, as I really know POKEY limitations in this matter.
Alex The margin of the notes in my demonstration is by far better than average POKEY tunes, just by the fact that the note dependency is really the modulo of frequencies, if you understand. It's no note dodging. So I wonder about the basis of your statement. SID allows to have "modulo" sounds using different channels and waveforms, but there is a stressed shift in the basses. So , if you hear more than 1 SID tune in a row, you'd need some ear correcting contrast to not break the ears. From that old chips, I only would name it music, when it comes to the NES Chip. The 5 channels do their single part.
Well, there exists a pokey version. Using standard instrumentations.
While it sounds a little peaky, the overall tune is ok. But the melodical "lead" has real problems. Some notes sound like broken stairs of an old ruin, which destroys the melodical flow.
You are right. And it is build on the notation of the video by sack-c0s . I'm using the notes and pattern with my own instruments. And it is by far not finished, as some parts of the tune still need to be added.
But what is the issue?
I'm using "impossible" sound on POKEY to resemble a lot of PWM sounds on a REAL machine with just minimal CPU usage. The tune is definitely there. Also, the bass, as you hear it in the tune is "impossible" already. Most relevant part: RMT is in no way actively supporting that.
Imagine the PWM active programmable including ONE software synth channel...
BTW: The loop at the end is intentional to approve the stability of the used sounds. They sound the same, every cycle. It's only my tunes that show those "fully controlled" sounds.
Can we assume several aspects with that latest tune?
-POKEY can sound better than every book has written in . -A real POKEY Tracker could go much more ahead for even better musically qualities. Have a look at the CPU indicator. -Such a tune can run on a minimalistic player. also -With some cooperation, we could have such a tune 30 years ago. -This means that possibly the quality of POKEY music was today at a much higher level.
btw.
Someone interested to use the tune in a prod. Let's say for the ABBUC contest or else?
I cannot promise that the tune sounds 100% the same on the real thing. But it sounds in all available emulations correct. Actually , it might sound even better on the real Atari, because the Altirra "non linear mixing" is enabled, which is a little over.
Actually, I'm not good in doing 3 channel compositions for sound chips. This time I'm using "real" SID notations. Doing the "fine tune" to bring the sounds to the needed note , using POKEY tricks , is real fun then.
If someone could provide the real notation of Hippel's "Buzzer" tune. I'm pretty sure, it will also sound cool after adding the correct POKEY sounds :)
Btw: Here is a minor update. Some more fluent changing in the tune, and a cancelling point is removed.
New decoder from Algorithm: Update to the low cpu usage audio decoder for the C64. Some major quality and accuracy improvements in the high frequency components. * Under 600 bytes a second on average - Check * Under 32 raster lines total usage per frame - Check * No massive wavetables as additional data - infact there is none - Check What you are hearing here is 70 seconds of non sequenced audio packed into 40k Raster bars show that there is all that cpu time free (and no, no inbetween interleaving tricks Utilises new playback technique for the low frequency audio in Sid channel 3 with the high frequency components in sid channels 1 and 2.
@Emkey I was reluctant to answer here but I'm listening to this tune cover done by You again and I have one word... Wrong!
JCH author of original "Chordian" on C64 ( and one of the best musicians on C64 ever) is a friend of my group Bonzai. I'm going to ask him what his thoughts are about this. But man... This is sounds so Wrong!
If you think this is not possible with POKEY, you are wrong ;)
Not to say that this "Decoder" does nothing but playing the waveforms of SID instead from the wave itself. The real clever stuff is to calculate the available digitized source and to exchange them with the available waves that were played by the soundchip.
POKEY has 4 channels, so if someone was taking care of some similar player for the Atari...
Just because people claim that the tune only sounds correct on a 6581 , while others claim it sounds only correct on the 8580. So, be kind and safe further words on it.
@emkay first, I must say that your tune conversion above is really schmick, I love it!
As for the Algorithm's trickery - this is something I am thinking for a long time now. Unfortunately my early attempts are just awful due to 2 factors: 1. no sinus waves that are basis for FFT derived magic used by Algorithm 2. very low frequency resolution of Pokey. even switching to 16-bit does not work very well for higher freq tones, possibly with some 64Khz pokey magic it would be better.
I have a whole new idea of using machine learning and stuff to get some similar sounds, but I am too weak in this department yet.
the real problem with those algorithms is that PWM still isn't fully discovered. Actually it is almost undiscovered. See my tune, remember that RMT isn't even supporting what you hear. It's all trickery POKEY has the 1.79MHz. POKEY has 15kHz clocking Both can run at the same time, using programmed PWM .
Sine wave isn't really needed. As you hear in the SID. there are real gaps in the frequency range, also there is no high tone available. Using the 15kHz PWM results in sawtooth when the sound substract. And it turns to triangle, when the sounds add together. The high notes get clean and useful sawtooth backward and forward were possible. PWM is not only a combination of 256x256 different pitches. It is also a combination of 256x256 different offsets in the wave creation. The calculation of this behavior is calculatable, just by the fact that POKEY plays the next wave after the current wave is completed. That "played wave" is the time difference (the offset) between the main voice and the "filter" voice.
Koledzy niech mi będzie wolno podlinkować coś co może być w temacie tego wątku. na kommodorowskim CSDB leci konkurs na muzykę bez użycia filtrów. wiem że filtry często są wyciągane w dyskusji POKEY vs SID więc tu można posłuchać co da się bez nich na SIDzie zrobić.
wymagany niestety emulator bo nie ma jeszcze na YT link jest do CSDB a tam trzeba sobie pobrać pliczki najlepiej .prg
The modified version is far away from the first approach, unfortunately in a negative way. Wrong is at least the added part on the third voice. It actually doubles the lead, without any significant harmonic relations. That results with a sort of "unison mush". To be perectly honest - first release seems to be better, but... oh well. It is always great to hear POKEY generating interesting sound, different from all those (well known and just boring) "beeps" and "plums" only.
It's about the music theory (not preferences only). In fact, I am a musician (even if not so active today, but still I am) and (for example) I used to love to hear, when bassist in my group just knows where to double notes of my part/riff, and where not to do so. Nevertheless, I think that you brought something new, fresh and very interesting. Keep up the good work.
