Hardsynth

[emkay]

interesting theory you have mentioned...pokey is unable to play correct "in tune" notes... that would stress the reason why pokey was used for sound fx in atari arcades....

 

2nd theory....every atari tracker does have this problem... CMC, MPT, TMC...would be nice to hear from Jaskier as coder of TMC what he thinks about this problem and how he might solved it

 

same goes to Miker, Xray and other Atari musicians what they think about the OOT-problem (Out Of Tune... ) and how they try to make it right with their trackers...

 

this is interesting... some clever guys have to think about that...

929162[/snapback]

 

Portamento helps.... vibrato too .... arpeggio helps either...

To have "in tune" notes, filter & 16 Bit are available, but we have no tracker that handles it correct.

 

 

On the other hand...

Technically, no one likes music because it is 100% in tune. Remember singers using vibratos with their voice....

Same with "flanger" which is mostly "off" from the correct pitch.

The essential is the correct note at the correct time. And, the "step width" between the semitones have to be correct.

Especially when creating "wide-sounds" (flächensounds), that uses more than one frequency, the real pitch has to be corrected "by ear".

When you buy a synthesizer Keyboard, alls sounds are tuned, before you even think about playing a melody.

 

So the musican can do "workarounds" to have an "off"-note sounding nice. And that is what is done with all 4/8 voice tunes played with the (double)-pokey.

 

The biggest problem, POKEY has, is a missing high tone cutting external circuitry. So you have to play high notes at a lower volume than deeper notes for correcting this.

 

Everything else is the same, as with all other electronic sound generators.

Even "real" instruments have this "problem":

A Trumpet plays only music in a mid-high range, while a Trombone can play deeper notes.... both use almost the same waveform.

Sample player are using multiple instruments, to have several octaves stable sounding.

 

The question is, particular with the "trumpet" sound I used in the last tunes, no one really seems to realize. It has the same "colour" over 3 octaves, which is outstanding. It's done by the filter offset, that adds a real noise to the higher notes, without having false harmonics.

 

Well... I wonder if someone really is recognizing the difference between the "standard" Filter usage and that "trumpet" sound...

[Tezz]

I'm not sure if you got it already, but the filtered sounds I used, are using a higher frequency range already

No, I've not got that

It's done by the filter offset, that adds a real noise to the higher notes, without having false harmonics.

Do we sacrifice a channel by joining with these filter settings? My main lead instruments use filtering so Ch1+3 joined, so if the bass sounds also require 2 channels joined 2+4 for filter settings, I have no more channels for percussion unless the song uses an "in-between bass"

 

Sample player are using multiple instruments, to have several octaves stable sounding

Yes, with sample playback you can get away with around an octave of real sounding range with a single instrument key sampled. Most pro's use multisampling mapped across the keyboard so that C then E then G are separate to give the full range becuase there is a lot more to the instruments characteristics than simply it's pitch played higher and lower. I've experimented with multiple instruments in RMT to possibly solve the OOT problem but this doesn't help because the steps are not fine tuning

 

Technically, no one likes music because it is 100% in tune. Remember singers using vibratos with their voice....

Same with "flanger" which is mostly "off" from the correct pitch.

The essential is the correct note at the correct time. And, the "step width" between the semitones have to be correct.

Yes, this is true but I think we are using the term "off key" loosly here, the scale or octave of notes must be the correct pitch or else it will sound like someone playing an out of tune piano.

 

Emkay>To have "in tune" notes, filter & 16 Bit are available, but we have no tracker that handles it correct

Yes, part of our difficulty is that the trackers and emulation are not perfect yet.

 

Heaven/TQA some clever guys have to think about that...

Yes, I think there are coders out there on the A8 scene capable of solving this.

[emkay]

Do we sacrifice a channel by joining with these filter settings? My main lead instruments use filtering so Ch1+3 joined, so if the bass sounds also require 2 channels joined 2+4 for filter settings, I have no more channels for percussion unless the song uses an "in-between bass"

 

 

You may check my latest project.... It's a long lasting tune, so be patient.

 

Well... I like the original much... (you know "gooseneck" )

 

To have this "gooseneck", I had to put nearly every instrument a pitch higher.

(This IS a bug in RMT that the 15kHz frequencies are slightly wrong, as the 1.79MHz is wrong)

So the "Main" part (refrain) is very very good under the current circumstances.

There is again one "weird" instrument which I tried to handle with 1.79MHz. So those parts are really "a bit weird".

Some parts are using 2 filter voices, but you don't miss the bass.... guess why!

Some parts are full 4 voice...

Edited September 11, 2005 by emkay

[emkay]

Someone said, it "multi" instruments cannot enhance pokey's music abilites...

 

not really finished, but a demonstration.

 

 

 

 

updated file

 

 

... another timing issue

Edited September 11, 2005 by emkay

[Tezz]

Interesting Emkay. can you post the RMT too.

[emkay]

Interesting Emkay. can you post the RMT too.

930085[/snapback]

 

Ofcourse

 

 

And the "finished" Silkworm title...

 

 

 

 

I would recommend recordings, to compare the original hardware ...

Edited September 12, 2005 by emkay

[analmux]

Interesting Emkay. can you post the RMT too.

930085[/snapback]

 

Ofcourse

 

 

And the "finished" Silkworm title...

 

 

 

 

I would recommend recordings, to compare the original hardware ...

930098[/snapback]

 

emkay, I had a look inside your turl1.rmt file. In instruments 7,8,9 e.a. you chose a table of freqs of F9. Why? This results in notes that are slightly out of pitch (i.e. it's not a clear tonal scale anymore)

[analmux]

Here is a suggestion for the turl1 tune.

 

I've changed instrument 09 and 0A a bit, and tried to get the music in pitch.

[emkay]

emkay, I had a look inside your turl1.rmt file. In instruments 7,8,9 e.a. you chose a table of freqs of F9. Why? This results in notes that are slightly out of pitch (i.e. it's not a clear tonal scale anymore)

931153[/snapback]

 

 

 

Note... a bit down... (as discussed 1000 times before)

 

Setting note a pith higher is too high

Starting with filter "down" sets the pitch a half tone down..

So... the tone itself is closer than the original pitch.

[emkay]

Here a version of the "Last Samurai" tune.

This time full 15kHz filter & some "fx" on the sounds...

The quality of the tune depends on the "FX" which can be "enhanced".

 

The Result could be mulitiple times better, on the real thing with cycle exact filter handling.... which easily can be done in VBI.

 

I did a "per pattern filter-finetune" which is fair enough.... for demonstration purpose.

[emkay]

It fits somehow

[emkay]

One question:

 

If one, who plays the notation correctly, plays a melody on a piano, and the song is played wrong...

Is it a mistake of the player or a malfunction of the piano?

 

 

I tried to create a "multi sample" tune and had the "Mad Tracker" to adjust every instrument. Having a simple sound fx gives an allmost "right" note.

Switching to 15kHz it's simply "off".

 

So while "Madtracker" gives a sound at "A2" , RMT plays it at "A3". Setting POKEY to 15kHz, the sound isn't anymore "A2". You have to play "H2" to have "A2".

It seems, the whole notation even on simple generator A is not only "off", the "steps" between the notes aren't correct also.

 

Any chance to have the notation corrected in RMT? This false notation setting explains why my experiments at 15 khz Filter aren't giving the "logical" results.

[emkay]

Well... having the whole tune "tuned" by the ears, almost every note is now "off" from RMT's Table but the melodies are better fitting to the original tune... exept the part with the portamentos.

This time, the 16 bit generator is used in some parts of the tune.

Particular the "guitar" sound shows what would have to be done:

Building a corrected notation table with the programmed sounds.

 

It seems, only the standard 64kHz notation table is ok... so it seems to be over the top to want corrected tables for special sounds, because we even don't have correct notation tables for the "simple" pokey sounds yet.

[Tezz]

Any chance to have the notation corrected in RMT?

Yes indeed. I think that RMT could enforce this although perhaps Raster will suggest that RMT is an "open" development tool which simply allows users to access the Pokey registers. The octave ranges can differ so an A2 note needs to be input as A3 in the RMT track which isn't really a problem but what is a prolem is when note A is not actually playing in the key of A at all.

[emkay]

Found an interesting site, belonging to the correct notation frequencies.

 

http://www.physik.uni-muenchen.de/leifiphy...e/tonleiter.htm

 

It's in german, but the tables speak for themself.

[emkay]

Found an interesting site, belonging to the correct notation frequencies.

 

http://www.physik.uni-muenchen.de/leifiphy...e/tonleiter.htm

 

It's in german, but the tables speak for themself.

934264[/snapback]

 

 

As the site says:

 

You don't need to have the exact 440Hz, you have to have the correct relation between the notes:

 

 

 

f0 = 1 · f0 Prim

f1 = 16/15 · f0

f2 = 9/8 · f0 Grosse Sekund

f3 = 6/5 · f0 Kleine Terz

f4 = 5/4 · f0 Grosse Terz

f5 = 4/3 · f0 Quarte

f6 = 45/32 · f0

f7 = 3/2 · f0 Quinte

f8 = 8/5 · f0

f9 = 5/3 · f0 Sext

f10 = 9/5 · f0

f11 = 15/8 · f0 Sept

f12 = 2 · f0 Oktave

 

 

With the pokey we have the 16 bit to controll the relation between the notes (as we can have almost 440Hz (with 65536 frequency steps for 7 octaves))

Or, we can use 15kHz Filter for exact programming of the relations between the notes, even if "A" sounds like 448Hz...(with 65536 filter combinations at 50Hz for adjusting the main frequency)

 

To say it in one sentence: RMT's weak spot is in the incorrect halftone relation.

 

 

And, I have to admit that I was wrong. I thought, POKEY usage was at 50% with the available trackers.

Today I assume about 10% of the given possibilites.

 

So, until POKEY usage is given, here the real last test tune (showing some filter possibilites/someone seems to play with the notes).

Edited September 19, 2005 by emkay

[emkay]

No Test, No Demo...

Just the question, what sound you prefer more?

It's one short tune with different filter settings at the first instrument, which means: 5 settings = 5 files.

[analmux]

I don't know, they all sound great

 

But if you really want to know: Maybe the 2nd one: blocked banjo.

[emkay]

I don't know, they all sound great

 

But if you really want to know: Maybe the 2nd one: blocked banjo.

936324[/snapback]

 

 

Hm... I'm tending between B and C .

 

Grenslev sounds a bit more spacy, but imho also a bit more sick.

[emkay]

Well... sometimes there is an idea....

 

It's all with 15kHz.

 

Please have a listen particular to the filter voice.

 

100Hz making it possible to play short filter notes and to set the pitch higher.

 

From the point of a chiptune view it is outstanding. Every note is played as it have to be and the "between sounding" is correct.

[Cybernoid]

Well... sometimes there is an idea....

 

It's all with 15kHz.

 

Please have a listen particular to the filter voice.

 

100Hz making it possible to play short filter notes and to set the pitch higher. 

 

From the point of a chiptune view it is outstanding. Every note is played as it have to be and the "between sounding" is correct.

998206[/snapback]

 

 

Awesome sounds! Is the out-of-tune sound a result of lack of RMT abilities? It sounds better on the real hardware versus the emulator.

[emkay]

Awesome sounds!  Is the out-of-tune sound a result of lack of RMT abilities?  It sounds better on the real hardware versus the emulator.

998313[/snapback]

 

The sounds are awesome, and the "attendance" isn't correct.

But, the main filter-voice cannot be out of tune. Perhaps it doesn't sound as you await pokey has to sound?

The real pokey has to sound better, because it produces less noise compared to the emulation.

[Cybernoid]

The sounds are awesome, and the "attendance" isn't correct.

 

Hmm, what do you mean by "attendance"? I think something is getting lost across the German/English boundary.

 

 

But, the main filter-voice cannot be out of tune. Perhaps it doesn't sound as you await pokey has to sound?

 

I think that the main voice is in tune with respect to all of its notes, but it may be out of tune with the other voices... I think... not sure. It sounded like the arpeggio notes were off.

 

The real pokey has to sound better, because it produces less noise compared to the emulation.

998322[/snapback]

 

Yes, this makes sense. Great tune, BTW.

[emkay]

 

I think that the main voice is in tune with respect to all of its notes, but it may be out of tune with the other voices... I think... not sure.  It sounded like the arpeggio notes were off.

 

 

998335[/snapback]

 

 

That's what I mean with "attendance" the sounds beside the main voice (sometimes named accord) .

I really would like to have a tracker with 2 of those synth-voices plus a digi-channel.

[emkay]

Great tune, BTW.

998335[/snapback]

 

 

It's a part of the red sector theme (as the title shows)

 

 

Here the rmt.... if one has a good idea for the arpeggios.

 

Here the link to the original mod:

 

 

http://www.modarchive.com/cgi-bin/download.../R/redsectr.mod