World: r3wp

yes there's a loop

I'm filling a buffer of 256 bytes so I can change parametres in realtime 
(buffer is filled, GUI works again, I can fiddle with knobs) and 
then I add buffer's content to some sound [block!]

buffer: context [
	length: 256
	data: make block! length
	init: does [
		data: make block! length
	]
	fill: func [fnc][
		init
		loop length [
			insert tail data do fnc
			data
		]
	]
]

then just >>insert tail sound buffer/data

but because there's no difference in speed of integer and FP math 
in REBOL, everything is done in FP sou it's needed to convert FP 
to 16-bit integer

sou=so

math with fpu should be comparable in speed today. i try it.

I'm using this code (and I really hope there's some better solution): 
>> insert tail sample do rejoin ["#{" skip to-hex to integer! (sample-value 
+ 1) * 32767.5 4 "}"]

If floating point is fast enough today, it's better to use FP because 
of resolution

I know Rebol quite well but I don't know music. Can anybody point 
me to a white paper/tutorial/newbie docs/samples that explains how 
to valorize sound/data to play some notes? Thank you.

luca: take a look at http://www.rebol.com/docs/sound.html, it's 
googd start. there are examples how to synthetize simple sine wavew 
and how to apply some envelope on it.

Thank you, it is exactly what I was looking for.

Bob Moog, inventor of famous synthesizers, died at age 71. Sad day 
for electronic music. http://www.moogmusic.com

Moog producs: http://www.moogmusic.com/detail.php
Amazing!

Volker so you're interested what's possible with better sound? Me 
too. But I can tell you, what is possible with today's sound. This 
- http://krutek.info/pix/pdsynth.png;)))

That is soo cool! :-)

Does the sound sound "green" like this? :)

two or three things still missing: WAV saver, Vibrato and probably 
something else. At least WAV saver must be finished before release.

Sounds very green ;))

But it's slow. rebcode! will be great for this

if you have something to translate to c, tell me.

Volker lot of things, really :) I'd like to finish features mentioned 
above and then translate all computations to C. I'll let you now, 
thank you.

Thanks :) maybe if you make it ready for c, i could do the rest. 
means make it use structures, not relying on contexts etc. for the 
time-critical things.

Mmm.. nice picture...

kru, would you mind if we use this as a VID showcase in the Wikibook?

Henrik why not. But the GUI is not 100% finished so I hope it can 
be later replaced with better picture.

will see if I can upload it in a few mins.

uploaded. if you get a new version with a new screenshot, you can 
upload it yourself, but you need an account to upload images... otherwise 
throw the link in the wikibook group

When I create a sound using Rebol, with 2 channels, is it possible 
to access the to channels independently? So as to create a stereo 
effect?

You must create sound with stereo effect. Once you insert something 
to sound:// you lost control and cannot change even the volume.

OK, is there an example anywhere that shows how to create stereo 
sound? So far I haven't found one.

http://www.rebol.com/docs/sound.html#section-7

For sound samples with two channels both channels are interleaved, 
i.e. samples for each channel alternate within the binary.

Example based on example in Carl's sound guide:

print "Generating sound..."
    bounceL: #{}
	bounceR: #{}
    repeat modulate 20 [
        delta: negate log-2 modulate
        for amplitude 400 1 delta [
            for phase 1 360 16 [
                val: 128 + to-integer 127 * sine phase
                val: amplitude * val / 400
                append bounceL to-char val
				append bounceL to-char 400 - val
            ]
        ]
    ]

	bounce: #{}
	repeat i 400 [
		insert tail bounce bounceL/:i
		insert tail bounce bounceR/:i
	]

    ; Set up the sound sample parameters:
    sample: make sound [
        rate: 44100 / 2
        channels: 2
        bits: 8
        volume: 0.5
        data: bounce
    ]

    ; Play the sound now:
    sound-port: open sound://
    insert sound-port sample
    wait sound-port
    close sound-port

The problem is that it does not work (at least here).

And here's why:

>> sample/channels
== 1
>> sample/channels: 2
== 2
>> insert sound-port sample
== 199
>> sample/channels
== 1

There's some strange conversion always when you play a sound.

Sound is buggy as something really buggy.

RAMBO it!

Ah there's bug in example, it should be "128 - val" instead of "400 
- val" but that's not the problem

Thanks Kru, I reread that part, and actually found Carl's explanation.

Now that's interesting ... 

if I paste your example into a console, I get a very short high beep

if I paste it into a script and do the script, I get a math error! 
(Math or number overflow)

Ingo: yes there are some bugs ;)

I didn't test it before pasting it here

Wait a minute, I'll post fixed version

OK

print "Play a ping sound..." ; (init event system too)

    ; Generate a ping waveform as a binary value:
    ping: #{}
    for amplitude 500 1 -1 [
        for phase 1 360 16 [
            val: 128 + to-integer 127 * sine phase
            val: amplitude * val / 500
            append ping to-char val
            ;append ping to-char 256 - val
        ]
    ]

    ; Set up the sound sample parameters:
    sample: make sound [
        rate: 44100 / 2
        channels: 1
        bits: 8
        volume: 0.5
        data: #{}
    ]

    ; Make the ping sound 10 times:
    loop 10 [append sample/data ping]

    ; Play the sound now:
    sound-port: open sound://
    insert sound-port sample
    wait sound-port
    close sound-port

this is mono example - plays ping sound