RO Sound System document

You might gain some enlightenment from looking at my file at
http://squeakvm.org/cgi-bin/viewvc.cgi/squeak/trunk/platforms/RiscOS/plugins/SoundPlugin/sqRPCSound.c?view=markup
(hope that actually works and gets you to the file?). The early part is wrapping of the module calls that would likely be useful to pretty much any user (in C, anyway) but the later code is rather Squeak specific and likely to be worse that Greek to you.

I did download that (and some of the related files) but as yet I’ve not been able to make sense of it. :-/ Not a very good programmer, alas. I can see various ‘C’ proceedures, etc, but not work out how they are used and in what order. And I know zero about ‘squeak’.

What you said may help a bit, though, so I’ll have more of a puzzle.

FWIW I’me trying to cobble together something akin to the demos I did based on DigitalRenderer. I can then in the document use example chunks like “start by using this to setup the process” and give a few lines of ‘C’, then say " then call this every time whatever happens " and give another short chunk of ‘C’. All to show the ordering/use of the swis or whatever.

I’d also love to example the use of SoundFS. But that doesn’t seem to have been updated (?) and again, I have no real documentation – although it would probably be much easier to work out and use.

I’ve emailed Andrew Sellors at the address given on his webpage, but no reply as yet. I’d also email John Duffel if I knew his email address. So if either of them read this, a response would be welcome!

I’ll be busy today but will try again given what you’ve explained here about how to proceed.

Thanks

Jim

Are there any ideas for the 8bit sound system to use 16bit samples? – ie for programs like ‘Rhapsody4’ and ‘Sibelius’.

It would be possible to take a 16 bit (linear) sound sample and use a voice handler to output it as 8 bit log (which gives about 12 bit linear quality). In fact it is simply a look-up table for each sample value. Where it gets tricky is if the ‘WAV’ sound sample is at 22kHz and you want to play it back at (say) 90kHz (Pandaboard ES). The processing required might make it impossible to do it ‘on the fly’. To do it properly you don’t just repeat each 22kHz sample value four times as this would introduce audible distortion.

Where it gets tricky is if the ‘WAV’ sound sample is at 22kHz and you want to play it back at (say) 90kHz (Pandaboard ES). The processing required might make it impossible to do it ‘on the fly’. To do it properly you don’t just repeat each 22kHz sample value four times as this would introduce audible distortion.

The PandaBoard does support 88.2k as well as 96k. This may be useful because doing x4 upsampling to 88.2k from 22.05k would be much easier than upsampling it to 96k. Simple integer upsampling ratios are about the easiest to do decently.

Jim

Just to let people know I’ve released a newer and extended version of the ROSS Document at

http://jcgl.orpheusweb.co.uk/temp/ROSSDocument.pdf

It is still a ‘beta’ as I’ve not finished the editing, etc. But it covers more than before, so I’d welcome feedback on errors and things I’ve missed. I’ll then set about doing a more complete and accurate version.

Jim

I note that you ‘do not say much about the 8bit voice generators’. It seems odd not to cover them as they allow eight independent channels of sound (so you can have notes still decaying while the next one is played).

Each voice generator can trigger a wave shape with harmonics plus sound envelope and tremelo or just play a sampled sound.

I note that you ‘do not say much about the 8bit voice generators’. It seems odd not to cover them…

They are covered in the PRMs in a way that 16-bit is not. I also don’t have much interest in them or experience with them. If someone else wants to do an ‘8-bit’ ROSS document it would be fine with me. But TBH my main focus is on shining a light into the 16-bit areas that haven’t been well documented and which will aid improved audio quality for purposes like hifi and good recordings.

There are also various potential traps and confusions which your comment example. In more modern terms the RO audio system has two channels – left and right – to produce stereo. And if software can generate stereo output it could place an arbitrary number of ‘positioned’ contributions by other means. So there is a risk of confusion over terms, etc.

I’m not trying to dismiss or criticise 8-bit and ‘VoiceGenerators’ per se.
I’m just personally adopting the approach that it is a different topic to the 16-bit parts of the ROSS I’m focussing upon. I’m afraid that it can be quite hard to cover even 16-bit with clarity and accuracy for a number of reasons. I’m delimiting the coverage to optimise what I can personally manage.

Jim

I’ve emailed Andrew Sellors at the address given on his webpage, but no reply as yet. I’d also email John Duffel if I knew his email address.

Afraid I’ve not had any reply to this. So still don’t know much about the modern Sound Buffers work started by John Duffel.

Jim

I have done a simple programme to play a sound sample on the Pandaboard (using the 8-bit voice generators). It takes a ‘.WAV’ file and generates a voice handler that will play that sample at the prevailing sample rate. A module is loaded that will play that sample using a SOUND 1,-15,&4660,40 command from BASIC at 88.2kHz.

The example ‘.WAV’ file is the sound of a bell produced using a programme called WaveAnal with the input frequencies ‘fudged’ so that although generated at 44.1kHz, it is designed to play back correctly at 88.2kHz.

Try it here

OK JIm, I found a little time to try to make a more useful file for you; I don’t have time to
a) create the code to generate the samples, so you get to work that out! maybe even load a sample file?
b) actually compile it, so there may be bugs. It is a simple hack of my working sound driver for Squeak though.
It’s intended to do a very simple play of ten samples of a quarter second each right now.
https://www.dropbox.com/s/3yigmrflis2g5o3/RISCOS-Sound-Demo.c

OK JIm, I found a little time to try to make a more useful file for you;…

I get “dropbox site requires javascript” when I use NetSurf on my ARMiniX. I’ll try FF on a linux box later.

Jim

I get “dropbox site requires javascript” when I use NetSurf on my ARMiniX. I’ll try FF on a linux box later.

You could try the direct download link here

You could try the direct download link here

Thanks! Yes, that worked. :-)

It occurs to me that I need to remember to add to the ROSS document some URLs for when people can find items like the SoundBuffers (and DigitalRenderer) modules, etc. If I’ve forgotten by the time I do version 1.00 of the document, someone please remind me!

Jim

It occurs to me that I need to remember to add to the ROSS document some URLs for when people can find items like the SoundBuffers (and DigitalRenderer) modules, etc.

Even better, whenever possible it would be nice to get them gathered up into the ROOL cvs tree where they become more visible, more accessible and maybe, just maybe, more likely to get updated.

Even better, whenever possible it would be nice to get them gathered up into the ROOL cvs tree where they become more visible, more accessible and maybe, just maybe, more likely to get updated.

Agreed. I’d be happy to see DigitalRenderer ‘adopted’ as well. Although I’m not sure what Chris Martin would think of that at present! :-)

Jim

Just to update people:

I’m currently doing an extended version of my ‘upsampling’ demo progs. Once that is done I will return to working on the ROSS document. I’ll try to add SoundBuffer coverage. Then set about turning the text into something more like English and correcting the typos, etc. As yet no-one has pointed out specific detailed errors and omissions in the technical content. So failing that happening I’ll assume for now that the content is OK.

Jim

Two updates.

1) I’m about to finish an article I’ve been writing for a magazine. Hope this will be sent off later today. Once it has gone I’ll resume work on the ROSS document and sort out the typos and my ‘english’ . Then put up a better-written version. I won’t add more sections at this point, but may do in the future.

2) In case anyone is interested, the ARMiniX and RO get a page in the latest issue of ‘Hi Fi News’ magazine. Page 119 of the “Yearbook 2013” issue. :-)

Jim