News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
T

WIP: Now That's What I Call Chip Tunes (WinAPE)

Started by tastefulmrship, 11:10, 29 June 11

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

tastefulmrship

Is there a specific reason why the AY-chip volumes are log values rather than linear progression? And is this the same for the Spectrum/AtariST/etc chips as well?

Plus, I agree with Executioner; I don't remember the CPC volumes being so 'top heavy' back in the 80s/90s and that was in the peak of my sound/music programming on a native machine. (All my olde music is lost to time, now... but is no real shame as it was all crap, to be honest!)

Devilmarkus

Yes, the AY D/A conversation is logarithmic.
You can see this clearly in the AY's datasheet:
When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

MaV

Quote from: tastefulmrship on 08:40, 05 October 11
All my olde music is lost to time, now... but is no real shame as it was all crap, to be honest!

I don't buy that anymore, since you've repeatedly given nice inputs to various topics.

Ye olde musick may have been good, and a loss it is that it doesn't exist anymore.
Black Mesa Transit Announcement System:
"Work safe, work smart. Your future depends on it."

Executioner

Quote from: Devilmarkus on 08:59, 05 October 11
You can see this clearly in the AY's datasheet:

That may well be the case, but in reality, there's a combination of channels on each output, some other electronics before the output gets to the speaker, and who's to say that the D/A converter in a modern PC is totally linear? The only way to really get accurate sounds levels would be to measure the decibel output from each channel on both a real CPC and emulator at each different level, and then I suspect that would vary quite a bit between PCs.

Devilmarkus

#54
Quote from: Executioner on 22:39, 05 October 11
That may well be the case, but in reality, there's a combination of channels on each output, some other electronics before the output gets to the speaker, and who's to say that the D/A converter in a modern PC is totally linear? The only way to really get accurate sounds levels would be to measure the decibel output from each channel on both a real CPC and emulator at each different level, and then I suspect that would vary quite a bit between PCs.

Sure, true.
But somewhere we must find a way to produce the output as realistic as possible.
When you can trust the docs, it also says, that "noises" and "tones" aren't affected by this curve.
So perhaps a mix of both would be the good way?
I really have no idea but the sources from AY_Emul use these tables and I thought this couldn't be too bad to adapt for CPC emus...
In fact, JEMU outputs values between 0 and 15 here, so why not put them into an array of logarithmic values? ;)
When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

TFM

Well, the human sense of hearing works in a logarithmic way, so it makes sense to keep devices logarithmic too. Actually decibel is a logarithmic unit. I don't even know linear stuff by myself, but I'm not a PC freak. Love the CPC  :-*
TFM of FutureSoft
Also visit the CPC and Plus users favorite OS: FutureOS - The Revolution on CPC6128 and 6128Plus

robcfg

I would start by making a small program that performs some tests with the AY so it can be recorded and compared to other recordings.


For example, choose a fixed frequency like 1000hz and play it at all the volume levels, and then play some mixed channels.


The idea is that more people can record the test on different machines and compare the results to get sort of a standard measure (should that be possible).

tastefulmrship

Do you mean something simple like this?
(attached below)

Obviously it needs more tones/envs/noises to properly test the AY, but what do you think as a starting point?
The resulting .wav file will be pretty big, though. Maybe .mp3ing the file (stereo, 128k) once completed?

Devilmarkus

I think another big problem is the difference between the 464 and the 664/6128.
Where the 664/6128 uses volumes 0-15, the 464 just uses volumes 0-7!!!
The 464 simply repeats them!
0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7 -> 16 values.
The 664/6128 has 16 real values:
0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 -> 16 real values...

This makes it very hard for sound developers to create things working on all machines the same way!!!
When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

tastefulmrship

Strange. I used to program 'music' on my CPC-464 and always used ENV 1,15,-1,5; it always sounded like it decayed 16 times to silence.
There was never a time when I thought it was just playing 2x8 volumes.

Devilmarkus

Well try this with an emulator using 464 and also 6128 roms:
10 FOR t=0 TO 15:SOUND 1,400,100,t:CALL &BB18:NEXT


When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

tastefulmrship

#61
Quote from: Devilmarkus on 12:43, 06 October 11
Well try this with an emulator using 464 and also 6128 roms:
Ok, this is surreal! Try my -TESTSND.BAS proggy with 464 ROM. It plays normally. And ENV 1,15,-1,5 works too. It just seems to be the volume in the SOUND command itself that's affected, not the AY chip (as such).
Good olde Amstrad! ^_^


EDIT: Here's a fun one, then, for ya!

ENV 1,15,-1,50

now in 464 use
SOUND 2,100,800,15,1
and it slowly decays to silence...
however, type in this
SOUND 2,100,800,7,15,1
and see what happens.

!WEIRD! Especially when compared with the program supplied (above) by Devilmarkus.

Executioner

Quote from: tastefulmrship on 12:56, 06 October 11
It just seems to be the volume in the SOUND command itself that's affected, not the AY chip (as such).

Actually, it's almost as if Locomotive mis-read the AY documentation when they did the 464 OS and thought there were only 8 volumes, then fixed it forr the later versions. It's the same AY and 99% of games do direct I/O to the 8255/AY so it doesn't affect most games.

tastefulmrship

#63
Time to murder some more classic Amiga chiptunes as only I can do!
I present to you all "This Musicdisk is now about 20% Cooler!", as it features my all-time favourite pony of all time!




- Additional credits;

(O) Apidya (The Pond) arranged by Rapture (Fabian Del Priore).
(V) Battle Squadron (Highscore) arranged by Okeanos (Olivier Degand).
(W) Uridium 2 (Loader) arranged by Dreamfish (Ian Ford).

Rainbow Dash image painted by Nethear (http://nethear.deviantart.com/gallery/#/d419usr)
Music converted using STarkos v1.2 by Targhan (Julien Nevo) on WinAPE2A18 by Executioner (Richard Wilson) and then edited using Arkos v1.0 by Targhan (Julien Nevo) in mono mode.

Music replay routine by Targhan (Julien Nevo).
Screen squashed using BITBuster1.2 by Team Bomba, Gasman (Matthew Westcott) and TomEtJerry (Hervé Monchatre).

The tunes have been tested and tweaked by MaV on real hardware. Also thanks to Martin and Jelena for deciding the background colour for me after I found a bunch of fuck-ups in my pixelling! (@rexbeng; you will not be surprised by that! ^_^)


- Notes;

If possible, please listen to these through the internal speaker and not through a stereo. You people told me to use real-hardware only, now I'm telling you to listen to these via real-hardware only! (Ironically, none of them have even seen real-hardware during conversion!)

And if you don't like ponies, then tough! Grow up and move on before we Bronies love and tolerate the shit out of you! Then you'll have something to complain about, yeah? Right!


Well, I'm off for now!
- JTMS...

Devilmarkus

#64
Nice compilation!

Added it :P


http://java.cpc-live.com/demo.php?id=-T-&title=thismusicdisk

Edit: This fucking javascript to show images corrupts the URLs you add...
When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

rexbeng

Yaaaa! I like the Rainbow Dash (that's her name, isn't it? If not I humbly apologize) theme a lot more to the Bart one. She fulfils my thirst for color and, oh my, she fits soooo well with the CPC palette  ;D


Now I have to re-listen to all the tracks since for me this is a whole new production Λ_Λ <3


rb

TFM

#66
Does the CPC really a song with the title of H?                       CPC rulez - period!
TFM of FutureSoft
Also visit the CPC and Plus users favorite OS: FutureOS - The Revolution on CPC6128 and 6128Plus

MacDeath

#67
sorry to be such a ponerd but...
Rainbow dash if properly paletted, has to get the inked lines of the body a different shade than the lines of the mane.
(sorry can't remember the exact name in english)... line shape ? contour lines ?


ok, outlines...


Also I downloaded your screenshots and they are like shittons of colours, so despite not being jpg it is as if.


anyway, there I did a mock up.
[attachimg=1]


To me, with CPC palette it means you use the sky blue for hair's outlines, and  Pastel Blue for body outlines... perhaps DarkCyan could work too...


I kept the Sky blue for darkened/shadowed parts...


Of course some other colours may be suitable, but my point is that a proper Rainbow crash need to differenciates those outlinings and the CPC palette can actually enable it.


Perhaps something could be done concerning hilights with Cyan and pastel cyan...  : cyan for shadowed parts, plus dithering of course, and pastel cyan for the foregrounds parts of the body.


Bro-Hoof.


MLP FIM: Proud to be a Brony [PMV] (Fixed Audio)



Post Edit : oops my bad, sorry, forget what I told about the high number of screwed colours, I used the JavaCPC picture, and JavaCPC emulate a CRT (somewhat) so is not quite "clean"...you actual picture is clean, with only 16 colours.
Still I maintain what I told about the outlines...
Of course it can be even more accurate on a PLUS, pastel blue being used because there is only one shade of "Skyblue" (which is cross-colour between cyan and blue).


Applejack is a pain in the Ass on CPC-Old though, due to the limited range of Oranges and browns..


Next time you'll choose Fluttershy, she takes less colours and is 200% cuter.

TFM

TFM of FutureSoft
Also visit the CPC and Plus users favorite OS: FutureOS - The Revolution on CPC6128 and 6128Plus

khisanth

Bloody excellent work mate!! This is simply superb.  More more more!!!  :D

Has anyone done C64s Wizball on the AY chip?

ivarf

Quote from: Executioner on 00:14, 07 October 11
Actually, it's almost as if Locomotive mis-read the AY documentation when they did the 464 OS and thought there were only 8 volumes, then fixed it forr the later versions. It's the same AY and 99% of games do direct I/O to the 8255/AY so it doesn't affect most games.


Ah, thats explains why I never got anything decent out of the CPC when playing with the envelope-commands back at the time. (or perhaps lack of talent :) )

ivarf

 At the time real hardware was for me a CPC connected to amplifier and speakers. My impression is that more sound was made for that setup in the earlier games. Later games often had over the top treble, probably for the internal speaker,  which sounded awful on a decent speaker setup. Still, I see what you mean :)

Devilmarkus

Quote from: MacDeath on 01:15, 18 October 12
Also I downloaded your screenshots and they are like shittons of colours, so despite not being jpg it is as if.

Good, then I did a good job ;)
They are modified with glossy and a link to prevent image stealing ;)
When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

tastefulmrship

#73
Quote from: MacDeath on 01:15, 18 October 12
To me, with CPC palette it means you use the sky blue for hair's outlines, and  Pastel Blue for body outlines... perhaps DarkCyan could work too...
If you look at the top of RD's mane, you'll see that I did originally use PASTEL BLUE (#14) for the outlines; I missed a few pixels and they are still using #14.

I changed all the outlines (apart from the few pixels I missed) to SKY BLUE (#11) because I was originally going to change PASTEL BLUE to another colour for the music text... but never got around to it. So A, H, O & V are still using #14.


Quote from: khisanth on 01:31, 30 October 12
Has anyone done C64s Wizball on the AY chip?
Yes, I am currently working through the entire .sid file, but a lot of the music uses some unique tricks that Rob Hubbard and Martin Galway used regularily back in the day. For example; manipulating the PULSE to create (what sounds like) two notes from one channel. If you use the old version of SIDPlayer that allows you to split the voices into stereo, then you can hear this being used at the beginning of the Spellbound tune (by Rob Hubbard).

Also, Martin Galway seems to be running the SID at x6 speed (ie 6 notes played per frame) as previously discussed in the ShoutBox in relation to the end music to this C64 demo; http://www.pouet.net/prod.php?which=60631

I will continue to trudge through it... expect a release eventually!

arnoldemu

Quote from: tastefulmrship on 15:21, 01 November 12
If you look at the top of RD's mane, you'll see that I did originally use PASTEL BLUE (#14) for the outlines; I missed a few pixels and they are still using #14.

I changed all the outlines (apart from the few pixels I missed) to SKY BLUE (#11) because I was originally going to change PASTEL BLUE to another colour for the music text... but never got around to it. So A, H, O & V are still using #14.

Yes, I am currently working through the entire .sid file, but a lot of the music uses some unique tricks that Rob Hubbard and Martin Galway used regularily back in the day. For example; manipulating the PULSE to create (what sounds like) two notes from one channel. If you use the old version of SIDPlayer that allows you to split the voices into stereo, then you can hear this being used at the beginning of the Spellbound tune (by Rob Hubbard).

Also, Martin Galway seems to be running the SID at x6 speed (ie 6 notes played per frame) as previously discussed in the ShoutBox in relation to the end music to this C64 demo; http://www.pouet.net/prod.php?which=60631

I will continue to trudge through it... expect a release eventually!
You can do 6 speed ays on cpc too? run the player every interrupt.

I thought starkos and arkos support that?
My games. My Games
My website with coding examples: Unofficial Amstrad WWW Resource

Powered by SMFPacks Menu Editor Mod