| |
cadaver
Registered: Feb 2002
Posts: 1153 |
Testbit hardrestart - when first?
Anyone have opinion on this, when did it become popular to stuff $09 to the waveform register on the first frame of a note to stabilize the attack, in addition to pure ADSR hardrestart? And who perhaps invented it first?
At least in 1988/1989 many tunes were still content with using no hardrestart whatsoever. |
|
... 9 posts hidden. Click here to view all posts....
|
| |
cadaver
Registered: Feb 2002
Posts: 1153 |
Yeah, it's the "gatebit on" in $09 that makes the difference. Of course it's only a kind of preference, to shorten the attack, oldschool sounds become impossible :( |
| |
Laxity
Registered: Aug 2005
Posts: 459 |
My understanding exactly!..
Hard restart is:
$d404 set to get off (fx. $40), while the ADSR is set to something like $0f00 or whatever.. A hard release at least for two frames!
After that at least two (or is it three.. don't remember) frames of gate on. Sound will not be audiable before about 25-30ms after the gate on, which will make the first frame of the sound really short and hard!.. Isn't it?.. Argh.. It's been a while since then, although I recently wrote a new player .. actually a JCH new player 21.gX.. think I got to 21.g3.. Well.. sidestepping into the insignificant there.. sorry!
The oscillator reset is.. eh.. resetting the oscilator, which will reset the random generator for the white noise generation.. This will make any noise produced sound the same after the reset was issued.. or am I wrong?.. I was actually never a big fan of this feature.. Did I really have that in my player per default?.. Hmm..
Anyway.. Hard restart is a wonderfull thing.. Love it, but imperfection can be a very nice thing with music. I think I overdid the HR in the past :) .. Actually everything was hard restart in my last player.. anno 1989.. (Jezz.. that's long ago..)..
Gotta take care of the kid... (that's jr.. not me .. hehe) |
| |
Laxity
Registered: Aug 2005
Posts: 459 |
ViruzTracker?.. Ah.. sound's good!.. Looking forward to trying it out.. If I may, ofcourse :).. Is it for public release? |
| |
Soren
Registered: Dec 2001
Posts: 547 |
Laxity:
I as far as I understood it, Oscillator Reset just makes sure that a wave starts from the same offset phase.
About the ViruzTracker, yep it will be released later on, so it would make me happy if you would test it a bit :-)
|
| |
Frantic
Registered: Mar 2003
Posts: 1627 |
Laxity said:
After that at least two (or is it three.. don't remember) frames of gate on. Sound will not be audiable before about 25-30ms after the gate on, which will make the first frame of the sound really short and hard!.. Isn't it?.. Argh.. It's been a while since then,
---
I don't think this is correct. The 25-30 ms (or rather one frame I think) "delay" is caused by the specific way of combining ADSR-hardrestart with oscillator reset that's commonly used, I think. It looks like this:
frame1: Gate off, ADSR = 0f00 (or so)
frame2: same..
frame3: testbit on, gate on, wavefom set, new ADSR set.
frame4: SOUND TRIG! (by unsetting testbit)
...because the ADSR is actually trigged in frame 3 already and not at the same time as the sound will become audible (frame 4) because testbit is "released" then and the oscillator starts.
So, I think it's wrong to state that ADSR-hardrestart makes the sound start 25ms "into" the ADSR envelope. I think it simply is more about making ADSR stable (by giving the internal counters time to "count out" to avoid the ADSR bug).
If gate on wasn't set until frame 4, then there would be no extra-snappy attack. Rather it would just be stable.
|
| |
Laxity
Registered: Aug 2005
Posts: 459 |
Autoboy: I might be mistaken.. But you don't need to reset the osc. to get proper hard restart.. or maybe we're not talking about the same thing...
If I do hard restart without resetting the oscs, noise (wave $81) just comes out a little different... I usually don't use the reset thing very much but hard restart I use - even a lot! ;)
Jeff: It also resets the noise generators, I'm pretty sure about this.. I could send you an example.. if I only had you email adress :) .. Noise played after an osc. reset sound a lot more metallic than without it.. |
| |
Frantic
Registered: Mar 2003
Posts: 1627 |
Laxity: No, we don't seem to talk about exactly the same thing. I just meant that the way many ppl use the testbit for osc reset when they (also) do hardrestart, makes the hardrestart even snappier, since in effect ADSR begins one frame before the oscillator is turned on. |
| |
Laxity
Registered: Aug 2005
Posts: 459 |
I don't think the osc. reset affects the ADSR engine of the SID at all..
This would still produce a stable ADSR:
frame 1: gate off, ADSR to 0x0f00
frame 2: no change
frame 3: gate on, waveform $01 (not $09)
frame 4: waveform $81, freq to something etc.
You'll not hear the sound that was gated on on frame 3, before well into frame 4 - and the ADSR is stable... right?... But in this case the oscs were not reset.. They would have been if gate on value was $09..
Do we agree on this?.. (don't mean to be snotty!.. Just want to make sure we take about the same thing) |
| |
Laxity
Registered: Aug 2005
Posts: 459 |
I think I might be mixing two things together here.. I'll have to dig back in the old disks.. |
| |
Laxity
Registered: Aug 2005
Posts: 459 |
Autoboy said:
[cut]
...because the ADSR is actually trigged in frame 3 already and not at the same time as the sound will become audible (frame 4) because testbit is "released" then and the oscillator starts.
So, I think it's wrong to state that ADSR-hardrestart makes the sound start 25ms "into" the ADSR envelope. I think it simply is more about making ADSR stable (by giving the internal counters time to "count out" to avoid the ADSR bug).
-------------------
Actually it's not wrong.. The delay of attack (of app. 25-30ms from gate on) is caused by the ADSR engine being at a complete halt.
I've been investigating this a bit, and bit 3 of the waveform reg. ($d404) does reset the oscillators. That means that the whitenoise randomizer is reset as well, and therefore creates a consistent noise that will be exactly the same every time triggered after an osc-reset. In fact, the noise is so consistent that you can actually distinguish a tone in it - and it sounds rather metallic.. But, the oscillators are not delayed when turned on, as suggested (as far as I can tell)!.. The gate on delay (of app. 1½ frame PAL - 25-30ms) that one expiriences with the hard restart is however caused by the ADSR engine of the SID, thus the hard restart has nothing to do with bit 3 of the waveform... It is however true that you can have a seemingly harder hard restart by applying the bit 3 reset as the whitenoise generator is reset and noise is more "metallic" and consistent... I haven't tested this, but I wouldn't suspect that I'd not be able to hear the differance wether or not the osc reset was applied to anything else but an instrument using a noise waveform in the first frame after... |
| Previous - 1 | 2 - Next |
