I'd like to synthesize a "ting" sound, sort of like a glockenspiel. Specifically, I'm trying to get something that sounds like this: http://dagobah.net/flash/ToneMatrix.swf though maybe with a little less echo.

Can anyone give me some advice? I've started by putting a sine wave through really high attack and decay. It's a start, but the sine wave still doesn't give me the sound I'd like.

You only need to square the amplitude envelope.

Experiment with this patch, (which I have written, somewhat clearly, for other Pd'ers to explore it too).

Download and, as always, be careful with your audio levels when adjusting feedback, no warranties whatsoever:
glockensting_01.pd

It should start playing as soon as you turn DSP on.

Landon is right about the squared envelope, try to use some fm synthesis too to get the metallic timbre you're looking for.

@coloscope Well, fm would be great. However, I analyzed the spectrum of the sounds at http://dagobah.net/flash/ToneMatrix.swf and they are just sine tones...a piece of cake !

Thanks Landon! I definitely wouldn't have thought of squaring the sine wave. I changed the envelope slightly, and it's now very close to what I'm aiming for. I would still like to get a little more depth, though. I tried [freeverb~] but it seems to dirty the sound. Any other suggestions?

Hi,

I tried Landon's algorithm with [freeverb~] using [roomsize 0.05( and [damping 0.9( and the results were quite nice (it sounds a bit like a mixture of a xylophone and a glockenspiel). I also modified the message fed to [vline~] from [1 0, 0 100 0( to [1 1, 0 100 1(, which makes the attack a bit less harsh.

I also tried using [rev3~ 110 50 100 80] instead of [freeverb~] and it also sounded all right to me. Give it a try and see if it fits your purpose.

Take care,
Gilberto

BTW, if you are interested, this is the patch I have been playing with using Landon's algorithm: 74 - glock.pd

Curiously, you (@gsagostinho) felt the attack was too harsh, and that some depth was needed (@LiamG). I felt that too, mostly the attack thing problem. So I analyzed the sound a bit further...

This is the data:

1. In the game, when you click on a light pad, there's a sound emission every ≈ 2 seconds.

2. An emission is formed by a main sound, let's call it W, and 3 echoes...let's call them w. So you can think of an emission as formed by 4 beats, the 1st one being stronger. So an emission is: Wwww.

3. W seems to last 500 milliseconds, and it has a typical decaying exponential envelope.

4. Obviously, there's some ping-pong delay effect. Its delay time ≈ 375 milliseconds (≈ 160 bpm).
   (This means that echoes happen every 375 milliseconds).
   

5. Phase is fixed. W always starts start in the positive domain.

6. Curiously, there is no (obvious) slow attack, W starts straightforwardly as a normal sine tone (and decays exponentially over 500 ms). So the attack is as fast as possible.
   

7. No reverb going on, or so it seems.

8. The stereo balance of W (i.e. "position"), is not easy to understand. It seems to be randomly drifting from left to right, a sort of random LFO thing? For example, clicking on the left zone of the game does not automatically produce a louder sound at your left ear.
   

CONCLUSION (rough!):
we may have felt the attack was too harsh and that there was little depth due to the absence of the (2-stereo) ping-pong effect. (Also, having seen flashing lights may have had some weird effect in the way we initially processed the sounds). So, if you want to recreate that specific kind of sound...the answer is neither slow attack nor reverb. Ping-pong delay is the answer.

@LandonPD

Really nice analysis, Landon

6. Curiously, there is no (obvious) slow attack, W starts straightforwardly as a normal sine tone (and decays exponentially over 500 ms). So the attack is as fast as possible.

I am not so sure about this as I can't zoom in in those images you posted. But the patch I shared before showed that an attack that grows from 0 until 1 in only 1ms was already much softer sounding than a sound with no attack. That means that even a very very very short attack period from 0 until 1 could make a difference here.

Take care and thanks for posting this!
Gilberto

Thank you, Gilberto!

Let me show you why I think that the sound designer of the game may have not applied a slow attack.

If you apply a slow attack, precisely a 1ms exponential attack, such as the following:

...to various tones, you get waveforms like the following ones (all within a single cycle time window):

The higher the tone, the easier it becomes to see and measure distortion (within those 1ms periods). (Here I'm using the term distortion meaning deviation from a sine tone, and not that it is a bad sounding thing!).

In the sounds of the game, there's nothing like any of the above waveforms. As I wrote, the sounds of the game have a frozen phase, W always starts start in the positive domain. I've only found clean repetitive sine tones. The sound designer just wanted to make a simple game, not dsp fireworks!

I hope these last pics help you to see what I've seen.

By the way, when I say "Ping-pong delay is the answer", there's no arrogance in it. That's just a rough conclusion, a quick one, for the sake of clarity and simplicity! There can be many answers, maybe the more answers the better. Ping-pong delay is just one of many possible answers.

@LandonPD Hi Landon,

My bad, you are indeed completely right. I was too focused on the image under your point no. 4 (which cannot be zoomed and could show some attack in the envelope) that I got completely oblivious to the meaning of the figure under point no, 6 (which had the answer all along). Thanks for pointing it out.

Take care!
Gilberto

Thanks for this excellent work, Landon! I've learned a lot from reading this. I'll be trying all of these things when I get some time (hopefully next week).

Is there any object ready made for ping pong delay? It would be easy for me to build it, but there wouldn't be any point if it is sitting there in an external.

It's such a pleasure to share these things with other people and work on it together.

@gsagostinho
Great! I'm glad that the last pics made things easier to understand.

@LiamG
Your're welcome, Liam. And it's a pleasure for me too.
As regards the ping-pong object, I don't have a classic one, but there must be some synth in this forum, or somewhere, that has one of those already baked.