|
 Kyma Forum
 Tips & Techniques
 Modal Filter and Cross synthesis
|
| next newest topic | next oldest topic |
| Author | Topic: Modal Filter and Cross synthesis |
|
speedo Member |
posted 10 July 2010 21:13
 
Is there anywhere that I can learn more about how the modal filter and cross synthesis works and how I can create my own samples for processing? The prototypes reference doesn't have anything from what I can tell. Thanks! IP: Logged |
|
SSC Administrator |
posted 10 July 2010 22:33
You may have already tried this, but if you click on the name of the Sound in the lower right corner, it opens a window describing the Sound and each of its parameters. Do you have some questions about the Modal Filter and Cross Filter? IP: Logged |
|
speedo Member |
posted 11 July 2010 06:57
Yes of course, I'm very familiar with this by now  I'm not sitting in front of Kyma at the moment but I'm mainly interested in preparing my own excitation files for physical modeling. Would I just use RE analysis? I'll see if the info boxes discuss this when I get to the studio today. Edit: and yes, I meant cross filter, or whichever one has the turtle and rabbit icon [This message has been edited by speedo (edited 11 July 2010).] IP: Logged |
|
speedo Member |
posted 11 July 2010 12:27
OK, I've spent the last few hours studying up on RE Analysis and Cross Filtering. I've also checked out some physical modeled instruments from other users. I'm still pretty confused although making progress I want to create some hybrid instruments - for example, a vocal attack melded with a woody sound and then the sustain portion has the response of a water jug. Can I simply use pre-existing sounds that have similar characteristics and then plug in my own samples for excitation and response? Or do I need to go more in depth with the modal filters etc? I assume I need to prepare IR samples, correct? I tried using the RE analyzed files loaded into the Crossfilter's Response but it doesn't seem to work so well. Overall, I'm not quite sure how the IR samples are prepared and function. I understand how they work with convolution reverb, but not so much with voice or instruments. EDIT: so to make my ramblings more concise: I want to use pre-existing physical models of instruments as a template, but use my own recorded sounds to alter the characteristics. What is the best way to do this? [This message has been edited by speedo (edited 11 July 2010).] IP: Logged |
|
keph Member |
posted 11 July 2010 13:21
RE processed sounds won't provide you what you are looking for in the CrossFilter. Those are for use in specific sounds (the RE Resonator). Just use your unprocessed sounds in the response input. Btw, CrossFilter is not convolution, though it can produce somewhat similar results. Here is a bit of info on convolution including notes from Pete who wrote the CrossFilter Prototypes: You may want to look at the RE Resonator sound, it works well for time varying resonators (like voice or a mix of sounds like you mentioned). Also, since it has a Time Index function, you can scan through your analysis file at differing speeds which may be helpful depending on your uses. IP: Logged |
|
SSC Administrator |
posted 11 July 2010 13:47
quote: OK, one approach might be to get a water jug and make a recording of yourself hitting it with various objects (some sharp and hard, some softer). Look at the recording and find one that looks like a nice impulse response (should look like an attack with 2 or 3 reflections after it). Trim the sample to include just this one impulse. Create a FunctionGenerator to read that impulse sample. Feed it into the Response field of the CrossFilter. To test it, feed the rhythmic pulse generator into the Input. You should hear the response of the water jug each time you hit it with a pulse. Next, I would create a Mixer including a FunctionGenerator on another sample of the start of a vocalization, the CrossFilter, and (perhaps) the ModalFilter set up to simulate a wood block)? In answer to your question, there is no special preparation necessary for the Response input to the CrossFilter. You can feed in any Sound there (including a live input if you like). [This message has been edited by SSC (edited 11 July 2010).] IP: Logged |
|
speedo Member |
posted 11 July 2010 19:07
excellent. thanks for the info. I'll try this all out tomorrow. So an instrument IR is just a few cycles of the instrument's onset? Is it generally only for instruments with a fast attack? How, for instance, could I make an IR for a blown bottle? IP: Logged |
|
SSC Administrator |
posted 11 July 2010 19:16
The Impulse Response of a room or a water jug or a filter or any other system is the response of that system to a single impulse input. For example, if you have a filter and you feed it an Impulse (a signal that is a single 1 (followed by an infinite string of 0s)), the output will be that filter's Impulse Response. In the real world you can get close to this impulse signal by hitting something with a drumstick or a pen (or by shooting off a gun inside it if it is large enough, though I have only ever used balloon bursts). For example if you hit the water jug once, you can record its impulse response. Then if you feed different kinds of inputs into the CrossFilter (using the jug response), you can get different kinds of excitation of the water jug. For example, you might get something close to blowing over the rim of the jug if you feed low amplitude white noise into the CrossFilter (maybe?) Or if you feed different shaped pulses into the CrossFilter, you might get the sound of hitting it with different materials. [This message has been edited by SSC (edited 11 July 2010).] IP: Logged |
|
speedo Member |
posted 11 July 2010 19:43
That makes perfect sense. Thanks. I guess I was already familiar with the concept based on my experience with reverb. I'm surprised at the lack of information on the Google machine regarding IRs other than room responses. Seems there's plenty of application with reverb and very little discussion about other systems' and uses. IP: Logged |
|
CharlieNorton Member |
posted 12 July 2010 20:16
Balloon Burst! Fantastic Idea! :-) IP: Logged |
|
pete Member |
posted 13 July 2010 16:11
Another example of making a response is hitting the bridge of a cello with the strings dampened. This should give you that woody sound and using that as the impulse where the signal is a sawtooth (like the bowed string movement) should sound very cello ish. You could try recording your voice but just one pulse. That is open your mouth but close your throat. Then try to push air out (even though you can't because your throat is closed). Then open your throat fast and you will get one single impulse. It should sound a bit like a cough but much shorter and cleaner. Feed this into the crossfilters response input with a sample player making sure the sample trigger is the same as the impulse trigger in the crossfilter. Then put a pulse stream as the signal. Make sure that the pulse stream is pitch modulated and amplitude modulated and not too regular (like a voices pitch is) and you should get something like your own voice with a single sound. You could try to make different voice impulses by making different mouth shapes when you record you clean cough sound. hope this makes sense. Pete IP: Logged |
|
speedo Member |
posted 14 July 2010 20:10
Thanks Pete. I'll try that out. IP: Logged |
All times are CT (US) | next newest topic | next oldest topic |
  |
|
Contact Us | Symbolic Sound Home
This forum is provided solely for the support and edification of the customers of Symbolic Sound Corporation.
Ultimate Bulletin Board 5.45c