here dude, check this one:
http://puredata.hurleur.com/sujet-3520-speech-formant-synthesizer-bonus-turkish-vowels-pack
-
DIY2 - Effects, Sample players, Synths and Sound Synthesis.
-
actually, i just checked with cycling the delay back through the formant, and it didn't work that well... sorry, i think you had it right in the first place. formant, then delay, seems best.
-
mod, listening for the real KP3, I thinkt that the way to emulate it, is in that way. Because the formant filter is not modifying the feedback, only the delay (the first).
I had seen the vowel vocoder link and learned about it with some videos. But, I would need the formants of more materials. Or, any other way to get that "sound". I don't know, by adding another filter... what filter?
Let me understand... material filter is just an equalizer, with particular parameters?
PD: looking for vocoders, I found one you made, but diytanh.pd is not in the ZIP. So, it's not working. Can you upload in that thread?
-
this should do it.
although [tanh~] is fine if you don't mind using externals.
sorry on the formant advice. as i said, it's not something i have looked into that much.
-
Well... actually, it needs diytanh.pd, not diytanh~.pd (diytanh~ abstract is as a subpatch there).
tanh~ is an external... and what about [expr tanh($f1)]?
-
yeah, that would work.
will attach the diytanh anyway
-
aaahhh!!!
don't use that one, it doesn't work for negative numbers.
attaching fixed patch
-
Hey, MOD. I'll have a lot (someday, more than 50) FXs opened at the same time. All of them, except one (the one it is working), have [switch 0], to avoid processing signal. But the messages are being processed.
I need to know if any DIY2 FXs is consuming so much CPU using only its message connections.
Sorry, but I will keep asking about your FXs... because, as I told you, I'm using them a lot.
Thanks!
-
I think the amount of CPU used by messages should be rather minimal, BUT if you open 50 effects as they are, with the GUI attached, then the GUI redraws from all 50 modules is going to drain a lot of CPU power.
Most of these patches have a subpatch with the audio guts, so if you want to use so many at once, it might be an idea to separate out just the audio section, and then use a generic GUI that can be switched for all your effects. It's a lot of work, i know...but if saving CPU is your goal, then it would be the best way.
-
mod, I know that GUI consumes CPU... so, the main patch only shows the name of the FX. Look at the PNG.
And, the patch where all the FXs are, is the other capture... and it is a subpatch, inside an abstract (inside the patch that has the blue pad). Also, even each pad isn't modified, it is only a test pad, only with outputs, to use inside the FX.
I also attach a FX abstract example capture.
-
mod, I know that GUI consumes CPU... so, the main patch only shows the name of the FX. Look at the PNG.
And, the patch where all the FXs are, is the other capture... and it is a subpatch, inside an abstract (inside the patch that has the blue pad). Also, even each pad isn't modified, it is only a test pad, only with outputs, to use inside the FX.
I also attach a FX abstract example capture.
-
mod, I know that GUI consumes CPU... so, the main patch only shows the name of the FX. Look at the PNG.
And, the patch where all the FXs are, is the other capture... and it is a subpatch, inside an abstract (inside the patch that has the blue pad). Also, even each pad isn't modified, it is only a test pad, only with outputs, to use inside the FX.
I also attach a FX abstract example capture.
-
Hey, mod... how should I make the phaser and/or the flanger having more that "flanger noise" ("Swiiisshhhhhhwwwwwiiissshhhhhh...")?
Raising the Feedback supposedly would do that... but the sound becomes "robotic"... and I don't like it.
One more thing! As the FXs I want to emulate are, I need both, flanger and phaser in MANUAL PHASE (or delay time, or frequency). So, I modify the flanger patch and replace this:
[r $0-freq]
[expr pow($f1, 2)]
[+ 0.1]
[phasor~]
...to this:
[r $0-freq]
[sig~]
...The phasor is similar. Is this right? Is there a better way to do it?
Thanks!
-
hay man, sorry... i only have a minute on the internet right now... and i'm just about to fly off to Australia for a month. Won't be checking very much while i'm there.
you can just direct your questions at everyone... there are plenty of people here who can help you as well.
all the best
-
Thanks, mod. I posted in a new thread.
-
After several tries using DelayFB, mono and stereo... I realized that they have a bug.
When a sound enters in the abstract, it cames out at the same volume (direct). Then, it starts the delay-feedback... it should start with feedback applied... but it doesn't do that. The first delay is at the same volume as the direct.
The fix is by changing only a connection inside [pd working]:
This:
[inlet~ signal]
|\
| \ [*~] .... (receives from vd~ and $0-feedback)
| \ /
| [delwrite~ $0-fbbb 1000]
|
...To this:
[inlet~ signal]
|\
| \
| [*~] .... (receives from vd~ and $0-feedback)
| |
| [delwrite~ $0-fbbb 1000]
|
...So, the first feedback has the volume decrease.
Attached mono and stereo version.
If I'm mistaken, PLEASE, LET ME KNOW!
-
After several tries using DelayFB, mono and stereo... I realized that they have a bug.
When a sound enters in the abstract, it cames out at the same volume (direct). Then, it starts the delay-feedback... it should start with feedback applied... but it doesn't do that. The first delay is at the same volume as the direct.
The fix is by changing only a connection inside [pd working]:
This:
[inlet~ signal]
|\
| \ [*~] .... (receives from vd~ and $0-feedback)
| \ /
| [delwrite~ $0-fbbb 1000]
|
...To this:
[inlet~ signal]
|\
| \
| [*~] .... (receives from vd~ and $0-feedback)
| |
| [delwrite~ $0-fbbb 1000]
|
...So, the first feedback has the volume decrease.
Attached mono and stereo version.
If I'm mistaken, PLEASE, LET ME KNOW!
-
I realized that all the FXs gain are initialized to 0.49. Why?
For example, in my patch, I change lowpass value with a tablet pad. When I put my finger in the pad, the LowPass is activated. But, if I set cutoff to 1, it would sound the same as not pressing the pad... but the volume is lower. That's because the 0.49 default gain.
Is there an explanation of that 0.49?
-
the gains are mostly log scaled from 0.1 to 10, so that a value of 0.5 on the slider corresponds to a gain of 1.
the reason i used 0.49 instead of 0.5, is that i didn't like the appearance of the GUI sliders when it was exactly 0.5 (the slider line goes thicker)
a non-resonant lowpass only cuts frequencies without boosting anything, and the DIY lowpass uses 4 non-resonant lop~ objects in series. Pure bass tones shouldn't be affected too much, but higher frequencies will suffer some loss even with the cutoff set at max.
If you're making a kaoss pad style effect, then maybe you could do a hack to reduce the effect 'amount' when the cutoff is set very high. So, for example, when the cutoff slider goes above 0.75, the 'amount' is automatically reduced from 1 to 0 over the remaining range.
-
I knew that the slider gets thicker... and I though that it could be because of that...! I modified to use gain in 0.5 (I like thick line) and I understand what you say about the amount (maybe I'll hack that).
But then, I realized that the volume was lower because I was using st-lowpass with st-vcf (in series) to get the resonance sound. And, with cutoff at 1 and resonance at 0... st-vcf outputs with lower volume. I had to set gain in 0.6 to get the same volume.
Is that... normal?
