yeah, that would work.
will attach the diytanh anyway
-
DIY2 - Effects, Sample players, Synths and Sound Synthesis.
-
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?
-
yeah, all filters would need some degree of volume compensation, i guess.
BTW, the RIGHT outlet of the vcf~ filter is actually for resonant lowpass, so you can just modify st-vcf~ if you want res lowpass filter
-
Sorry, mod... I didn't understand about the "RIGHT outlet" of the vcf~...
Anyway, I would like to know if I did the correct FX "mix". I wanted a LPF and HPF with resonance. So, I put, in series, a st-lowpass (or st-hipass) and st-vcf.
No lowpass, no resonance:
lowpass cutoff 1
vcf cutoff 1
vcf resonance 0Low pass and resonance:
lowpass cutoff (goes to) 0
vcf cutoff (goes to) 0.5
vcf resonance (goes to) 0.8The same with hipass (cutoff inverted).
Because, with vcf alone, I can't get lowpass and hipass with resonance. And, with this mixture, I get a good FX.
-
i'd try to use the right outlet of vcf~ for your res lowpass, but for the highpass maybe that way you're doing it will be ok.
-
Hey, mod... I post here a thread in pd-list, because I want to use your compressor... and I want your opinion.
For my live looping-system, I do beatbox, using a Shure58. The mic has very good quality, but I realized that the kicks and snares, are recorded very loud (max signal = 4). So, I want to use a compressor (I though using tanh(), but some people told me that it distorts a lot and extra harmonics).
I want to use your compressor (almost my FXs uses DIY2 effects). After several tests, I concluded that it is very difficult to compress the first peak of a kick, for example. Then, the audio is compressed... but the first peak is still there.
When trying to save the original sound to disk, Pd has to normalize from 4 to 1. And, compressed with your compressor, too: 4 to 1 (more or less).
So, I thought about using the compressor with attack and release at lowest and then, for those peaks, use something like [expr~ (tanh($v1/1.5))*1.5], not to cut to 1, but 1.5. So, It don't distort the sound so much.
What do you think?
-
i think i answered this on pd list? you ok now?
