Signal Types Debugging Patch: Questions + Feedback

[ambian]

Hey,

I’m making a signal types debugging patch for myself. It includes 4 patches, each related to a signal type. I am uploading the first version of the patch below.

signal-types-alpha-01.audulus (545.3 KB)

The purpose of the patch is for me:

• to have an easy way to get a signal reading that I can easily understand (no matter what module the signal comes from)
• to have an easy reminder of how the different signal types work

It would be great if you could answer the questions I have + share any other feedback on what I am doing.

Please note that I did not know about the livestream on signal types and I am yet to watch it (I will soon)… Metaphorically speaking: I’m the type of person who tries to build everything without the manual first.

Patch 1: Gate

The readouts here are pretty straightforward: BPM, value, waveform meter.

The question I have is: why do I need to output from 1/4 to get the BPM in my zerocross converter? Is the idea that each beat has 4 steps in it? That makes sense to me… I just want to make sure, that’s the case and it’s not an error in the patch/logic.

Patch 2: Modulation

This was made by @biminiroad to illustrate what happens when modulation signal goes into a gate input. It’s a great reference IMO. I just added the explanation if why the waveform meter at the top right is a flat line. More details on this here.

No questions for this one. It’s all clarified in the thread mentioned above.

Patch 3: 1/Octave

Here I am simply outputting the signal to a value meter and a waveform meter (normalizing it to fit in the meter first). The question I have is:

Shouldn’t I be able to produce values between -4 and -5 with the Mod->Oct module? -4 is the lowest I can get it to output.

Patch 4: Audio

This needs the most work… Maybe I can share what I would like this part to do and you can tell me how to accomplish it?

I would like to have 2 meters:

1. a level/amplitude meter. This is simply to measure if the signal I’m outputting is too hot. What is the best way to do this? Is the meter I’ve thrown in enough? It would be great if I knew what value the ceiling of the meter is?

2. Pitch… I know this is way harder to visualize… and I also have ears for this… but it would be great if I could visualize pitch within the waveform/oscilloscope meter (without the signal going above/below the waveform meter) in a way that 's visually comprehensible.

A) To stay within the floor/ceiling of the waveform meter I just need a math expression (or a module which converts the signal to -1 to 1 range) correct?

B) As far as it being visually comprehensive I assume I would need some controls for the oscilloscope’s time range? I would basically need to be able to zoom in (the time) to get a useful visual for anything other than a continuous pitch without an envelope.

I feel like someone shared something along these lines (new oscilloscope for Audulus 4 maybe?!) the other day. But I can’t remember the details.

Your input is greatly appreciated!

If I can get all 4 parts to being as useful as I’d like them to be, I promise to record a short video tutorial for beginners like me… Which explains how to use this patch as a learning tool + to debug signals in other patches.

[stschoen]

@ambian

The question I have is: why do I need to output from 1/4 to get the BPM in my zerocross converter? Is the idea that each beat has 4 steps in it? That makes sense to me… I just want to make sure, that’s the case and it’s not an error in the patch/logic.

The master clock module assumes 4/4 time so each quarter note gets one beat. The 1b output is one bar not one beat

@ambian

Shouldn’t I be able to produce values between -4 and -5 with the Mod->Oct module? -4 is the lowest I can get it to output.

-4 is the lowest offset for the module att*8-4 when att is 0 expression is -4

The meter is 0 to 1. Input is rectified and some smoothing. It’s often useful to have a clip indicator since the meter is top stopped at one.
Waveform display is -1 to 1 with a fixed scan rate.
Here’s some VU meters I built https://discourse.audulus.com/t/mono-stereo-vu-meters/88?u=
And a scope hack https://discourse.audulus.com/t/waveform-node-not-working/419/17?u=stschoen

[ambian]

@stschoen

The master clock module assumes 4/4 time so each quarter note gets one beat. The 1b output is one bar not one beat

Great. Thank you!

@stschoen

-4 is the lowest offset for the module att*8-4 when att is 0 expression is -4

shouldn’t the expression be att*10-5 instead of att*8-4? the module description states that it translates 0 to 1 modulation signals into -5 and 5 octave signals… shouldn’t that mean that the module can produce a -5 value?

I don’t fully understand the description below.

@stschoen

The meter is 0 to 1. Input is rectified and some smoothing. It’s often useful to have a clip indicator since the meter is top stopped at one.
Waveform display is -1 to 1 with a fixed scan rate.
Here’s some VU meters I built Mono & Stereo VU Meters

EDIT: I included the VU meters and fixed the level/signal range in this part by using a level node instead of attenuator (what I had in there in the 2nd version).

@stschoen

And a scope hack Waveform Node Not Working

The scope hack is also what I was talking about in #4 B) in my post. Included them in my patch.

Unfortunately I don’t fully understand how the upper oscilloscope version works. I cannot tell what the purpose of the Hz and o inputs is in this context.

I’m attaching the version of the patch from the screenshots above [with my att*10-5 mod included hah].

Feel free to modify it.

EDIT: I am including the version with the fixed level/signal range for the audio signal.

signal-types-alpha-03.audulus (288.8 KB)

Thanks!!

[stschoen]

The VU meters are calibrated in dBs with +5 as clipping. They are logarithmic. There is a clipping indicator above the meter. The built in meter node is linear with a range of 0 to 1. a is an audio signal. You usually attenuate the audio using either an expression or more commonly a level node. For oscillators without an envelope, input follow the oscillator with a level node and modulate the level node.

To “o” input on the scope is a 1 per octave signal. You use it or the Hz input to set the sweep frequency of the scope. The other variant uses a zero crossing detector to set the frequency. It’s not as accurate for some waveforms.

[biminiroad]

I updated the modules and it seems like -4 to 4 is a more useful musical range, so I changed it to that but didn’t change the description inside yet

[ambian]

Haha. Now this makes sense… So if I want to have it in the 10 octave range are the edits I made to the patch (screenshot below), all I need to make the adjustment (back to 10 octaves)?

[ambian]

Let me start this reply by saying that the scope and VU meters are incredibly useful and are a complete lifesaver for me! Thank you again for making them!

I just want to make sure I fully understand how to use them.

@stschoen

The VU meters are calibrated in dBs with +5 as clipping. They are logarithmic. There is a clipping indicator above the meter.

1. If they are logarithmic wouldn’t the labels be non-linear (ex. 0, -5, -10, -20 instead of 0, -5, -10, -15)?
2. You mentioned that +5 is where the clipping occurs, but visually I see the clipping meter turn on around halfway between 0 and +5. See below. Am I reading it incorrectly?

not clipping:

clipping:

@stschoen

To “o” input on the scope is a 1 per octave signal. You use it or the Hz input to set the sweep frequency of the scope. The other variant uses a zero crossing detector to set the frequency. It’s not as accurate for some waveforms.

So the two additional inputs for the top scope meter (Hz and o) both set the sweep frequency just in different formats (and they are meant to be used as alternatives depending on context, not both at the same time)? And the reason for having a separate frequency input is that in some cases this method produces a more accurate waveform than determining the frequency using zerocross on the audio signal?

@stschoen

You usually attenuate the audio using either an expression or more commonly a level node. For oscillators without an envelope, input follow the oscillator with a level node and modulate the level node.

This makes sense.

[stschoen]

Because the dB scale is also logarithmic the markings are linear. VU meters do not display the peak signal, they display a filtered version. The *5 refers to the peak signal. I think the clip light comes on at 0.995 just a bit before the output clips. I originally had 0 VU as clip but I changed it to be closer to traditional VUs
You’re correct about the scope. With complex waveforms zero-crossing doesn’t work well. One input or the other

[ambian]

The scope meters now make sense.

@stschoen

Because the dB scale is also logarithmic the markings are linear. VU meters do not display the peak signal, they display a filtered version. The *5 refers to the peak signal. I think the clip light comes on at 0.995 just a bit before the output clips. I originally had 0 VU as clip but I changed it to be closer to traditional VUs

it sounds like you’ve done them consistently with traditional VUs. I just don’t understand how VU meters work well enough to follow the logic in everything you said. Could you recommend some good (beginner friendly) online reading on: VU meters + clipping + peak vs filtered signal?

[stschoen]

Wikipedia has pretty good article

VU meter

[ambian]

Thanks. This video also has a decent high level summary: https://www.youtube.com/watch?v=Y76tCk9HeDw

I think I understand everything except for…

@stschoen

The *5 refers to the peak signal.

What do you mean by this? Are you saying that the meters are calibrated so that:

0 (on the meters) == 1 (audulus value) - 5dB (leaving 5dB as headroom)

and

the clip indicator responds to peaks while the meters show filtered signal

?

[stschoen]

It should have read +5 not *5. The meters are currently calibrated so that clipping occurs when the signal peaks reach +5. Because the meters are displaying a smoothed signal, the level displayed on the meter when clipping occurs depends on the signal. For typical audio signals the clip indicator lights when the meters are at somewhere between 0 and +3 VU.
The clip indicator lights when the peaks of input signal exceed 0.995. Audulus hard clips audio out at 1 so I figured that was a reasonable number. The meters are meant to be a quick indicator of signal level for audio signals rather than a precise measurement tool. Audulus only clips the signal at final output (speaker or ADC node) so higher levels earlier in the signal chain are possible. Some modules also clip the signal, usually with the intent of introducing some type of distortion.
Depending on the standard used, traditional VUs are calibrated for clipping at either 0 dB or +6 dB with a typical audio signal. Analog audio equipment clips the signal less abruptly so having the occasional peak that clips is not as audible as the hard clip caused by digital signal processing. I originally chose 0 dB as the clipping point but later relabeled it to +5. Since we’re not dealing with a standard analog audio signal it’s a bit arbitrary.