Visualizing Audio Filters
Tutorial | Jan 08, 2024
In the video, I demonstrate how to visualize the behavior of different filters using a spectrum analyzer in the Bitwig Grid. By modulating a sine oscillator and applying various filters such as the comp filter and fizz filter, viewers can observe the changes in the frequency domain. This visual representation provides a helpful way to understand the effects of these filters even without deep knowledge of their underlying theory.
You can watch the Video on Youtube - support me on Patreon
This video tutorial demonstrates how to visualize and understand the behavior of different filters in Bitwig using a Poly-Grid and a spectrum analyzer. Here's a summary:
-
Setting Up the Test Environment:
- Utilize a Poly-Grid and a spectrum analyzer with specific settings (large resolution, slow speed, 3 dB/octave tilt for pink noise).
- Create a sine oscillator modulated by an LFO to observe filter effects.
-
Using Comp Filter:
- Apply a comb filter, adjusting volume and feedback.
- Observe the comb filter's effect in the frequency spectrum, particularly its ability to create overtones and harmonics, useful for physical modeling.
-
Exploring Different Filters:
- Experiment with various filters like 'Fizz' and observe changes in frequency domain.
- Adjust parameters to see real-time changes in the spectrum.
-
Visualizing Filter Effects:
- Compare visual representations of filters in Bitwig with other software like Serum.
- Note the use of reverb filters and convolution for different sound modeling.
-
Understanding Filter Characteristics:
- Analyze different filters, such as 'Rasp', to understand their unique features and frequency responses.
- Observe how adjusting filter settings alters the sound spectrum.
-
Investigating Further:
- Examine different modes and types of filters (e.g., bandpass, low pass) and their impact on sound.
- Notice how changing filter poles affects steepness and sound character.
-
Conclusion:
- Emphasizes the importance of visualizing filters for a better understanding of their impact on sound.
- Encourages experimentation with various settings to explore different sound textures.
This tutorial is useful for visual learners and those interested in the technical aspects of sound design in Bitwig.
Questions & Answers #
Maybe you dont watch the video, here are some important takeaways:
How can the science behind the filters in the grid be explained? #
I am not an expert in filter design, so I cannot explain the theory behind each filter. However, I can show you how to visualize what the filters do using a spectrum analyzer and a sine oscillator modulated with an LFO. This can give you an idea of how the filters affect the frequency spectrum.
Is there a video that explains the theory behind the filters in the grid? #
I have not explained the theory behind the filters in a video before, and I am uncertain if I will do so in the future. However, I have provided a visual demonstration in this video to help you understand the behavior of different filters in the frequency domain.
What is the purpose of visualizing the filters in the grid? #
While the primary focus when using the filters should be on how they sound, visualizing them can also be helpful for understanding their characteristics and behavior. Visuals can provide additional information that complements the auditory experience, allowing you to see how the filter affects the frequency spectrum.
How can the visualization of the filters be done in other software? #
In other software such as Serum, you can often find filter graphs that show how the filters look in the frequency domain. By switching between filters and observing the visual representation, you can gain insight into how each filter type works and how it affects the frequency spectrum.
Transcription #
This is what im talking about in this video. The text is transcribed by AI, so it might not be perfect. If you find any mistakes, please let me know.
You can also click on the timestamps to jump to the right part of the video, which should be helpful.
[00:00:00] The filters in the grid make different sounds.
[00:00:02] We do not know the theory behind each of them.
[00:00:05] Did you explain it in a video before or will soon?
[00:00:08] So I can't explain the science behind the filters
[00:00:12] or how they came to be.
[00:00:14] I'm not a filter design expert in these regards,
[00:00:19] but what I can do is I show you some
[00:00:22] a way to visualize actually what these filters do.
[00:00:27] So what I do then is I use a Poly-Grid here.
[00:00:30] And after the Poly-Grid, we just use a spectrum analyzer,
[00:00:36] extend here the view, resolution goes to large,
[00:00:39] maybe speed goes to slow, tilt goes to a 3 dB per octave,
[00:00:43] which is pink noise.
[00:00:45] And then in here, we just use a sine oscillator
[00:00:47] and you modulate the sine oscillator with an LFO.
[00:00:52] Going to the pitch input here, disable this
[00:00:56] and then go maybe for a saw, switch this to Hertz,
[00:01:01] let's say 10 Hertz or 12 Hertz, something like this.
[00:01:04] We modulate basically the pitch here.
[00:01:07] We go down here in volume and use an output.
[00:01:11] Sounds like this.
[00:01:13] Right, it's just a sine modulated.
[00:01:17] Then we use your bipolar.
[00:01:19] So we modulate over the full range.
[00:01:21] You can see it's more like a straight line
[00:01:23] that we have now here.
[00:01:25] And then we can implement here some filters.
[00:01:27] Let's say a comp filter,
[00:01:29] but in here, go down with the volume, increase the feedback.
[00:01:36] And you can see how the comp filter here
[00:01:38] basically plays out in the frequency spectrum.
[00:01:41] When we go to 1K, 1 kilohertz,
[00:01:44] you can see the lowest frequency is basically exactly
[00:01:48] on 1K, which is the resonant or the fundamental.
[00:01:52] And then we have here all these overtones
[00:01:54] or harmonic series we create with the comp filter,
[00:01:57] which makes it an interesting filter
[00:02:00] for physical modeling, right?
[00:02:02] Because you can just create a lot of overtones with them.
[00:02:05] With that, one of the Hertz here,
[00:02:09] yeah, we have the lowest frequency here.
[00:02:13] We can lower the frequency or the resonance here
[00:02:15] with the feedback.
[00:02:17] We can dial in here additional low pass with this.
[00:02:22] So if you're not happy with all the overtones here,
[00:02:25] you can just remove some of the overtones.
[00:02:27] It's basically like a low pass after the filter,
[00:02:32] a second low pass.
[00:02:33] So yeah, this is the comp filter,
[00:02:36] how it looks like inside of the frequency domain.
[00:02:39] You can also change this here
[00:02:41] to maybe some of the new filters, let's say Fizz.
[00:02:44] You can see Fizz looks a bit different.
[00:02:49] You can change the frequency.
[00:02:50] Let's go to 1K.
[00:02:52] And 1K is here, the highest,
[00:02:55] one of the highest resonances.
[00:02:57] We have here resonance here.
[00:02:59] And when you change it to this knob,
[00:03:01] you can see how the filter kind of behaves
[00:03:04] in the frequency domain, what it actually does.
[00:03:07] And you get more like an idea how the filter sounds like.
[00:03:15] I guess they aimed here inside of the grid
[00:03:18] with all of these devices,
[00:03:19] more for like that you should hear what's going on.
[00:03:22] So you have to trust your ears more
[00:03:25] instead of focusing on these visuals here.
[00:03:29] But sometimes it's nice actually
[00:03:30] to see how it looks like.
[00:03:32] We have something like this in serum here.
[00:03:36] Where you have like this filter graph here on the right side.
[00:03:39] You can switch between filters
[00:03:40] and you can see, this is the comp filter here.
[00:03:44] You can see how they look like in the frequency domain.
[00:03:47] I guess a lot of people actually want that.
[00:03:50] Let's see all parts here.
[00:03:51] So all parts looks like there's a reverb filter here.
[00:03:56] It looks more like an impulse response they used
[00:04:03] for the filter types here, I guess.
[00:04:05] And then you change basically the tuning
[00:04:07] of the impulse response.
[00:04:11] You can do something like this in Bitwig also here
[00:04:13] with convolution, right?
[00:04:17] And then you can change basically the frequency
[00:04:19] with the tuning knob.
[00:04:21] But the problem here is it's not updated in hertz.
[00:04:26] It's more like, they watch you move this
[00:04:30] and when you stop moving
[00:04:31] then it actually updates the filter frequency in here.
[00:04:34] So you can't really use this to filter anything.
[00:04:38] Maybe they change this in the future.
[00:04:39] So you could use your tuning knob
[00:04:41] actually as a cutoff knob, right?
[00:04:44] Something like this.
[00:04:45] So yeah, this is a reverb filter here.
[00:04:49] You can see how it looks like in the frequency domain.
[00:04:52] And I guess a lot of people actually want this also
[00:04:56] inside of the grid here.
[00:04:57] So you can see what some filters are actually doing.
[00:05:00] So this is FIS.
[00:05:01] You can see how it looks like.
[00:05:03] You can try it out for yourself.
[00:05:05] Always switch it to rasp maybe.
[00:05:08] If a lot of overtones here,
[00:05:10] let's go to 1K.
[00:05:12] It's 1K is this here, the lowest.
[00:05:15] Let's increase the resonance.
[00:05:18] The lowest ripple here.
[00:05:19] So this is a low pass here.
[00:05:24] And then we can use this knob
[00:05:26] and it introduces a lot of different frequencies
[00:05:30] here around 1K.
[00:05:38] Interesting that they named this rasp.
[00:05:40] It actually looks more like ripple.
[00:05:42] So then we can switch this here, the brightness mode.
[00:05:47] Oh, interesting.
[00:05:51] So this one here keeps the resonance here on 1K
[00:05:56] and it changes all the other resonances here.
[00:05:59] And this one introduces some resonance below 1K.
[00:06:03] It's more like a mirror.
[00:06:07] Yeah, it's this one stays in the middle at 1K here.
[00:06:10] And then you spread out to the lower frequencies
[00:06:15] and to the higher frequencies.
[00:06:17] Interesting.
[00:06:18] Then we have this one also kind of the same feel.
[00:06:23] It's also like a mirror.
[00:06:25] Resonance, band pass.
[00:06:29] Yeah.
[00:06:34] Yeah, so this is how you can visualize basically
[00:06:37] what these filter are actually doing
[00:06:39] or how I do it sometimes.
[00:06:41] Let's go to ripple.
[00:06:45] Feedback, feed forward.
[00:06:47] Let's increase here the resonance.
[00:06:50] 1K, so 1K is here, right?
[00:06:54] But the resonance is here and that's interesting.
[00:06:56] Yes, you have to move this knob here all the way up.
[00:06:59] Yeah.
[00:07:00] Yeah, interesting.
[00:07:08] Then we have different modes here
[00:07:11] with different resonant frequencies, different points.
[00:07:15] Yeah, this is how we can visualize it.
[00:07:26] You can also try it out here with all of the others.
[00:07:29] Let's say XP, the peak one just introduces here
[00:07:34] a frequency push at, let's say, 1K
[00:07:39] or maybe a notch.
[00:07:45] So you cut frequencies here.
[00:07:50] Or let's go to selling key and use the BPA tier.
[00:07:59] This I think the steepest,
[00:08:00] BPA, this is the steepest band pass
[00:08:04] you can get in between here.
[00:08:05] Eight pole band pass filter.
[00:08:08] If you go to BP six, you can see
[00:08:12] how it's not so steep anymore.
[00:08:15] That's a low pass.
[00:08:22] LP eight, LP six, four, three,
[00:08:29] two and so on.
[00:08:30] So you can see the steepness here changes
[00:08:32] with the change of these poles.
[00:08:35] It's even one pole filter.
[00:08:37] Yeah, this is how you can visualize it in the spectrum.
[00:08:44] I wanted to show you this
[00:08:45] because I think a lot of people watching my channel
[00:08:48] are actually visualized.
[00:08:53] They want to see the visuals, how it looks like
[00:08:55] instead of how it sounds like.
[00:08:57] So this is how we can do it.
[00:08:59] Okay, I hope I could bring some knowledge out there.
[00:09:04] Thanks for watching.
[00:09:05] Leave a like if you liked the video,
[00:09:06] subscribe to the channel
[00:09:07] and I'll see you in the next video.
[00:09:09] Bye.