Tags: posts polarity-music Bitwig Tutorial Sampling

Bitwig Grid Sampler Analysis Outputs for Pitch and Onsets

Tutorial | Jul 03, 2026

Learn how to use Bitwig Studio 6.1 Grid Sampler analysis out ports to extract pitch, onsets, and envelope for sample-based sequencing and pitch correction.

Quick Answer

Short Overview

Bitwig Studio 6.1 Grid Sampler analysis out ports add direct access to onset, pitch, and envelope data from loaded audio, turning a sample into a control source for modular patching. This matters in practice because detected notes and transients can drive oscillators, envelopes, triggers, and timing logic inside The Grid without manual MIDI programming.

With pitch and onset extraction, audio can be quantized to scales, aligned to the beat grid, or reused to repitch another sampler layer for generative sequencing, glitch textures, and sample-aware modulation. The feature is especially useful for experimental sound design, even though current pitch and transient detection may still require refinement for precise tracking.

Key Takeaways

Bitwig Studio 6.1: Using Sampler Analysis Outputs Inside The Grid

Bitwig Studio 6.1 adds a useful new capability to Sampler inside The Grid: analysis output ports for onsets, pitch, and envelope. Once enabled, Sampler can expose information extracted from its loaded audio and make that data available as modulation and control signals inside a Grid patch.

This turns a sample into more than a playback source. It becomes a generator of note, rhythm, and amplitude data that can drive oscillators, envelopes, quantizers, and other modules.

What It Does

In the Sampler inspector, there is a new checkbox: Enable analysis out ports.

When enabled, Sampler exposes three additional outputs:

These outputs can be patched anywhere in The Grid, making it possible to:

How It Works

Pitch output becomes useful once Sampler has analyzed the source audio. In practice, this means enabling dynamic pitch detection at least once so the sample is analyzed.

After analysis, the pitch output can be sent into modules such as:

For example, a sample containing separate pitched hits can produce a changing pitch control signal as playback moves through each hit.

The onset output behaves like a transient detector. Each detected attack creates a control spike that can be used to trigger envelopes, clocks, or logic.

The envelope output follows the amplitude shape of the sample and can be used directly as a modulation source, for example to control oscillator level.

Basic Workflow

A typical patch looks like this:

  1. Load a sample into Sampler in The Grid.
  2. Enable analysis out ports in the inspector.
  3. Enable dynamic pitch detection so Sampler analyzes the file.
  4. Patch:
  1. Use the outputs to drive synthesis or transform the original material.

A simple example is to mute Sampler’s audio output, then:

The result is a crude resynthesis layer that follows the original sample’s melodic and dynamic behavior.

Using Thresholds On Onset Detection

Onset detection can generate too many triggers, especially with complex material. One way to clean this up is to insert a threshold stage:

This filters out weaker or unwanted transient detections and produces more usable rhythmic events.

Quantizing Rhythm To The Beat Grid

Because onset detection is based on audio events, the resulting triggers are not necessarily aligned to the project grid. To force them into time:

This converts irregular transient timing into musically synchronized triggers.

In effect, the sample provides the source rhythm, while the patch snaps that rhythm to the track tempo.

Quantizing Pitch To A Scale

Detected pitch can also be constrained musically:

This is especially useful when the source audio contains melodic content that is close to, but not exactly in, the target scale.

Correcting Sample Playback With Detected Pitch

A more advanced technique is to use two Samplers:

  1. The first Sampler analyzes the audio and outputs pitch.
  2. That pitch is sent through a Scale module or other pitch-correction stage.
  3. The corrected pitch is then used to drive a second Sampler playing the same material.

This effectively repitches playback based on the analyzed notes, allowing rough pitch correction of sampled hits to a target scale.

The method depends heavily on accurate pitch tracking. When detection is unstable, the correction will also be unstable.

Working With Audio Output Alongside Analysis

The analysis outputs do not replace normal audio playback. Sampler audio can still be used in parallel.

One useful technique is to:

This adds overtones to an otherwise pure sine and produces more complex timbres tied to the original sample.

Sliced Mode And Onset Markers

Sampler’s Sliced mode can display automatically detected onsets. Switching slice mode to Onsets shows where Bitwig believes transient events occur.

Manual editing of these markers is possible visually, but there is an important limitation: those edits do not appear to persist in a way that affects the analysis outputs when switching back to other modes. Returning to onset slicing restores the automatic detection rather than preserving manual corrections for broader analysis use.

That means onset output currently reflects the automatic detector rather than a user-corrected onset map.

Spectral And Looping Modes

Analysis outputs are also useful beyond one-shot sample playback.

If Sampler is set to loop, the pitch and onset outputs become a repeating control sequence derived from the looped audio.

In Spectral mode, playback can be slowed, frozen, or scanned through the sample. This creates another way to extract pitch and gating behavior from a transformed playback engine rather than from simple linear playback.

For example:

This can produce semi-random melodic structures based on frozen fragments of the sample.

Practical Uses

These analysis outputs open several strong patching workflows inside The Grid:

Limits

There are some clear limitations in the current implementation.

Pitch Detection Accuracy

Pitch tracking is usable, but not always precise. Even relatively simple monophonic material can produce unstable or incorrect note detection. Noisier sounds, strong transients, overtones, DC offset, and other spectral complications make this harder.

As a result, the feature currently excels more at experimentation, glitch processing, and approximate control than at surgical pitch extraction.

Onset Detection Control

Onset detection is useful, but there is no clear way to manually correct the analysis and have those corrections persist as analysis data. A workflow for editing and preserving onset decisions would make the feature significantly more reliable.

Recording Into Sampler

A notable workflow gap remains: there is still no direct way to record straight into Sampler inside The Grid. The current process involves recording elsewhere, locating the file, and dragging it back into Sampler.

Direct recording into Sampler would make this analysis-driven workflow much faster and more self-contained, especially for live experimentation inside Grid patches.

Why This Matters

The key advantage is modular access. Outside The Grid, Sampler does not expose this kind of note, onset, and envelope data in the same flexible way. Inside The Grid, those outputs become patchable signals, which means any sample can function as:

That makes Sampler in Bitwig Studio 6.1 not just a playback device, but a bridge between recorded audio and modular control.

Conclusion

Bitwig Studio 6.1’s new analysis outputs for Sampler in The Grid are a powerful addition. They allow audio material to be interpreted as control data and reused for synthesis, timing, pitch mapping, and generative patching.

The feature is already musically useful, especially for experimental workflows. Better pitch accuracy, editable persistent onset correction, and direct recording into Sampler would make it much stronger. Even in its current state, it adds a compelling new way to turn audio into structure inside The Grid.

Transcript

This is the transcript of the video. The text was generated automatically and may contain small mistakes. The timestamps jump to the matching part of the video.

Click to expand transcript

[00:00:00] Yo guys, welcome back. I want to show you something in Bitwig Studio 6.1 here. The new sampler inside of the grid has some
[00:00:07] Yeah, additional features. So if you have the sampler in the grid and you select it in the inspector on the left side here
[00:00:14] We have a new checkbox called enable analysis out ports if you enable this
[00:00:19] You can see on the sampler itself
[00:00:21] We have now three new output ports here the onsets the pitch and the envelope of the audio signal here
[00:00:29] That you have inside of the sampler. So this means if you
[00:00:33] yeah
[00:00:35] Analyze here the sample by using dynamic pitch detection or
[00:00:40] Even if you haven't
[00:00:43] Disabled you get here a pitch signal so we can
[00:00:47] Check this here. I'll maybe use this one here
[00:00:50] By just using a quantizer and if I play the sample
[00:00:56] You can see we get here the pitch signals from the sound
[00:01:00] Right
[00:01:07] Which means we can put just just any sample in that
[00:01:11] Detect the pitch or the onsets and then work with that information
[00:01:17] inside of the grid and create I don't know patches or
[00:01:22] Generative music or whatever you want to do with that
[00:01:25] So I want to show you here something
[00:01:28] That I did recently so I want to record here another
[00:01:33] Sampler patch or just some music into here the master recording thing
[00:01:38] So maybe I just hit record and play something
[00:01:54] So feel like that so we can go here to the
[00:01:57] Thing and put this into the grid
[00:02:02] in here and
[00:02:05] We have now multiple single
[00:02:08] Musical hits here. So what we can do now is we can check I think at the moment
[00:02:15] Disable this at the moment no pitch is detected. Maybe let's try
[00:02:24] Yeah, nothing really happens here you need to enable at least once this dynamic pitch detection here
[00:02:31] So it's analyzing the audio
[00:02:33] So you can see we have different pitches here
[00:02:36] as a single line and now when we play
[00:02:42] And see you get different pitches here
[00:02:54] You can see sometimes the pitch detection is not
[00:02:57] Highly precise. I hope they work on that because this material here is not very complex. I would say it's a pretty clean
[00:03:08] note I'm playing there with a very basic sound
[00:03:13] So the more noisier the sound gets the more problems the pitch detection has
[00:03:20] I hope they work a bit on that at the moment. It's it's really
[00:03:24] Okay, to play with it to get some glitched sounds. But if you want to have precise
[00:03:30] Pitch tracking, it's not that but I get it. It's a very hard thing to do in
[00:03:36] In processing that you detect a pitch because they have so many things that
[00:03:42] throws the
[00:03:45] Algorithm of right some overtones and transients and DC offsets and stuff like this
[00:03:51] So yeah, it's a very hot problem. But maybe we get some better versions over the beta period. I hope so at least
[00:03:58] Then we have also here some onsets and you can't see the onsets also problem. I have here is when you disable the
[00:04:06] Dynamic pitch tracking you still now receive the pitches of this sound if this is on
[00:04:14] the sampler basically tries to
[00:04:16] Offset here the pitch so it if I play for instance if I play a C it tries to
[00:04:23] Pitch all these individual hits here. So it matches to C
[00:04:28] Right so it tries to play everything at C even though some of these
[00:04:41] Notes here on a different note if I disable this and I play it it plays all of these hits in the original pitch
[00:04:48] And you can see we receive down here basically the note of each hit on the pitch quantizer
[00:05:01] We also get on set so maybe you see an oscilloscope
[00:05:07] and
[00:05:10] Maybe move this up with this here makes longer and then switch this to slow. So let's see all this looks like
[00:05:18] Right so you can see here for each onset we get some kind of trigger not everything is
[00:05:31] 100% at one
[00:05:35] So you can't see also your the onsets, but what you can do is you can switch this to sliced mode and then switch this here do
[00:05:43] Onsets and then you can see where I detect some onsets. I believe you can
[00:05:51] Remove your some of these onsets by double-clicking
[00:05:56] But the problem is
[00:05:59] You can hit you can't hit play anymore because it's now a sliced sample, right?
[00:06:04] You have to hit here these notes to actually play
[00:06:08] What's in there?
[00:06:10] So my initial thinking was oh, I can disable here actually the onsets and just take what I want
[00:06:17] For instance this one here, right and then switch back to single mode and then it hit play
[00:06:27] No, it doesn't take these manually
[00:06:33] corrected onsets if you go back to sliced it goes back to divisions here and when you go to onsets it just takes the
[00:06:41] automatic
[00:06:43] Yeah, detection of these onsets. So all your manual
[00:06:46] Changes are gone
[00:06:50] So maybe they need to find a way here where we can influence the onsets and can influence how the pitch
[00:06:56] detection works, right so manually
[00:07:01] correcting some of these things and then
[00:07:03] Persistence in a way so we can then take the output of that
[00:07:08] Because this automatic detection can always go wrong or it's mostly wrong
[00:07:14] So that would it would be nice to have a way to correct us actually in a certain manual way
[00:07:20] Anyway, so we get here the
[00:07:23] The onsets and the pitch from that so we can for instance for just for simple
[00:07:32] Simple example here you can use an os os oscillator
[00:07:36] Or maybe take a sign sign oscillator here for now so we can take the pitch
[00:07:44] To an sign oscillator disable this and then we have also here a trigger so we can use an ad
[00:07:51] You can use this trigger
[00:07:54] You can use here
[00:07:57] maybe use this output and
[00:08:01] Let's see all the sounds when I hit play here on the sampler. There's no audio output. We only use the pitch and
[00:08:07] envelope
[00:08:11] So tries to play
[00:08:18] So kind of rocks
[00:08:25] So what we can do is we can
[00:08:28] Try to make some kind of threshold here because you can see we have here a lot of onsets we can filter out some of these things and
[00:08:37] Take here logic. Maybe is bigger
[00:08:42] Right, so is bigger than a certain value say this one here and we use a constant
[00:08:55] Something like this make it a bit smaller put this in here. Maybe we use here different color something like this
[00:09:01] And hit play
[00:09:04] So only
[00:09:07] These yellow lines here that are over this line are a trigger
[00:09:12] So this does much better. Maybe make it a bit higher something like that
[00:09:17] And we can also take for instance instead of using here the ad and trigger that
[00:09:25] We can take multiply
[00:09:28] Connectors and use the envelope out here
[00:09:34] Multiplied us. So now it takes the envelope from this
[00:09:38] waveform here and
[00:09:41] Uses that to change the volume of the sinus so that also let's see all the sounds
[00:09:55] Right. So this is also possible
[00:09:57] We can also try and use the audio output of the sampler
[00:10:02] And just face modulate here the sign oscillator to get some overtones because the sign at the moment is just very pure
[00:10:09] I don't know overtones at all. So let's try this one
[00:10:25] So right we have here some
[00:10:30] Very weird pitch detection here would be nice to actually straighten this out in a manual way
[00:10:36] But it's no way at the moment. I'm I'm 100% sure they do something like this in the beta period before they release this to
[00:10:43] stable
[00:10:46] So we can change the take detection here or make some
[00:10:51] Corrections. So anyway, so yeah, we can take your the audio output and then face modulate the sign to get some overtones
[00:10:58] We use the pitch here. We can also
[00:11:00] Of course change the
[00:11:03] Scale to something else, right?
[00:11:06] So instead of using the notes coming from here and
[00:11:09] Quantizing this to a real note. We can quantize this to a scale. So it's now in a certain scale
[00:11:22] Or we can use what's the name the scale module here instead of this and then say we want to have
[00:11:29] C major. Yeah, let's just leave C major
[00:11:33] You can also see that here some of these triggers are
[00:11:48] Completely random. So they are not on a beat grid. So we can also do something like this. We can use a clock quantizer
[00:11:56] Clock quantizer we quantize everything to I don't know a
[00:12:01] 16 note grid
[00:12:05] go to 16 notes or 16 triggers per bar and
[00:12:09] We go here with the
[00:12:15] With the onset output into this. So now everything is quantized to a beat grid
[00:12:19] And then we have to use here, of course the AD
[00:12:24] Can I replace this here? No AD
[00:12:27] Back to this and
[00:12:30] We get here. Oh, we don't use this one use this one. That's better
[00:12:36] So we have a threshold everything that goes above the threshold gets clock quantized to 16 notes
[00:12:44] And then we can use this as a trigger here
[00:12:47] No triggers too high. Okay, let's try again
[00:13:01] No triggers, okay
[00:13:09] But you get the idea right so we can quantize here
[00:13:13] What's coming out of this?
[00:13:15] Samples so it's not randomly triggering all over the place so we can
[00:13:19] Synchronize it actually to the BPM of a track. We can also
[00:13:23] Quantize this here to a scale of the track. So we can really take an audio sample
[00:13:28] analyze it and then use whatever is in there as a source for notes and
[00:13:35] rhythm, which is pretty cool
[00:13:38] in my opinion and
[00:13:41] Yeah, why not do it?
[00:13:43] Okay, so we have this maybe remove here this
[00:13:46] You go to trigger and then what else can we do? We can also
[00:13:53] Correct actually the output of the sample
[00:13:56] Because you can see here
[00:13:59] It's not playing in the current scale. We have C major
[00:14:03] So we can try to correct this by just using
[00:14:09] The sampler twice
[00:14:11] So instead of using here this audio output we use the pitch we correct the pitch and then we go back into the sampler here and
[00:14:21] then play
[00:14:25] Each note at the different thing. So here we
[00:14:29] Use pitch detection we switch this off
[00:14:34] And then it uses from every single hit the pitch and then place this pitch
[00:14:40] At the correct pitch because we use the scaler here. Okay, so this sounds then maybe
[00:14:46] We remove this here for a while
[00:14:49] Let's move this over here
[00:14:53] And what do we use here we use an
[00:14:58] Multiplier
[00:15:04] And we use than here this one that sets you the songs
[00:15:08] So instead of instead of it's a mixer here maybe
[00:15:34] So
[00:15:36] You can hear it's not the same pitch it's correcting basically
[00:15:43] The pitch to the scale
[00:15:47] So again, what we did here is analyzing the audio signal
[00:15:52] extracting the pitch
[00:15:54] Then we use the pitch from each of these hits go into a scaler or by scale module
[00:16:01] that corrects every pitch output here and
[00:16:04] Yeah brings it up to the next note on the scale
[00:16:09] And then we use this note coming out of this and playing back here the sample again
[00:16:15] But with a different pitch because we are on a different scale so you can pitch correct more or less
[00:16:21] Samples this way
[00:16:24] When the pitch is correctly detected at least that's the that's the yeah
[00:16:31] Kind of what you need to do and what what needs to be correct
[00:16:34] I also try to use this in a way where you can
[00:16:38] Go back into the sampler which doesn't work
[00:16:41] You can try to use here a long delay, but you probably get very weird results with this
[00:16:47] Maybe you can delay it here by three sixteen notes
[00:16:50] Or two sixteen notes. So the next note
[00:16:54] Or the note is played back by the pitch of the next note or something like this
[00:17:01] This could be also interesting
[00:17:03] We go in there. Let's play it back
[00:17:07] So right it takes basically two sixteen notes for the pitch to adapt and then you have this weird
[00:17:21] Clitch in there, but you can also go down at you the lowest
[00:17:31] Yeah, it gives you also some kind of result so this should be also possible
[00:17:37] But you probably run into a small glitches because the delay time here is not super low in my case
[00:17:45] It's I don't know what it is
[00:17:47] audio
[00:17:49] Five milliseconds. This is the lowest
[00:17:52] Yeah, it's it's probably okay for some results
[00:17:56] But you can use this to pitch correct some of the things to a scale if you have a sample
[00:18:01] That's not in scale you can try use the pitch detection for that and then just correct it, but still I would be
[00:18:07] very happy if Bitwig actually allowed to
[00:18:12] Record directly into the sampler, I mean just imagine you play or you generate some sounds inside of the
[00:18:21] Grid here and then you go directly into the sampler record some stuff
[00:18:26] analyze the pitch
[00:18:29] Make the transient detection on and you know just work with the audio
[00:18:34] I don't know why it's so complicated. Maybe it's a
[00:18:37] It's not that easy to do in Bitwig. I have no idea, but just let me record with a sampler
[00:18:44] It would be so nice to have and I hope Bitwig
[00:18:48] Dostas at some point in the future at least
[00:18:51] So yeah, so these are some things you can do inside of the grid with a sampler you can't do
[00:18:59] Outside of the grid just with the native device because there is no
[00:19:03] No modularity basically there or there's no note output no
[00:19:08] Onset output and also here the envelope output very nice
[00:19:14] Okay, so before I end the video
[00:19:17] I want to give you also an example here what you can do with the other modes because at the moment
[00:19:23] We only use three pitch and spectral but the interesting part is also that you not only can play the sample
[00:19:30] And then extract the pitch and the onsets or the gates
[00:19:33] We can also just loop this. So let's say I put this here into a loop mode
[00:19:40] One way looping and then hit play on the keyboard
[00:19:44] We get continuously the pitch and
[00:19:47] Maybe I delete this here
[00:19:53] Right we get more or less. It's a sequence now and load sequence with gates and pitches
[00:20:00] But we can also go to maybe the spectral mode here and
[00:20:08] Then we say there's a playback speed. Maybe we bring down the playback speed here
[00:20:14] And then it stays in place right so it's not triggering all the time so we can increase the speed
[00:20:23] Oh
[00:20:25] Maybe switch this to fragments density
[00:20:48] Well, it's actually only giving away here the pitch and the gate from the playhead not all of these small little
[00:20:56] Grainy things would be nice to have
[00:20:59] Could be very chaotic
[00:21:01] So if we remove here the scale we can also mix in the output
[00:21:11] Let's use a mixer here
[00:21:19] Oh
[00:21:21] But we can use the audio output here too
[00:21:34] We increase the density
[00:21:48] Oh
[00:21:50] Yeah, also interesting when we
[00:22:09] Bring this down here and say we want to freeze actually in the playhead
[00:22:18] So now you can choose this one and go through the sample manually, right?
[00:22:23] so with this
[00:22:25] I need to activate actually here one time
[00:22:30] Yeah, with this we can just use a trigger here
[00:22:49] Four triggers on each bar
[00:22:52] That's okay, then use the dice get some random values
[00:22:59] modulator output here
[00:23:02] And go maybe to the beginning
[00:23:05] Then use a leg
[00:23:18] Oh
[00:23:20] Spectral let's see all the sounds only have to activate it
[00:23:48] Oh
[00:23:50] There's too much or maybe does too too fast yeah, it's called 85
[00:24:18] Oh
[00:24:20] Quantize to global scale, maybe it was here loved e b minor
[00:24:48] Oh
[00:24:50] Let's do another one
[00:24:59] And then you see a band
[00:25:18] And we use maybe this one here 16
[00:25:47] Something like that
[00:25:49] And then we probably also need to quantize this or we quantize already and maybe two eight
[00:26:18] We can put this here just to repitch
[00:26:22] Remove the modulation
[00:26:30] And then go to zero here
[00:26:35] And then try to loop it we loop this already use
[00:26:46] Scale
[00:26:48] Something like that
[00:27:07] This one takes takes the pitch here pitches the sampler and then this sampler
[00:27:15] Pitches here the sign and everything is quantized to the scale
[00:27:19] We play this backwards
[00:27:45] Yeah, why not?
[00:27:47] so yeah, this is just a very stupid idea to
[00:27:51] Use samples as a source inside of the grid with the new sampler
[00:27:56] It's very nice very fun to do
[00:27:59] I hope they improve the pitch detection or the pitch tracking and onset detection
[00:28:07] If not, it's still fun for some glitchy stuff
[00:28:11] but with the better pitch detection, of course, it would be
[00:28:16] Yet more real use cases for I don't know vocals and lead sounds or any samples
[00:28:25] Would be nice to have also recording inside of the sampler would be very nice
[00:28:32] Inside of the grid where you can just make some experiments and directly record into the sampler without using a you know
[00:28:39] Global thing and then searching for the right folder and dragging it back in and so on
[00:28:44] So why not just let me record inside of the sampler?
[00:28:48] I mean it just it looks actually like that it looks like oh where where is the input? Where's the audio input?
[00:28:57] It looks like it it has to have some kind of recording capability
[00:29:02] Anyway, so that's that I want to make a video about this and it's very fun
[00:29:07] Try it out. Let me know what you think in the comments down below and
[00:29:10] Yeah, see you in the next video. Bye