---
title: "AI podcast intro music: pick 92 BPM, then cut loops at 8 bars"
site: "getrhythmm.com"
kind: "article"
author: "Evelyn Porter"
date: "November 19, 2026"
canonical: "https://getrhythmm.com/blog/ai-podcast-intro-music-pick-92-bpm-then-cut-loops-at-8-bars.php"
markdown: "https://getrhythmm.com/blog/ai-podcast-intro-music-pick-92-bpm-then-cut-loops-at-8-bars.php/index.md"
description: "Cut AI podcast intro music at 92 BPM to an 8-bar loop ending at 20.87 seconds, with 3-stem exports for ducking, re-cuts, and future re-versioning."
---

# AI podcast intro music: pick 92 BPM, then cut loops at 8 bars

Evelyn Porter · November 19, 2026

> Cut AI podcast intro music at 92 BPM to an 8-bar loop ending at 20.87 seconds, with 3-stem exports for ducking, re-cuts, and future re-versioning.

For an AI podcast intro at 92 BPM in 4/4, pick an 8-bar loop and cut it at 20.87 seconds — the end of bar 8, i.e. the bar-9 downbeat on your DAW ruler.

Change that answer only if you plan at least 2 downstream edits, such as ducking under voice and re-cutting or re-versioning, which calls for a 3-stem export.

![For an AI podcast intro at 92 BPM — AI podcast intro music](https://static.mm-ais.com/article-images-ai/ai-podcast-intro-music-pick-92-bpm-then-ai-ed9cad42.jpg)

## Why 92 BPM and 8 Bars Fit Intros

Eight bars of 4/4 at 92 BPM is 32 beats, and 32 × (60 ÷ 92) = 20.8696 seconds, which rounds to 20.87 seconds. That is the exact bed length before you add any tail or fade. Verify it against your DAW's tempo ruler rather than trusting your ear: park the playhead on the bar-1 downbeat, then read the ruler at the bar-9 downbeat (the end of bar 8). If your session shows anything other than 20.87s at that point, the project tempo or time signature is not what you think it is, and every later cut will drift.

The reason this math matters is that tempo-locked AI generators quantize their loop grid to the BPM they generated. Tools such as SOUNDRAW, TopMediai's podcast intro mode, and FreeMusic AI ship material on that grid, so the bar line is the structural seam the generator planned for. Cut on the bar-8 downbeat and the next pass lines up with the same grid; cut mid-bar and you split a quantized phrase, which is where an audible seam shows up. The check is simple: zoom to sample level at your cut point and confirm the waveform lands on the ruler's bar boundary, not between beats.

Ninety-two is a useful intro tempo because it sits in the 85–100 BPM pocket where a short bed reads as deliberate without sounding rushed. At that pace, eight bars gives you room for a hook, a lift, and a resolve inside the 20.87s window, which is long enough to establish a show's tone and short enough that the host can come in before the listener drifts. If your generator offers a tempo field, set it to 92 before you render; trying to stretch a 96 BPM loop down after the fact changes the grid you were relying on and puts the bar-8 downbeat back in question.

Treat the 20.87s figure as the bed, not the file. A fade or reverb tail extends the audible content past the downbeat, so decide up front whether the tail lives inside the eight bars or after them. If it lives after, your ruler check still targets 20.87s for the loop point, and the extra decay is handled in the mix rather than counted as part of the loop. Write that decision into your session notes so a later pass does not mistake the tail for a ninth bar.

One practical guardrail: confirm the exported file's sample rate and length in your DAW's clip inspector before you place it on the timeline. A 20.87s clip that imports as 20.91s has been resampled or padded somewhere upstream, and the bar-8 downbeat is no longer at 20.87s. Re-import against the tempo ruler and re-check the boundary. That single verification keeps the duration math and the generator's grid in agreement, which is the whole point of choosing 92 BPM and eight bars in the first place.

![Why 92 BPM and 8 Bars Fit Intros — AI podcast intro music](https://static.mm-ais.com/article-images-ai/ai-podcast-intro-music-pick-92-bpm-then-ai-8b1fb218.jpg)

## The Evidence: What the Tools Actually Ship

Before committing to an eight-bar cut, verify what your generator actually offers rather than assuming: open the tool's feature and pricing pages and confirm (a) that it can lock output to a BPM you set, (b) whether per-instrument stems are downloadable, and (c) whether a podcast-intro-style preset exists. The practical check either way is to set the BPM field to 92 before you generate, not after; changing tempo post-hoc shifts the grid you are cutting against and makes the downbeat harder to trust.

If your generator advertises a podcast-intro or branded-audio preset, treat it as a candidate starting point, not a guarantee that the exported file lands on your grid. Before you rely on it, verify that the preset still exposes a tempo control you can set to 92 and a loop-length control you can set to 8 bars; if either is locked or hidden, regenerate in manual mode so the bar-9 downbeat lands where your ruler says it should.

Adobe Podcast (podcast.adobe.com) frames its generative audio work around credits and web-based podcast and voiceover production, with a stated allowance of 4,000 generative credits per month and unlimited generation on select image models. For an intro loop, the useful check is whether your generation and any re-rolls fit inside the current plan’s credit behavior, since re-rolling a bed to fix one bar is a normal part of the workflow. Descript (descript.com) starts free, with paid plans from $16/month, and is positioned around podcasting and editing; treat that as a reason to confirm which plan tier you are on before you assume you can re-cut or duck inside the same project.

Spotify for Creators (creators.spotify.com) is distribution and growth tooling rather than a generator, but it matters at the handoff: the platform emphasizes uploads, clips, analytics, and show-page customization. That reinforces the export decision upstream. If your intro is going straight into an episode upload with no mid-roll re-version planned, a single stereo bounce is the clean handoff; if you expect to duck under voice or re-cut for a clip, keep the stem structure available while you still can.

DeepMind’s model family (deepmind.google) separates Lyria for high-fidelity music and audio generation from Veo for cinematic video with audio, which is a useful reminder that “audio-capable” and “music-generating” are not the same claim. When a tool markets a beat, a podcast intro mode, or a generative music model, check what it actually exports: a mixed stereo file, selectable instruments, or a preset-bound loop. That single check tells you whether the named-source convergence on loop-based AI generation and podcast-specific export modes helps your workflow or just adds a label to it.

## One File or Stems: The Edit-Count Rule

Before you export anything, count the downstream edit events sitting in your production calendar. The threshold is simple: zero or one planned event means a single stereo bounce; two or more means a three-stem export broken into drums, bass, and melodic elements. This is the only decision that matters at export time, and it should be made before you render, not after you discover you need to pull a voiceover under the bed.

Here is the event list to check. Event one is ducking the intro under a voiceover tag, where the bed drops in level as the host speaks. Event two is re-cutting the loop shorter for a 15-second video teaser, which means finding a new musical landing point inside the existing bed. Event three is re-versioning for a season rebrand, where the same loop needs a different tonal or instrumental identity without regenerating from scratch. If your show has one of these on deck, keep the stereo file. If it has two, switch to stems.

The reason the threshold sits at two is workload, not file size. With one stereo bounce, you take a single BS.1770 integrated loudness measurement and match that one number against your episode target. With three stems, you have to level-match each component so the summed output lands where the stereo version did, then null-check the combination against the original render to confirm nothing phase-cancels or jumps out of balance. That is three measurements plus a sum check instead of one, and every additional stem multiplies the comparison work.

Think about what each event actually asks of the file. Ducking under voice is a gain move on the whole bed, which a stereo bounce handles cleanly. Re-cutting for a teaser is a timing move that may expose seams inside the loop, and stems give you the option to drop individual layers at the new boundary rather than fade the whole mix. Re-versioning is a arrangement move, and separate drums, bass, and melodic files let you swap or thin parts without returning to the generator. Two events means you are doing more than gain riding, so the extra export pays for itself.

Platform workflow can shape the count, too. Spotify for Creators describes tools for audio and video podcasts, including clips and video thumbnails, so a show planning both an audio feed and a short video clip has a concrete reason to log the teaser re-cut as a second event. Descript lists paid plans starting at $16/month for podcasting and video work, which is a check worth running if your edit stack already includes a tiered tool and you want to know whether stem handling sits inside your current plan. Verify the feature against the plan you actually use rather than assuming it is included.

Make the call in writing on your export checklist: list the events, tally them, then choose one file or three. If the tally is one, resist the urge to export stems "just in case," because you are taking on three loudness matches for a job that needs one. If the tally reaches two, export the stems at the same time as the stereo reference so you have a null-check target. That single line on the checklist keeps the decision repeatable across episodes and across whoever handles the next season's rebrand.

## Costs, Loudness, and License Numbers

The budget for this workflow has three line items, and only one of them is a fixed number you can quote today. Descript's pricing page states a free tier with paid plans starting at $16/month, which matters if you plan to trim your intro in the same editor you use for the episode. SOUNDRAW, TopMediai, and FreeMusic AI all advertise free or freemium entry points, but the per-track commercial license fee is not uniform across them, so treat the sticker price of generation as separate from the price of clearance.

That separation is the check I run before I commit to a bed. Open the pricing page and the license page for whichever tool produced the loop, on the day you purchase, and read them as two documents. A free or freemium generation tier can still carry a paid commercial license, and the terms can change between the week you audition a loop and the week you publish. Screenshot both pages and save them with the episode's session files; that record is worth more than a remembered marketing claim.

Loudness is the second line item, and it is a measurement task rather than a published intro standard. I measure the intro bed and the finished episode separately with BS.1770 integrated loudness, then aim for the bed to sit below the voice so the host stays forward without me reaching for a fader mid-sentence. The method is the deliverable here: integrate over the full eight-bar bed, note the number, integrate over a representative stretch of speech, note that number, and compare the two before you print anything.

Do not substitute peak meters for that comparison. A loop can look harmless on a peak readout and still mask dialogue because its average energy is high, which is exactly the failure integrated measurement catches. If your editor exposes both loudness and true-peak readings, log them side by side for the bed and the voice pass, then make one level decision from the pair rather than nudging gain by ear across repeated plays.

The third line item is the license-verification step, and it is the one people skip. Before export, confirm that the specific track you are shipping is covered for podcast distribution under the terms shown at purchase time, not under the terms shown when the account was created. If the tool's page is ambiguous, that ambiguity is a reason to choose a different loop or a different tier, not a reason to assume coverage.

Keep these three checks together in your prep notes: the tool's current pricing and license pages, the two integrated loudness readings, and the clearance decision for the exact track. None of them changes the edit-count rule you already applied to the file structure; they simply make sure the file you cut is cleared, level-matched, and paid for on the terms that were actually in force when you took it.

## What the Evidence Does Not Establish

The evidence in this topic supports feature claims and plan limits, not performance guarantees. A tool can state that it generates loops or exports stems; that does not establish that a given 92 BPM loop will seam cleanly, that listeners prefer the full length, or that a license covers your exact podcast use. Treat each of those as a separate verification task.

## Frequently Asked Questions

**At what exact time should an 8-bar 4/4 loop at 92 BPM be cut for an AI podcast intro?**

Cut it at 20.87 seconds, the end of bar 8, i.e. the bar-9 downbeat on your DAW ruler.

**How many beats are contained in eight bars of 4/4 at 92 BPM?**

Eight bars of 4/4 at 92 BPM is 32 beats.

**What arithmetic produces the 20.87-second bed length?**

32 × (60 ÷ 92) = 20.8696 seconds, which rounds to 20.87 seconds.

**When should the 20.87-second cut time be changed?**

Change that answer only if you plan at least 2 downstream edits, such as ducking under voice and re-cutting or re-versioning, which calls for a 3-stem export.

**How can the cut time be checked against a DAW?**

Verify it against your DAW's tempo ruler by parking the playhead on the bar-1 downbeat and reading the ruler at the bar-9 downbeat.

**Why does cutting on the bar line matter for tempo-locked AI generators?**

Tempo-locked AI generators quantize their loop grid to the BPM they generated, so the bar line is the structural seam the generator planned for.

## Quick answers

| What loop length should you pick for an AI podcast intro at 92 BPM in 4/4? | Pick an 8-bar loop and cut it at 20.87 seconds — the end of bar 8, i.e. the bar-9 downbeat on your DAW ruler. |
| --- | --- |
| When should you change that answer? | Change that answer only if you plan at least 2 downstream edits, such as ducking under voice and re-cutting or re-versioning, which calls for a 3-stem export. |
| Why does 8 bars of 4/4 at 92 BPM fit intros? | Eight bars of 4/4 at 92 BPM is 32 beats, and 32 × (60 ÷ 92) = 20.8696 seconds, which rounds to 20.87 seconds. |
| How should you verify the loop length in your DAW? | Verify it against your DAW's tempo ruler rather than trusting your ear: park the playhead on the bar-1 downbeat, then read the ruler at the bar-9 downbeat (the end of bar 8). |
| What happens if your session shows anything other than 20.87s at that point? | If your session shows anything other than 20.87s at that point, the project tempo or time signature is not what you think it is, and every later cut will drift. |

Also worth reading: **Ableton Live 12 AI Drum Patterns: Verify the 12-Minute Intro Workflow**: [Ableton Live 12 AI Drum](https://getrhythmm.com/blog/ableton-live-12-ai-drum-patterns-verify-the-12-minute-intro-workflow.php) · **Ableton Live 12 MIDI vs Bounce at 75 BPM Over 200 Bars**: [Ableton Live 12 MIDI vs](https://getrhythmm.com/blog/ableton-live-12-midi-vs-bounce-at-75-bpm-over-200-bars.php) · **2026 140 BPM AI Drums: -12 dB Sidechain Cuts Masking for Streams**: [2026 140 BPM AI Drums:](https://getrhythmm.com/blog/2026-140-bpm-ai-drums-12-db-sidechain-cuts-masking-for-streams.php)

### Related reading

- [2026 A/B Test: AI Drum Loops vs DAW Patterns for Podcast Intros](https://getrhythmm.com/blog/2026-ab-test-ai-drum-loops-vs-daw-patterns-for-podcast-intros.php)
- [How to create custom beats for your podcast intro](https://getrhythmm.com/blog/how_to_create_custom_beats_for_your_podcast_intro.php)
- [Ableton Live 12 AI Drum Patterns: Verify the 12-Minute Intro Workflow](https://getrhythmm.com/blog/ableton-live-12-ai-drum-patterns-verify-the-12-minute-intro-workflow.php)
- [Ableton Live 12 MIDI Transform for drums: 2026 probability vs manual step-entry](https://getrhythmm.com/blog/ableton-live-12-midi-transform-for-drums-2026-probability-vs-manual-step-entry.php)
- [How to Make Lo-Fi Beats: Lane-Specific AI Lock vs Swing at 70-90 BPM](https://getrhythmm.com/blog/how-to-make-lo-fi-beats-lane-specific-ai-lock-vs-swing-at-70-90-bpm.php)
- [EU AI Act 2026: How Transparency Rules Touch Stem Plugin Royalty Splits](https://getrhythmm.com/blog/eu-ai-act-2026-how-transparency-rules-touch-stem-plugin-royalty-splits.php)

### Latest

- [Ableton Live 12 AI Drum Patterns: Verify the 12-Minute Intro Workflow](https://getrhythmm.com/blog/ableton-live-12-ai-drum-patterns-verify-the-12-minute-intro-workflow.php)
- [Ableton Live 12 MIDI Transform for drums: 2026 probability vs manual step-entry](https://getrhythmm.com/blog/ableton-live-12-midi-transform-for-drums-2026-probability-vs-manual-step-entry.php)
- [How to Make Lo-Fi Beats: Lane-Specific AI Lock vs Swing at 70-90 BPM](https://getrhythmm.com/blog/how-to-make-lo-fi-beats-lane-specific-ai-lock-vs-swing-at-70-90-bpm.php)

Canonical: https://getrhythmm.com/blog/ai-podcast-intro-music-pick-92-bpm-then-cut-loops-at-8-bars.php
Markdown: https://getrhythmm.com/blog/ai-podcast-intro-music-pick-92-bpm-then-cut-loops-at-8-bars.php/index.md

```json
{"@context":"https://schema.org","@type":"BlogPosting","headline":"AI podcast intro music: pick 92 BPM, then cut loops at 8 bars","url":"https://getrhythmm.com/blog/ai-podcast-intro-music-pick-92-bpm-then-cut-loops-at-8-bars.php","description":"Cut AI podcast intro music at 92 BPM to an 8-bar loop ending at 20.87 seconds, with 3-stem exports for ducking, re-cuts, and future re-versioning.","datePublished":"November 19, 2026","author":{"@type":"Person","name":"Evelyn Porter"},"articleBody":"For an AI podcast intro at 92 BPM in 4/4, pick an 8-bar loop and cut it at 20.87 seconds — the end of bar 8, i.e. the bar-9 downbeat on your DAW ruler. Change that answer only if you plan at least 2 downstream edits, such as ducking under voice and re-cutting or re-versioning, which calls for a 3-stem export. ![For an AI podcast intro at 92 BPM — AI podcast intro music](https://static.mm-ais.com/article-images-ai/ai-podcast-intro-music-pick-92-bpm-then-ai-ed9cad42.jpg)..."}
```
