Show spoken script
Welcome back. In this lesson we’re going to level up one of the most underrated “finish-the-track” skills in Ableton Live 12: track freeze strategies, specifically for drum and bass workflow.
Because if you’re making DnB, you already know the pain. You’ve got a reese with five stages of distortion, a multiband, a huge reverb tail, maybe some spectral stuff… then you add parallel drum smack, automation everywhere, and suddenly your CPU meter is screaming, your buffer has to go up, and every creative decision turns into a technical delay.
Freezing fixes that. But more importantly, freezing is a mindset: commit the tone, then edit the audio like a sampler. That’s where tracks actually get finished.
By the end of this lesson, you’ll know when to freeze, when to flatten, and when to resample, and you’ll have a practical “commit workflow” you can use every time you build a rolling DnB drop.
Alright, let’s set up the session in a way that makes freezing painless.
First: set a realistic buffer size while producing. Go to Preferences, Audio, and set Buffer Size somewhere around 256 to 512 samples. If you’re recording MIDI live and you want it to feel tight, lean toward 256. If your project’s already heavy, 512 will save you.
Next: keep an eye on the CPU meter in the top right. Don’t obsess over it, but treat it like a warning light. If it starts spiking when you hit play, that’s your sign it’s time to commit.
Now group your tracks like a normal DnB template. Make a DRUMS group with Kick, Snare, Break, Hats or Tops, Perc. Make a BASS group with Sub, Mid or Reese, and maybe a Bass FX lane. Then a MUSIC or ATMOS group, and an FX or IMPACTS group.
And do the boring part now: rename and color code. This matters because once you start printing and committing, you’ll create extra tracks and versions. If you don’t name things clearly, you’ll lose the best take, or flatten the wrong thing, or just stop trusting your session.
Now, what does Freeze actually do in Live 12?
When you freeze a track, Ableton renders it to a temporary audio file including the instrument, the audio on the track, and the device chain. The track becomes basically “playback only” for the sound design side. You can still adjust volume and pan, and you can mute and solo. But you can’t tweak the synth, you can’t change the device parameters, and you can’t edit MIDI notes unless you unfreeze.
Here’s the rule of thumb for drum and bass: freeze when you’re about 80% happy with the sound and you want momentum. Not when it’s perfect. When it’s close enough that continuing to tweak is more about procrastination than progress.
Now let’s talk about what to freeze first, because some tracks are just CPU criminals.
The biggest one is usually your mid bass, your reese, your tech bass rack. A typical chain might be Wavetable or Operator with unison, then a saturator, maybe Roar or an amp and cab for grit, an auto filter with LFO movement, multiband dynamics to glue and add aggression, and then some kind of wide reverb like Hybrid Reverb for those techy tails.
That chain sounds amazing. It also eats CPU.
So the move is: once the tone is close, freeze the MID BASS track. Not the sub, we’ll come back to that. The mid.
Next offender: reverbs, delays, and long FX. If you’ve got Hybrid Reverb, convolution style spaces, or massive Echo feedback throws, freeze them early. These are often “set and forget.” They’re not the musical core of the track. They’re vibe support. Commit them.
Third offender: anything in high quality or oversampling mode, especially third-party plugins. The moment you approve the sound, freeze it. And here’s an extra coach note: freezing is also a latency fix tool, not just a CPU tool. If you’ve got lookahead limiting, linear-phase EQ, or oversampled processing causing monitoring lag, freezing can make MIDI performance feel tight again, especially at 128 to 256 buffer.
Now we need to be super clear on three different commit options: Freeze, Flatten, and Resample. In DnB, each one has a different job.
Option one is Freeze. Freeze is “keep flexibility.” You freeze when you still might need to tweak the patch later, but you want CPU relief so you can keep arranging. Right-click the track and choose Freeze Track, then keep moving.
Option two is Flatten. Flatten is “commit to audio and start editing.” This is the jungle mindset. If you want chops, reverses, gates, tight call-and-response hits, flatten the frozen track. Right-click, Flatten, and now you have an audio clip you can slice up.
A classic DnB move is to flatten your bass, then slice it into eighth-note and sixteenth-note chunks in the drop. Suddenly, your bass arrangement becomes fast, visual, and super intentional. You stop endlessly tweaking the synth and start writing the record.
Option three is Resample. Resampling is “print creatively,” especially when you want to capture a vibe that includes sidechain pumping and bus processing.
Here’s the clean way to do it: create a new audio track and name it something like BASS PRINT. Set its input not to Resampling from the master unless you really need everything; instead, set Audio From to your bass group. That’s usually cleaner and easier to manage. Arm the track and record an 8 to 16 bar pass.
Now you’ve got a performance you can chop into fills and switchups.
And here’s a huge tip: if you resample the bass group after sidechain, you imprint that bounce. That rolling feel is sometimes the whole identity of the drop. Printing it can lock in the groove.
But there’s also a smart advanced trick called the ghost sidechain print technique. Print your bass without sidechain first, because it’s easier to edit cleanly when the transients aren’t being dipped. Then apply sidechain on the printed audio with a simple compressor or volume shaper. That way you can still change the pump pattern later without reopening sound design.
Now let’s talk about freezing with groups, because it can save you, or ruin your day.
If you freeze a group track, you’re committing the whole stack under it. That’s sometimes useful, but in DnB it’s usually too aggressive early on.
For drums, freeze the expensive layers, not the core groove. Keep your kick and snare editable longer because you’ll keep adjusting punch, transient shape, and balance. But freeze break processing chains once the character is right.
A typical break chain might be Drum Buss for drive and crunch, then a saturator soft clipping, EQ Eight to high-pass around 30 to 50 hertz and maybe dip harshness around 3 to 6k, then Glue Compressor doing one to three dB of gain reduction, and maybe Redux lightly for jungle texture.
Once the break is vibing, freeze that break track. It stabilizes the character and saves CPU, while you keep the kick and snare free to evolve.
For bass: split sub and mid so you can commit intelligently. Keep the sub as MIDI longer. A simple Operator sine or triangle is cheap on CPU and it’s critical to keep it in tune with the track as you arrange. But freeze the mid or reese early, because that’s where the heavy chain lives.
Now here’s where freezing turns into a workflow weapon: arrangement blocks.
Build Drop A as a 16 bar block. Get it rolling. Then freeze or flatten the mid bass and the atmos and FX. Duplicate the section to make Drop A2. Now you’re editing audio, not reinventing sound design. That’s how you make variation quickly.
Some quick edit ideas you can do immediately after flattening: reverse a bass hit at the end of bar 15 leading into bar 16. Gate the bass rhythmically using Auto Pan with phase at zero degrees so it acts like a volume gate, synced to an eighth note for stutters. Or create call and response by muting every second bar chunk. These are fast, musical decisions that come from committing.
And there’s a deeper reason this works: committing audio forces arrangement decisions. In drum and bass, that’s the difference between an eight-bar loop and a finished record.
Now let’s protect you from regret with a simple professional habit: the pre-freeze safety track.
Before you freeze, duplicate the track, rename the copy something like MID BASS SAFE, and turn it off using the track activator. Then freeze the working version.
This is huge. It means you can keep moving forward without feeling like you’re burning bridges. You’ve got the patch preserved without having to unfreeze, re-freeze, and potentially break your flow.
Another powerful use: freezing automation and modulation.
In DnB you’re constantly automating filter cutoff sweeps, distortion drive, reverb sends, FM amounts, unison movement, all that. When you finally nail a moment, like the last two bars into the drop, freeze it or flatten it and treat it like a sampled performance. It turns something complex and CPU-heavy into a stable piece of audio you can place and repeat with confidence.
Quick warning though: don’t trust freeze to catch external randomness. If you’re using devices or plugins with random seeds, generative modulation, or free-running LFOs, unfreezing and refreezing might produce different audio each time. If you want to truly lock a specific performance, resample it in real time and keep that take.
Now a few common mistakes to avoid.
First: freezing too early. If you’re still unsure about the notes or the groove, freezing will slow you down, because you’ll have to unfreeze just to fix musical problems. Freeze tone once the music is right.
Second: flattening before checking transitions. Long reverb tails and delay throws can get chopped awkwardly. A pro move is: when you flatten or resample, print an extra bar past the section so the tails have room. Then trim with fades intentionally.
Third: freezing the whole drums group too soon. If you commit the entire drum stack, swapping a snare or adjusting the kick transient becomes a headache. Commit layers, not the whole foundation.
Fourth: printing too hot. If your bass is slamming near zero pre-master, once it’s audio you’ve limited your headroom choices. Aim for sensible peaks on busy groups, roughly minus 12 to minus 6 dBFS before they hit your master chain. Think of freezing as taking a mix snapshot. You want it healthy, not clipped.
Fifth: forgetting sidechain relationships. If your drop relies on the bass pumping around the kick and snare, and you print it without sidechain, the energy can collapse. Decide up front whether you’re printing with pump, or using the ghost sidechain technique and adding pump later.
Now let’s do a tight 15-minute practice to lock this in.
Make a MID BASS using Wavetable. Start simple: Oscillator one on a saw. Add unison, maybe four to seven voices depending on how wide and heavy you want it.
Add a chain: Saturator with about six dB of drive and soft clip on. Then Auto Filter on a low-pass, and automate the cutoff over eight bars. Then Multiband Dynamics lightly, just one to three dB of gain reduction, more like glue than destruction.
Write a two-bar rolling bassline. Syncopated. Leave spaces for the snare. That negative space is part of the groove.
Duplicate it out to 16 bars. That’s Drop A.
Now freeze the MID BASS track. If you’re happy, flatten it.
Then do two to four edits: reverse the last hit of bar eight. Create a one-sixteenth stutter in bar fifteen by duplicating a tiny slice. Add a silent gap before a hit for impact, even just 10 to 40 milliseconds can make things punch harder.
Play it back with your drums. Listen for the bounce. If the groove feels better after committing, you’re doing it right.
Before we wrap, let’s add a couple advanced ideas you can steal immediately.
Try multi-print bass. Duplicate your mid bass chain into two or three variants, like one with a different filter position, one with different distortion, one with extra chorus. Print each for 16 bars. In arrangement, swap between prints every four bars. You get evolution without touching the synth again.
Try two-stage bass printing. First print the core timbre with minimal movement, basically committing the identity. Then on the audio, add movement effects like phaser, autofilter, frequency shifter, grain tools, and if it gets heavy, print again. It’s stable but still experimental.
And use a naming convention so you don’t lose your mind. Something like: MID BASS for the MIDI and devices, MID BASS_PRINT_01 for the audio commit, and MID BASS_PRINT_01_EDITS for your chopped arrangement. That’s how you iterate fast without destroying good takes.
Recap time.
Freeze is your CPU and momentum tool in Ableton Live 12. In drum and bass, freeze your mid bass chains, big reverbs, spectral effects, and heavy break processing. Flatten when you’re ready to do jungle-style chops, gating, and call-and-response edits. Resample when you want to capture the vibe, especially sidechain bounce and bus processing.
Work in 8 or 16 bar blocks. Commit. Then arrange with audio like you’re using a classic sampler.
If you want to go one level deeper, tell me what your biggest bottleneck is right now: bass CPU, drum bus, FX returns, or master chain. And tell me if your bass is mostly Wavetable or Operator, or third-party like Serum or Vital. I’ll map out a clean freeze and print plan that fits your template.