How to Fix Latency When Recording Guitar Through an Audio Interface
Latency makes recording guitar miserable. Here is how to diagnose the delay, fix it through buffer settings, drivers, and monitoring, and hit near-zero latency.
Mike Reynolds
Professional Guitarist & Audio Engineer · 20+ years
ℹ️ Affiliate Disclosure: Music Gear Specialist earns from qualifying purchases through Amazon and other partner links. This doesn't affect our recommendations—we only suggest gear we'd use ourselves.
ℹ️ Affiliate Disclosure: Music Gear Specialist earns from qualifying purchases through Amazon and other partner links. This doesn't affect our recommendations—we only suggest gear we'd use ourselves.
You strum a chord on your guitar, and the sound comes back through your headphones a split second later. That split second - the latency - makes it impossible to play naturally. Your timing falls apart, your feel disappears, and recording becomes an exercise in frustration.
Latency is the single biggest workflow killer in home recording, and it is entirely fixable. The problem is not your interface or your computer - it is how they are configured.
Understanding Where Latency Comes From
Latency in a recording setup has three components:
Input latency: The time for your guitar signal to travel from the interface’s input through the analog-to-digital converter into the computer. This is typically 1-3 milliseconds and is fixed by hardware - you cannot change it.
Processing latency (buffer): The time the computer takes to process the audio data through your DAW and any plugins. This is the biggest source of delay and the one you can control. It is determined by your buffer size setting.
Output latency: The time for the processed audio to travel from the computer back through the digital-to-analog converter in the interface to your headphones. This is also typically 1-3 milliseconds and is hardware-fixed.
Total round-trip latency = input latency + processing latency + output latency.
The processing latency (buffer) is where 80-90% of perceived delay lives. Fixing the buffer fixes the problem.
Step 1: Use the Correct Audio Driver
This is the single most impactful fix, and it is the step most beginners miss.
Windows: Use ASIO (Not WASAPI, MME, or DirectSound)
Windows has multiple audio driver frameworks, and most of them add significant latency:
| Driver Type | Typical Latency | Use Case |
|---|---|---|
| MME/DirectSound | 50-100ms | Old, legacy - never use for recording |
| WASAPI | 15-30ms | Adequate for playback, too slow for recording |
| ASIO (manufacturer) | 3-10ms | The only acceptable option for recording |
| ASIO4ALL | 10-20ms | Fallback if no manufacturer ASIO driver exists |
Your interface manufacturer’s ASIO driver is mandatory for low-latency recording on Windows. If you are recording through WASAPI or MME, you will always have perceptible latency regardless of buffer settings.
Install your manufacturer’s ASIO driver:
- Focusrite: Focusrite Control (includes ASIO driver)
- PreSonus: Universal Control
- Behringer: Behringer ASIO driver (or ASIO4ALL as fallback)
- Universal Audio: UA Connect
- Audient: Audient iD app
In your DAW, select the ASIO driver in audio preferences. In Ableton, go to Preferences > Audio > Driver Type > ASIO. In Reaper, go to Options > Preferences > Audio > Device > Audio System: ASIO.
macOS: CoreAudio Works Well
macOS’s CoreAudio driver is efficient and low-latency by default. You do not need a separate ASIO driver on Mac. Just select your interface in your DAW’s audio preferences and adjust the buffer size.
Step 2: Set the Buffer Size Correctly
The buffer size determines how many audio samples the computer processes at once. Smaller buffers = less latency but more CPU load. Larger buffers = more latency but more CPU headroom for plugins.
| Buffer Size (samples) | Latency at 44.1kHz | Latency at 96kHz | Best For |
|---|---|---|---|
| 32 | ~1.5ms | ~0.7ms | Only works on fast computers, few plugins |
| 64 | ~3ms | ~1.3ms | Recording with amp sims - ideal if stable |
| 128 | ~6ms | ~3ms | Recording - best balance for most setups |
| 256 | ~12ms | ~6ms | Recording with moderate plugin load |
| 512 | ~23ms | ~12ms | Mixing - perceptible delay for recording |
| 1024 | ~46ms | ~23ms | Mixing only - too laggy for live input |
| 2048 | ~93ms | ~46ms | Heavy mixing sessions - unusable for recording |
For recording guitar, set your buffer to 128 samples. This gives roughly 6ms of processing latency at a 44.1kHz sample rate. Combined with input/output conversion latency, your total round-trip will be approximately 8-12ms - below the threshold where most musicians perceive delay.
If 128 samples causes audio crackling or dropouts, increase to 256. If your system handles it without glitches, try 64 for even tighter response.
Where to change buffer size:
- In your DAW’s audio preferences - most DAWs let you set buffer size directly
- In your interface’s control software - some interfaces (like Focusrite Scarlett via Focusrite Control) set the buffer size at the driver level
- In ASIO settings - if your DAW has an “ASIO Settings” or “Control Panel” button in audio preferences, it opens the driver’s buffer configuration
Step 3: Optimize Your System for Low Latency
A buffer size of 128 samples demands that your computer processes audio fast enough to fill each buffer before the next one is needed. If it cannot keep up, you get crackling, popping, or audio dropout - the dreaded “buffer underrun.”
Close Background Applications
Every application running on your computer competes for CPU time. Before recording:
- Close web browsers (Chrome is particularly CPU-hungry)
- Quit Spotify, Discord, and other audio applications - these can conflict with your interface’s ASIO driver
- Disable cloud sync services (Dropbox, OneDrive, Google Drive) that perform background disk activity
- Close antivirus real-time scanning temporarily, or add your DAW and audio driver to the exclusion list
Windows-Specific Optimizations
Disable USB Selective Suspend: This power-saving feature interrupts USB device communication and causes audio glitches. Control Panel > Power Options > Change Plan Settings > Change Advanced Power Settings > USB Settings > USB Selective Suspend > Disabled.
Set Power Plan to High Performance: Control Panel > Power Options > High Performance. The “Balanced” plan throttles CPU speed to save power, which increases audio processing time.
Disable Wi-Fi during recording: Wi-Fi adapters generate interrupt requests (IRQs) that compete with USB audio for CPU attention. If you do not need internet during recording, disable Wi-Fi.
Check DPC Latency: Download LatencyMon (free) and run it. This tool shows whether your system has driver issues causing latency spikes. If LatencyMon shows red warnings, specific drivers (often network, GPU, or USB drivers) are causing interrupt latency that affects audio performance. Update or disable the offending drivers.
macOS-Specific Optimizations
macOS handles audio processing more efficiently than Windows, but heavy background tasks still interfere:
- Disable Spotlight indexing for your audio project drives
- Close Time Machine if it is backing up during a session
- Disable Bluetooth if you are not using wireless peripherals
Step 4: Use Direct Monitoring for Zero-Latency Playing
If you cannot get your buffer low enough without crackling, or if you need absolute zero latency while playing, direct monitoring is the solution.
Direct monitoring routes your guitar signal from the interface’s input straight to your headphones output, bypassing the computer entirely. The signal goes: guitar > interface input > interface headphone output. No computer, no DAW, no latency.
How to enable direct monitoring:
Most interfaces have a hardware direct monitor switch or knob:
- Focusrite Scarlett 2i2: The “DIRECT” button on the front panel
- Audient iD4: The “Direct Monitor” button
- Behringer UMC202HD: The “DIRECT MON” switch
- Universal Audio Volt 2: The “Monitor” knob blends direct and DAW signal
In software-controlled interfaces, direct monitoring is configured in the companion app (Focusrite Control’s mixer, UA Connect’s monitoring section).
The trade-off: With direct monitoring, you hear your dry guitar signal - no amp simulation, no effects, no reverb. If you are using amp modeling software like Neural DSP, Guitar Rig, or Amplitube, direct monitoring bypasses those plugins.
Workaround: Some interfaces and DAWs support “hybrid monitoring” - hearing the direct signal with a small amount of DAW-processed signal blended in. Focusrite Control’s mixer allows this. Alternatively, use a hardware amp modeler or multi-effects pedal (like a Line 6 HX Stomp or Boss GT-1000) before the interface. The hardware processing adds no computer-related latency.
Step 5: Reduce Plugin Latency
Even with a small buffer, certain plugins add their own processing latency:
Linear-phase EQs: Add significant latency (sometimes hundreds of milliseconds) for phase-accurate processing. Never use these on your monitor bus while recording.
Look-ahead compressors and limiters: Add latency equal to their look-ahead time (typically 1-5ms). Disable during tracking, enable during mixing.
Convolution reverbs with large IR files: Can add latency depending on implementation. Use algorithmic reverbs during recording sessions.
Oversampled plugins: Some amp sims and distortion plugins oversample at 2x or 4x for better sound quality, which increases processing time. Check plugin settings for oversampling options and disable them during tracking.
General rule for recording sessions: Use the minimum number of plugins on your input monitoring chain. A simple amp sim plus a touch of reverb is all you need while tracking. Add the full effects chain during mixing, when you can increase the buffer size to 1024 or 2048 without any latency consequences.
Step 6: Sample Rate Considerations
Higher sample rates reduce latency at the same buffer size because each buffer fills faster:
At 128 samples buffer:
- 44.1kHz = ~6ms processing latency
- 96kHz = ~2.7ms processing latency
- 192kHz = ~1.3ms processing latency
Recording at 96kHz instead of 44.1kHz nearly halves your latency at the same buffer size. However, it also doubles CPU load, disk usage, and file sizes. For most home recording, 44.1kHz at a buffer of 64-128 samples provides low enough latency without the overhead of higher sample rates.
If you have a powerful computer and large storage, recording at 96kHz with a 128-sample buffer gives you approximately 3ms of processing latency - essentially imperceptible.
The Practical Recording Workflow
Here is the workflow that eliminates latency as a problem in your sessions:
Before recording:
- Set buffer to 128 samples (or 64 if your system handles it)
- Close unnecessary background apps
- Minimize plugins on the monitoring chain - one amp sim, one reverb max
- If using Windows, verify you are on the ASIO driver
During recording:
- If you hear unacceptable latency, enable direct monitoring and record dry
- Record multiple takes without stopping to adjust settings - tweaking mid-session costs more time than slightly imperfect monitoring
After recording (mixing):
- Increase buffer to 512 or 1024 samples - latency no longer matters because you are playing back recorded audio, not monitoring live input
- Add all the plugins, effects chains, and processing you want
- The higher buffer gives your CPU more headroom for complex plugin loads
This two-phase approach - low buffer for tracking, high buffer for mixing - is how every professional studio operates. It eliminates latency during performance while maximizing processing power during production.
If you are still building your recording setup, our home recording guide covers the full gear chain and explains how each component in the signal chain affects your sound and workflow. And if your interface is not showing up at all, check our interface recognition troubleshooting guide before diving into latency settings.
Mike Reynolds
•Editor & Lead Reviewer · 70+ articles published
Mike Reynolds covers guitars, amps, pedals, and recording gear for Music Gear Specialist. With 70+ articles published and hundreds of hours researching music equipment, he focuses on honest recommendations based on real user experiences, community feedback, and manufacturer specifications.