Home Vocal Recording and Mixing

What sample rate and bit depth should you record vocals at in REAPER

By StudioRescue Editorial

Published October 15, 2026 · Updated October 15, 2026

This question gets asked constantly, and the honest answer is boring, which is probably why so few articles give it plainly. For a voice recorded at home, your sample rate choice is one of the smallest factors in how the finished take sounds. The room, the mic position and your record level all matter far more. So the goal here is not to find a magic number. It is to make a defensible choice once, set it in REAPER, and never lose another session to it.

Bit depth is different. There is a real, practical reason to choose 24-bit over 16-bit, and it has nothing to do with the sound of silence. It is about how much room you have to get your record level wrong.

Why the question keeps getting asked

Two things make this confusing for beginners. First, the numbers look like quality settings, so bigger reads as better, and 96 kHz sounds like it must beat 48 kHz. Second, the advice you find is usually written for a different job than yours. Orchestral scoring, sound design with heavy pitch shifting, and archival transfer all have reasons to use higher rates. Recording one voice into a laptop does not inherit those reasons.

It also keeps getting asked because the cost of getting it wrong is real but delayed. You find out weeks later, when a file plays back at the wrong speed or a client rejects a delivery. The repair is usually possible, but it is work you did not need to do.

Sample rate: what it decides, and what it does not

Sample rate sets the highest frequency that can be captured. A sampled system can represent frequencies up to half its sample rate, so 44.1 kHz reaches about 22 kHz and 48 kHz reaches about 24 kHz. Healthy young human hearing tops out around 20 kHz and falls with age. Both standard rates already cover the audible range with room to spare, which is why neither is the reason your vocal sounds the way it does.

What sample rate genuinely decides is compatibility and conversion. That is the useful way to pick.

When 48 kHz is the better pick

Choose 48 kHz if anything you make might end up attached to video, including YouTube, Reels, course videos or voiceover work. Google's recommended upload encoding settings for YouTube specify 48 kHz audio, and video production has used 48 kHz as its working rate for decades. Recording at the rate your video will be delivered at removes a conversion step, and it also means audio and video timelines agree from the start.

When 44.1 kHz is the better pick

Choose 44.1 kHz if your output is music going to streaming services, or if you narrate audiobooks. Spotify's delivery requirements accept 44.1 kHz or higher and reduce anything above 24-bit down to a maximum of 44.1 kHz, 24-bit on their side. For audiobooks the requirement is firm rather than a preference: ACX requires each delivered file to be a 192 kbps or higher constant bit rate MP3 at 44.1 kHz. If that is your work, recording at 44.1 kHz removes a conversion step from every single delivery. The rest of the ACX numbers are covered in how to meet ACX audiobook specs in REAPER.

Why 96 kHz is usually the wrong pick for a voice

Doubling the sample rate doubles your disk use and roughly doubles the processing load per plugin. On a laptop that is already struggling, that is the difference between a session that records cleanly and one that crackles. In exchange, for a single voice, you get content above 24 kHz that neither you nor your listener can hear, and you will almost certainly convert back down to 44.1 or 48 kHz on the way out anyway.

Bit depth: this one has a real answer

Record 24-bit. The reason is headroom, not audible noise.

For an ideal converter, the signal-to-noise ratio works out to roughly 6.02 dB per bit plus 1.76 dB. That puts 16-bit near 98 dB and 24-bit near 146 dB. No real audio interface reaches the 24-bit figure, so the extra bits are not buying you silence you can hear. What they buy is permission to record conservatively.

At 24-bit you can track with peaks sitting around -18 to -12 dBFS, leaving plenty of room for a loud phrase, and still raise that take later without dragging up an audible noise floor. At 16-bit the same cautious level leaves you much closer to the floor, which pushes people to record hot, which is how takes get clipped. Setting that record level is its own skill, covered in how to gain stage vocals in REAPER.

What about 32-bit float?

REAPER can record 32-bit floating point as well. Its practical advantage is that samples above full scale are not permanently squared off in the file, so a take that overshoots can often be pulled back down intact. The costs are larger files and the occasional tool or delivery route that expects fixed-point audio. Spotify, for example, asks for 24-bit when that is your native master and converts anything above it. A reasonable position: 24-bit for normal sessions, and 32-bit float when you are capturing something unrepeatable at a level you cannot control, like a live take or an interview.

Set it once in REAPER

Three places need to agree, and the order matters.

On the way out, your render settings are a separate decision from your recording settings, and that is where the conversion to a delivery rate happens. See how to export audio from REAPER.

What these settings will not fix

This is the part most versions of this article leave out.

• They will not improve your room. Reflections and echo are captured identically at every rate and depth.
• They will not make a cheap mic sound expensive. Bit depth records what arrives at the converter. It does not change what the capsule heard.
• They will not undo clipping that already happened. A take recorded into the red at 24-bit is still a clipped take.
• They will not fix tone. Muddy, nasal or harsh are all tone problems and none of them are rate or depth problems.
• They will not survive a mismatch. The right numbers in REAPER with the wrong number on the interface still gives you a speed problem.

If something already went wrong

1. Playback is slow and deep, or fast and thin. That is a rate mismatch between the file and the project, and it is repairable. Walk through how to fix audio recorded at the wrong sample rate in REAPER.
2. The take is clean but far too quiet. At 24-bit this is usually recoverable. 16-bit quiet takes raise more noise with them.
3. Everything was recorded at 96 kHz and the session struggles. Finish at the rate you recorded rather than converting mid-project, then record the next session lower.
4. A client or platform rejected the file. Check the delivery spec, not the recording spec. Those are different numbers and the delivery one is usually what failed.