Sample Rates in Film and Audio Post: 48kHz vs. 96kHz and When Higher Actually Matters

The 96kHz Session Nobody Could Open

A composer finishes the score for a 90-minute documentary. The recording sessions were conducted at 96kHz / 24-bit -- the composer's default setting for all projects because "higher is better." The score sounds exceptional. The final stems are exported and sent to the re-recording mixer.

The mixer opens the files. Their session is configured at 48kHz because all the location audio, dialogue, and sound effects are at 48kHz. The 96kHz stems are automatically sample-rate converted by Pro Tools on import. The conversion is seamless and inaudible. The mixer completes the mix at 48kHz and delivers a 48kHz / 24-bit master.

The composer later asks whether the 96kHz recording "made a difference." The honest answer: no. At every stage after the recording session, the audio was converted down to 48kHz -- the broadcast delivery standard. The 96kHz session doubled the file size of every recording, required double the storage, increased the CPU load of every plug-in in the session, and produced no audible benefit in the final deliverable.

This isn't an argument against 96kHz in all contexts. There are situations where higher sample rates matter. This post explains what sample rate actually does, when going beyond 48kHz is justified, and when it adds cost and complexity without adding quality.

Technical standards referenced in this post draw from the AES (Audio Engineering Society) standards for professional audio, the ITU-R BS.1770 loudness standard, the EBU R68 standard for broadcast audio alignment, and published specification data from Sound Devices, Zoom, and Tascam field recorders.

What Sample Rate Actually Controls

Sample rate is the number of times per second a digital audio system measures (samples) the amplitude of an analog audio signal. At 48kHz, the system takes 48,000 amplitude measurements every second. At 96kHz, it takes 96,000. Each measurement is stored as a digital value whose precision is determined by bit depth.

The Nyquist theorem (formally the Nyquist-Shannon sampling theorem) states that a digital system can accurately reproduce any frequency up to exactly half the sample rate. At 48kHz, the system can accurately represent frequencies up to 24kHz. At 96kHz, the ceiling is 48kHz. At 192kHz, the ceiling is 96kHz.

Human hearing has an upper frequency limit of approximately 20kHz, and in adults over 30 that limit typically drops to 16-17kHz. A 48kHz sample rate can reproduce all audio frequencies within human hearing with full accuracy. The headroom above 20kHz (up to 24kHz at 48kHz sample rate) provides margin against the anti-aliasing filter used to prevent high-frequency aliasing artifacts.

Why, then, does 96kHz exist? Several reasons are given, with varying degrees of validity:

First, the anti-aliasing filter in a 48kHz system must roll off sharply between 20kHz and 24kHz to prevent aliasing. Aggressive filter slopes can cause subtle phase shifts in the upper audible range (16-20kHz). A 96kHz system can use a gentler filter slope with its rolloff above 40kHz -- well outside human hearing -- and the upper audible range benefits from the less aggressive filtering.

Second, some processing operations (pitch shifting, time stretching, certain harmonic saturation plug-ins) produce intermodulation products that benefit from headroom above 20kHz. If a distortion plug-in generates harmonics at 25kHz, at 48kHz those harmonics alias back into the audible range as distortion artifacts. At 96kHz, they remain above the Nyquist ceiling and are filtered out cleanly.

Third, ultrasonic content from certain acoustic sources (cymbals, tape saturation, certain analog electronics) exists above 20kHz and is captured more completely at 96kHz. Whether this contributes to audible quality is debated -- the evidence from controlled listening tests is mixed and does not consistently show audible improvement.

Sample Rate Comparison for Film Audio Applications

The file size multiplier is exact: a 96kHz file is exactly double the size of the same recording at 48kHz for the same bit depth and duration. A 192kHz file is exactly four times larger. The CPU multiplier is approximate -- plug-in processing load scales roughly with sample rate but varies by algorithm.

Three Real-World Sample Rate Decisions

Example 1: Location Sound Recording for a Feature Film

A production sound mixer on a 15-day feature. The sound recorder is a Sound Devices 888. The DP has asked whether recording at 96kHz for the location audio would "future-proof" the production.

Analysis: The deliverables spec for the feature includes: a 48kHz / 24-bit broadcast WAV for broadcast delivery, a 48kHz / 24-bit mix for the re-recording mixer, and a 48kHz DCP audio track for theatrical. No deliverable requires or benefits from 96kHz location audio.

Storage implication: The production records approximately 600 GB of 48kHz location audio over the shoot. At 96kHz, that doubles to 1.2 TB -- adding approximately $300-400 in additional hard drive costs and increasing the daily DIT offload time by roughly 30 minutes.

Decision: 48kHz at 24-bit. The location sound mixer noted that 24-bit depth (vs. 16-bit) provides far more practical benefit than a higher sample rate -- the additional dynamic range of 24-bit (144 dB theoretical vs. 96 dB for 16-bit) prevents clipping on unexpected loud events without requiring the mixer to reduce gain conservatively. Sample rate increase adds storage cost; bit depth increase adds dynamic range insurance at no storage cost beyond the 50% increase from 16-bit to 24-bit that is standard on all professional productions.

Example 2: Original Film Score Recording with Live Orchestra

A composer recording a 40-piece chamber orchestra for a mid-budget feature score. The scoring stage is Studio A at a major facility with NEVE 8078 console and a Studer A827 analog tape machine available for reference. Budget for the recording sessions: $18,000 over three days.

Analysis: Live acoustic instruments -- particularly strings and high transients from brass and woodwinds -- contain significant content between 20kHz and 40kHz. Some listening tests with orchestral recordings suggest that 96kHz captures a slightly more "open" spatial quality in the high frequency range, attributed to the less aggressive anti-aliasing filter. This is a plausible benefit, not a guaranteed one.

The real-world tradeoff: Recording 40 microphone channels at 96kHz on a three-day session generates approximately 1.8 TB of audio data. At 48kHz, the same session generates 900 GB. The additional storage cost is approximately $200. The additional Pro Tools session processing overhead at 96kHz requires either a larger buffer (adding monitoring latency for the musicians) or a more powerful host machine.

Decision: 96kHz / 24-bit for the orchestra recording sessions, delivered to the composer as 96kHz stems. The composer mixes the score session at 96kHz, exports 96kHz stems to the re-recording mixer, who sample-rate converts to 48kHz on import. The 96kHz sessions preserve maximum quality through the score's own mixing and processing stages. The final film delivery is 48kHz regardless. Cost: approximately $400 extra in storage; benefit: marginally better quality through multiple processing generations in the score mixing stage.

Example 3: Documentary Field Recording with Wildlife Audio

A documentary crew recording wildlife audio in the field -- birdsong, insect sounds, atmospheric ambience -- for a natural history series. The recordist uses a Zoom F8n Pro recorder.

Analysis: Birds and insects produce significant ultrasonic content above 20kHz. Several species of bat and some insects communicate primarily in the 20-80kHz range -- completely inaudible on a 48kHz recording and fully captured on a 96kHz recording (up to 48kHz) or 192kHz recording (up to 96kHz). For standard documentary delivery, this content has no value -- it's inaudible to humans and will be discarded in post. However, if the series is co-producing with a natural history broadcaster that specifically requests ultrasonic recordings for scientific use (BBC Natural History Unit commissions occasionally include this), 96kHz recording has genuine, non-placebo value.

Decision: 48kHz for standard delivery; 96kHz for sequences specifically requested to include ultrasonic content for scientific documentation. A dual-track workflow using the Zoom F8n's ability to simultaneously record at two sample rates (if supported) eliminates the need to decide per-setup.

When to Use Each Sample Rate: A Decision Framework

Step 1: Identify all delivery targets and their specified sample rates. Broadcast and streaming delivery: 48kHz. CD delivery: 44.1kHz. Hi-res streaming (Tidal Masters, Apple Lossless): 96kHz or 192kHz. Most film deliverables are 48kHz. If any deliverable requires 48kHz, your final output is 48kHz regardless of acquisition sample rate.

Step 2: Assess whether your signal chain involves multiple generations of processing. If audio will pass through more than 3-4 generations of processing (recording, mixing, stem mixing, final mix, loudness correction, delivery encode), recording at 96kHz means each generation operates at a higher resolution before the final 48kHz conversion. The benefit is real but subtle -- audible only in high-resolution monitoring chains and primarily in the high-frequency range above 14kHz.

Step 3: Check whether your plug-ins benefit from 96kHz operation. Certain plug-ins are specifically designed to operate at 96kHz or higher: advanced saturation/harmonic distortion processors, physical modeling synthesizers, high-quality pitch shifters. If your processing chain relies heavily on these tools, 96kHz sessions preserve their performance. Most EQs, compressors, and reverbs produce effectively identical results at 48kHz and 96kHz.

Step 4: Calculate the storage impact using the [Audio Bitrate Storage Calculator](/tools/audio-bitrate-storage). Enter your project's duration, channel count, bit depth, and both sample rate options. The calculator returns file sizes for both scenarios and the difference in GB. Confirm that your storage budget, backup plan, and drive write speeds can accommodate the higher sample rate before committing.

Step 5: Confirm your editing workstation can handle the CPU load at 96kHz with your planned plug-in load. Open your standard session template, switch the session to 96kHz, and run an audio performance test. If buffer underruns appear at your standard buffer size (typically 256 or 512 samples), you'll need either a larger buffer (adding monitoring latency) or hardware upgrades. For a mix session without live monitoring, a larger buffer is acceptable. For a recording session with performers monitoring, it's not.

Step 6: Set a consistent sample rate for all devices recording simultaneously. If two recorders or a recorder and a camera are capturing audio for the same scene, they must both operate at the same sample rate. Mixing 48kHz and 96kHz audio in the same NLE session causes the DAW to sample-rate convert one or the other in real time, which adds CPU load and can introduce subtle quality penalties if the conversion algorithm is low quality. The Sample Rate Converter shows the storage and quality implications of any sample rate conversion in your pipeline.

Pro Tips and Common Mistakes

Pro Tip: When recording at 96kHz and delivering at 48kHz, perform the sample rate conversion at the highest-quality SRC setting in your DAW before the final mix session begins -- not at export. High-quality SRC algorithms (iZotope 64-bit SRC, R8brain, Reaper's resampler at "Extreme" quality) produce slightly better results than the on-the-fly conversion that happens when a 96kHz file is imported into a 48kHz session. Convert your stems to 48kHz once at the start of the final mix, verify the conversion quality, and work in a clean 48kHz session.

Pro Tip: 44.1kHz audio should never be used in a video production context. When 44.1kHz audio is mixed with 48kHz video in an NLE, the NLE performs a sample rate conversion that -- at low-quality settings -- introduces artifacts. More problematically, some NLEs perform this conversion poorly or inconsistently, creating drift over long clips. 48kHz is the correct sample rate for all video audio, including music licensed from services that deliver at 44.1kHz. Convert licensed music to 48kHz before bringing it into a video project.

Pro Tip: For podcast and narration recording destined for YouTube or streaming platforms, 48kHz / 16-bit is technically sufficient and dramatically reduces file sizes compared to 48kHz / 24-bit. The 16-bit noise floor (-96 dB) exceeds the practical dynamic range of any narration recording made in a room that isn't a controlled anechoic chamber -- the room noise floor is always above -96 dBFS. Save 24-bit for production recordings where unexpected loud sounds (unexpected music playback, accidental shout, equipment noise) could push the signal close to clipping.

Common Mistake: Setting a film score session to 96kHz because "hi-res is better" without accounting for the sample rate mismatch in the final mix. A 96kHz score delivered to a 48kHz final mix session will be sample-rate converted automatically. The conversion is functionally transparent in any modern DAW with a quality SRC algorithm. The composer who recorded at 96kHz has gained nothing in the final delivery and has paid for it in storage costs and CPU overhead throughout the scoring process.

Common Mistake: Recording location audio at 44.1kHz on a camera or field recorder that defaults to it. Several consumer and prosumer cameras, Zoom recorders, and smartphone recording apps default to 44.1kHz. When this footage is imported into a 48kHz NLE session, the NLE sample-rate converts it automatically. Most modern NLEs handle this cleanly, but the mismatch is a potential source of subtle quality degradation and can introduce drift in some offline editing applications. Check your recorder's default sample rate before the shoot and set it to 48kHz explicitly.

Frequently Asked Questions

Can I hear the difference between 48kHz and 96kHz audio?

Under controlled listening conditions in a high-quality monitoring environment, some listeners report marginal differences in the extreme high-frequency range (above 14kHz). Under normal production monitoring conditions -- studio monitors in a treated room -- reliably identifying 48kHz vs. 96kHz audio in a blind test is difficult and inconsistently achieved even by experienced audio engineers. The consensus in the AES literature is that the perceptible benefit of 96kHz over 48kHz for standard music and dialogue content is minimal to non-existent at the delivery stage. The clearest benefit is in processing: within a production session where audio undergoes multiple processing operations, 96kHz provides headroom that reduces artifact accumulation.

Why does broadcast require 48kHz and not 44.1kHz?

The 48kHz standard for professional video audio was established in the 1980s when the EBU and SMPTE standardized digital audio for broadcast television. The choice of 48kHz (rather than 44.1kHz, which was adopted by the CD format) was partly technical -- the mathematical relationship between 48kHz and common video frame rates makes sample rate conversion between audio-only and audio-with-video formats simpler -- and partly political, establishing a separate professional standard from the consumer CD format. The EBU Tech 3250 recommendation and SMPTE ST 302 both specify 48kHz as the mandatory sample rate for broadcast audio. This standard has been consistent since the 1980s and remains unchanged for all broadcast, theatrical, and streaming delivery contexts.

What sample rate should I use for music created specifically for a film?

Record and produce the music at 96kHz if the scoring budget and equipment support it. The benefit in the production environment (better plug-in performance, more headroom through processing generations) is real, even if inaudible at delivery. Convert to 48kHz when delivering to the re-recording mixer for the final mix session. The conversion is lossless in quality terms at this stage. Never deliver a 44.1kHz music session to a film post environment -- the sample rate mismatch creates avoidable problems in the mix session.

Does sample rate affect the file size of my location audio backups?

Directly proportionally. A 1-hour, 6-channel, 24-bit recording at 48kHz is approximately 1.24 GB. The same recording at 96kHz is approximately 2.48 GB. The Audio Bitrate Storage Calculator computes exact file sizes for any combination of sample rate, bit depth, channel count, and duration. For a 10-day documentary shoot with 8 hours of daily recording on a 6-channel recorder, the storage difference between 48kHz and 96kHz is approximately 75 GB -- meaningful at scale but not prohibitive on modern storage hardware.

Is there any film delivery format that requires 96kHz audio?

A small number of high-resolution delivery formats specify 96kHz as an option or requirement: Dolby Atmos master sessions for theatrical exhibition can be delivered at 96kHz, and some high-resolution digital cinema masters for premium large format (PLF) theatrical screening include 96kHz audio. Netflix's hi-res audio tier accepts 96kHz stems for Atmos mixing. For standard theatrical DCP delivery, broadcast, and streaming, 48kHz is the universal standard and 96kHz provides no compliance benefit. If a specific platform or contract requires 96kHz, that requirement will appear explicitly in the delivery specification -- no guesswork is needed.

Related Tools

The Sample Rate Converter calculates the quality implications and output parameters of any sample rate conversion -- useful for planning the 96kHz to 48kHz conversion in a score-to-mix pipeline. The Audio Bitrate Storage Calculator shows you exactly how much more storage a 96kHz project requires compared to 48kHz across a full production schedule. For the broader context of how sample rate fits into a complete audio deliverables package, audio delivery standards for film and television covers the full specification set -- LUFS targets, channel configurations, and bit depth alongside sample rate requirements. The post on LUFS, dBFS, and loudness normalization covers the loudness compliance side of audio delivery that sample rate selection doesn't affect but that must be addressed before any mastering deliverable ships.

Conclusion

48kHz is the correct sample rate for all film and television audio delivery. 96kHz has legitimate uses in production and scoring sessions where audio will undergo significant processing before the 48kHz final delivery -- the headroom reduces processing artifacts through multiple generations. 192kHz has genuine value only in scientific recording and very specific archival use cases. For any project where the final deliverable is broadcast, streaming, or theatrical DCP, recording location audio, music, or sound effects above 48kHz adds storage and CPU cost without improving what the audience hears.

This post covers standard stereo and multichannel audio for film delivery. Object-based audio formats (Dolby Atmos, MPEG-H) have their own session requirements that include sample rate specifications within a more complex workflow than standard channel-based delivery.

What's the highest sample rate you've used on a production -- and looking back, did the deliverable justify the additional workflow complexity?