What Is Bit Depth?

Bit depth is a fundamental concept in digital audio that determines the precision with which amplitude values are stored for each sample. In simple terms, bit depth dictates how many distinct amplitude levels can be represented. For every additional bit, the number of possible quantization levels doubles. A 16-bit system offers 65,536 levels (216), a 24-bit system provides 16,777,216 levels, and a 32-bit float system can represent an enormous range, well beyond what human hearing can perceive.

The bit depth directly influences the dynamic range of a recording. Dynamic range is the ratio between the quietest and loudest parts of an audio signal before distortion occurs. Each bit contributes approximately 6 dB of dynamic range. Therefore, a 16-bit recording has a theoretical dynamic range of about 96 dB, while 24-bit delivers around 144 dB. This expanded range is why 24-bit audio is standard in professional recording and mixing environments, as it offers more headroom for capturing peaks without clipping.

It is important not to confuse bit depth with sample rate. Sample rate governs how often samples are taken per second (e.g., 44.1 kHz or 48 kHz), while bit depth determines the resolution of each sample’s amplitude. Both parameters independently affect file size and quality, but bit depth has a more direct impact on the noise floor and dynamic range. For a deeper dive into the mathematics of quantization, see the Wikipedia article on audio bit depth.

How Bit Depth Affects File Size

File size for uncompressed PCM audio (such as WAV or AIFF) is calculated using a straightforward formula: File Size (bytes) = Sample Rate × Bit Depth × Duration (seconds) × Number of Channels ÷ 8. The division by 8 converts bits to bytes. This formula shows that bit depth is a linear multiplier: doubling the bit depth doubles the file size, assuming all other factors remain constant.

Let’s work through a concrete example. A 3-minute stereo recording at 44.1 kHz and 16-bit depth would be: 44,100 × 16 × 180 × 2 ÷ 8 = 31,752,000 bytes, or roughly 30.3 MB. Increase the bit depth to 24-bit while keeping everything else the same: 44,100 × 24 × 180 × 2 ÷ 8 = 47,628,000 bytes, about 45.4 MB. That is a 50% increase in file size for a 50% increase in bit depth. For a 32-bit float recording with the same parameters, the file size jumps to 63.5 MB.

In practice, these differences accumulate quickly. A full album at 24-bit/96 kHz can easily consume several gigabytes. Understanding this relationship is critical when planning storage and when setting up recording sessions with constrained disk space or transfer bandwidth.

Additional Factors That Compound File Size

Beyond bit depth, sample rate and the number of channels also scale file size linearly. A 5.1 surround mix has six channels, so its file size will be three times that of a stereo mix at the same sample rate and bit depth. Similarly, doubling the sample rate from 44.1 kHz to 88.2 kHz doubles the file size because twice as many samples are captured per second.

When using compressed formats like MP3, AAC, or FLAC, the relationship becomes less straightforward. Lossy codecs discard information based on perceptual models, so bit depth still influences the source quality but the final file size is largely determined by the bitrate setting (e.g., 320 kbps). Lossless codecs like FLAC reduce file size by about 30–60% without losing information, but the original bit depth remains encoded. A 24-bit FLAC file will still be larger than its 16-bit version, though the difference is smaller than with uncompressed audio. For a practical overview of storage implications, consult the Sound On Sound article on bit depth.

Comparing Common Bit Depths: 16-bit, 24-bit, and 32-bit Float

16-bit Audio

16-bit is the standard for CD audio and most consumer distribution formats. With a dynamic range of 96 dB, it is sufficient for finished masters intended for streaming and playback. The noise floor is low enough that in typical listening environments, the quantization noise is inaudible. However, during recording and mixing, 16-bit offers very little headroom. Peaks must be carefully managed to avoid clipping, and any headroom left unused results in reduced effective dynamic range. For this reason, 16-bit is almost never used for tracking or mixing in professional studios.

24-bit Audio

24-bit audio has become the de facto standard for recording and mixing. The added 48 dB of dynamic range (144 dB theoretical) provides massive headroom. Engineers can record with average levels well below 0 dBFS, leaving generous margin for unexpected peaks. This reduces the risk of digital clipping and allows for more aggressive processing later. Additionally, the quantization noise floor is far below the noise of even the best analog equipment and microphones. In practical terms, a 24-bit recording captures the full analog signal without adding audible noise from the conversion process.

For editing and mixing, 24-bit also offers more precision when applying gain changes, fades, and effects. When summing multiple tracks, the higher resolution prevents rounding errors from accumulating. Many audio interfaces and digital audio workstations (DAWs) operate internally at 32-bit or 64-bit floating point to further protect against these errors, but capturing at 24-bit remains the industry benchmark.

32-bit Float Audio

32-bit floating point is a special format where amplitude values are stored with a sign, an exponent, and a mantissa. This allows an enormous dynamic range – roughly 1528 dB – which completely eliminates the risk of clipping within the digital domain. 32-bit float is used in some modern recorders and DAWs for intermediate processing, but it is not a distribution format. Files are significantly larger than 24-bit versions, and the extra precision is wasted for final delivery because playback systems are limited to 16-bit or 24-bit conversion. For multi-track recording, some engineers prefer 32-bit float to avoid having to set gain levels precisely, but the trade-off is increased storage and bandwidth requirements.

Bit Depth and Dynamic Range in Detail

The human auditory system can perceive sounds over a range of about 120 dB from the threshold of hearing to the threshold of pain. A 16-bit system with 96 dB of dynamic range theoretically cannot capture the full range of human hearing without some form of compression or limiting. In practice, most music and speech signals fall within a 50–70 dB range when properly mixed, so 16-bit is sufficient for listening. However, the margin for error is minimal.

Quantization noise is the key artifact introduced by limited bit depth. In a 16-bit system, the noise floor sits at approximately −96 dBFS. If the signal is recorded with peaks at −18 dBFS (common practice for analog mixers), the effective dynamic range above the noise floor is only 78 dB. With 24-bit, the noise floor is at −144 dBFS, so even with conservative recording levels, the noise is far below the analog noise floor. This is why 24-bit is considered safe for high-dynamic-range content like classical or acoustic recordings.

Dither is a technique of adding low-level noise to mask the distortion caused by quantization errors when reducing bit depth (e.g., from 24-bit to 16-bit). Proper dither ensures that the final 16-bit master sounds smoother and more natural. Without dither, the truncation creates harmonic distortion that can sound harsh, especially on quiet passages. The iZotope guide to dithering provides an excellent practical explanation of when and how to apply dither.

Practical Considerations for File Size

Storage and Archiving

A professional recording session can easily generate hundreds of tracks, each recorded at 24-bit/48 kHz or higher. A 2-hour multitrack project with 24 tracks can consume over 20 GB. For archival purposes, many studios keep 24-bit source files and create 16-bit or lossy versions for client delivery. Hard drive space is cheap, but transfer times and backup speed matter. When purchasing storage, consider that a 24-bit file system will require roughly 50% more capacity than 16-bit for the same project.

Streaming and Broadcasting

Streaming services universally use lossy compression (AAC, MP3, Ogg Vorbis) at bitrates ranging from 128 kbps to 320 kbps. These codecs take advantage of perceptual masking to discard data that is unlikely to be heard. The source material delivered to streaming platforms is typically 16-bit/44.1 kHz, although some services accept 24-bit masters for high-resolution tiers (e.g., Tidal, Qobuz). When encoding for streaming, using a higher bit depth source does not necessarily improve the final quality at low bitrates; the limiting factor is the codec’s bitrate budget. Nevertheless, a well-mastered 24-bit source provides a richer foundation for the encoding process and can yield marginally better results.

Editing and Processing Performance

Higher bit depth files require more CPU and memory bandwidth during playback and processing. Modern computers handle 24-bit audio easily, but when working with large sessions or heavy plugins, using 16-bit files for editing could improve performance. Some engineers convert multitrack files to 24-bit for mixing and then export stems or final mixes at 24-bit or 32-bit float to preserve quality. The performance impact is usually negligible on current hardware, but it is worth monitoring when a project pushes system limits.

Choosing the Right Bit Depth for Your Workflow

Recording

Always record at 24-bit. The extra dynamic range gives you peace of mind and eliminates the need to hit a precise level. Many recorders default to 24-bit, and audio interfaces with 24-bit converters are ubiquitous. Recording at 16-bit is only advisable if your hardware is limited or if you are capturing final mono dialogue for a specific low-bandwidth application.

Mixing and Mastering

Work in 24-bit or 32-bit float throughout your mixing chain. Most DAWs operate internally at 32-bit float, so your source files should be at least 24-bit to preserve headroom and prevent truncation errors. When bouncing mixdowns, use 24-bit or 32-bit float interleaved. For mastering, deliver 24-bit masters to clients or platforms that request them, and create 16-bit dither‑down versions for CD or standard streaming.

Distribution

For physical media like CD, you must create a 16-bit master (with dither) at 44.1 kHz. For digital storefronts and streaming, follow each platform’s specification. Many platforms accept 24-bit/48 kHz or 24-bit/44.1 kHz. High-resolution audio services may request 24-bit/96 kHz or 24-bit/192 kHz. In all cases, ensure proper dither when reducing bit depth to 16-bit. For lossy encoding, the codec will handle downconversion, but starting from a 24-bit master yields slightly better results.

Conclusion

Bit depth is a foundational parameter in digital audio that directly controls the dynamic range, noise floor, and file size of your recordings. A higher bit depth offers more resolution and headroom at the cost of larger files. For professional work, 24-bit is the minimum standard for acquisition and processing, while 16-bit remains the distribution end point for most consumer formats. Understanding the calculation of file sizes helps you plan storage and bandwidth, and knowing the trade-offs between bit depths allows you to make informed decisions at every stage of production. Whether you are recording a quiet acoustic performance or a loud rock band, the combination of sample rate and bit depth must be chosen with both quality and practicality in mind.

For further technical details on digital audio quantization and dynamic range, refer to the Audio Engineering Society standards or the classic textbook Handbook for Sound Engineers. Practical implementation guidelines are also available from the ITU‑R BS.1770 loudness standards document, which discusses dynamic range in broadcast contexts. For a historical perspective on bit depth evolution, the AES E‑Library hosts numerous papers on the subject.