Byte / Bit Converter

Eight, by modern convention. The byte was not always eight bits - early computers used 6, 7 or 9 bit "bytes". The 8-bit byte became standard with the IBM System/360 in 1964, was ratified in ISO/IEC 2382-1, and is now formalized as the "octet" in networking standards where the term "byte" is sometimes ambiguous. Every modern CPU, file format and network protocol assumes 8 bits per byte.

KB, in the SI sense, equals exactly 1000 bytes. KiB, the IEC binary prefix kibibyte, equals exactly 1024 bytes. They differ by about 2.4%. The distinction was codified in IEC 60027-2 in 1998 to eliminate the ambiguity of "KB" meaning either 1000 or 1024 depending on context. Similar pairs exist for all larger units: MB / MiB, GB / GiB, TB / TiB, PB / PiB.

No. The conversion runs inside a Preact component using plain JavaScript arithmetic and - at very high magnitudes - BigInt for precision. There is no fetch call, no websocket, no telemetry on your input values. PostHog logs the pageview but not the numbers you type. You can turn off networking after the page loads and the converter keeps working.

The electrical signaling on a wire or radio link transfers one bit at a time. Measuring in bits per second matches the physical reality of how fast symbols can be modulated onto the medium. A 1 Gbps Ethernet link moves 10^9 bits per second = 125,000,000 bytes per second = 125 MB/s. Dividing by 8 is the standard first step when comparing ISP speeds against file-download predictions.

32 bytes exactly. AES-256 uses a 256-bit key (256 / 8 = 32 bytes). AES-128 uses 16 bytes; AES-192 uses 24 bytes. An AES block is always 128 bits (16 bytes) regardless of key size. The initialization vector for AES-GCM is 12 bytes, and the authentication tag is usually 16 bytes. All those numbers are specified in FIPS 197 and NIST SP 800-38D.

JavaScript's Number type is an IEEE 754 double-precision float, which uses a 52-bit mantissa plus an implicit leading bit. That gives exact integer representation up to 2^53 - 1 (Number.MAX_SAFE_INTEGER = 9,007,199,254,740,991). Above that, consecutive integers start aliasing - 2^53 and 2^53 + 1 both map to the same bit pattern. For byte counts in the exabyte range, use BigInt or string-based arithmetic. The converter switches to BigInt internally past the safe range.

Half a byte - four bits. A single hexadecimal digit (0-F) holds exactly one nibble of information, which is why hex dumps pair up nicely with byte layouts. The term is informal rather than standards-track and rarely appears in modern technical writing; you still see it in retro-computing and low-level networking contexts. This converter does not list nibbles as a unit, but you can enter "0.5 B" if you ever need it.

Depends on the format. 8-bit grayscale uses 1 byte per pixel. 24-bit RGB (the usual sRGB image) uses 3 bytes. 32-bit RGBA adds an alpha channel for 4 bytes. HDR and scientific imaging use 10, 12 or 16 bits per channel, giving 30, 36 or 48 bits per pixel. A 4K (3840 x 2160) 24-bit frame is 3840 * 2160 * 3 = 24,883,200 bytes = 23.7 MiB uncompressed.

Exabyte (EB / EiB = 10^18 or 2^60 bytes), zettabyte (ZB / ZiB = 10^21 or 2^70), yottabyte (YB / YiB = 10^24 or 2^80). In 2022 the General Conference on Weights and Measures added ronna (10^27) and quetta (10^30) as the new SI super-large prefixes, with ronto and quecto as their submultiples. Global IP traffic crossed the zettabyte-per-year mark around 2016 according to Cisco VNI reports. This tool stops at petabyte because consumer-scale use cases rarely exceed it.

File size in bits, divided by link speed in bits per second, gives time in seconds. A 5 GB file (in binary, 5 * 2^30 bytes = 42,949,672,960 bits) over a 100 Mbps link takes 42,949,672,960 / 100,000,000 = 429.5 seconds, or about 7 minutes 10 seconds in theory. Real downloads are 10-30% slower because of TCP slow-start, TLS handshakes and server-side throttling. Use this converter to move between units before plugging into the division.

It covers both. Each row in the grid is one of b (bits), B (bytes), KB, MB, GB, TB, and PB. Type a value into any row and the rest of the grid - including the b-to-KB-to-MB-to-GB ladder and the bit / byte split - recomputes from a single internal bit total. The base toggle then re-renders every cell as binary (1024-based) or decimal (1000-based), which is the part most other "byte converter" or "bytes calculator converter" pages skip.

Splitting the conversion across narrow tools (KB to MB, then MB to GB) costs more keystrokes and risks rounding error at each step. Holding every unit in one row, recomputed from a shared bit total, eliminates the rounding chain and lets you read the answer for any unit instantly. The same panel doubles as the data size converter and the bit-rate sanity-check tool, which is why the page is structured as a multi-row grid rather than a single input-output pair.