SMPTE ST 2110 Bandwidth Calculator

SMPTE ST 2110 is the broadcast industry's standard suite for carrying professional video, audio and ancillary data as separate, individually routable essence streams over standard IP networks — a deliberate architectural departure from older baseband SDI (Serial Digital Interface) infrastructure, where video, audio and metadata traveled bundled together down a single dedicated coaxial cable per signal. Understanding why ST 2110 handles bandwidth so differently from SDI, and why the specific factors this calculator models (chroma subsampling and bit depth, alongside the more obvious resolution and frame rate) matter so much, is genuinely useful for anyone sizing or specifying broadcast IP infrastructure.

Under SDI, bandwidth per signal was effectively fixed by the interface standard itself — a given SDI generation (3G-SDI, 6G-SDI, 12G-SDI, and so on) supported a specific maximum bit rate, and a facility simply chose the appropriate SDI tier for its required resolution and frame rate, with audio and metadata embedded within that same fixed-bandwidth signal at no separately calculated cost. ST 2110's IP-based approach removes this fixed-tier structure entirely: video essence bandwidth is now a direct, calculable function of the actual pixel data being carried — resolution, frame rate, chroma subsampling and bit depth all directly and proportionally affect the real bandwidth required, and unlike SDI, there's no single fixed 'link speed' that bundles everything together; a facility instead sizes its actual IP network fabric (switches and links, typically 10GbE, 25GbE, 40GbE or 100GbE in modern broadcast facilities) to carry the specific, calculated aggregate bandwidth of however many simultaneous video, audio and ancillary essence flows the facility actually needs to move at once.

Chroma subsampling's effect on bandwidth is one of the more consequential, if less immediately obvious, factors this calculator explicitly models, and understanding the underlying color-sampling concept clarifies exactly why it matters so much. Digital video signals conventionally separate a pixel's brightness information (luma, denoted Y) from its color information (chroma, denoted Cb and Cr) — a separation that exploits the well-established fact that human vision is considerably more sensitive to fine detail in brightness than in color, meaning color information can be sampled at a lower spatial resolution than brightness without a correspondingly proportional loss in perceived image quality. 4:4:4 subsampling samples full-resolution color information for every single pixel (3 total samples per pixel: one luma, two chroma), while 4:2:2 samples chroma at half the horizontal resolution of luma (effectively 2 samples per pixel on average), and 4:2:0 samples chroma at half resolution in both horizontal and vertical directions (effectively 1.5 samples per pixel on average). This is exactly why moving from 4:2:2 (the broadcast-standard default, chosen specifically as a well-established, historically SDI-compatible balance between quality and bandwidth) to full 4:4:4 increases required bandwidth by a full 50% at identical resolution, frame rate and bit depth — a genuinely large cost for the additional full-resolution color fidelity that 4:4:4 provides, and exactly why 4:4:4 is reserved for specific applications (like real-time graphics/chroma-key compositing, where full-resolution color edges genuinely matter) rather than used as the broadcast-wide default.

Bit depth's effect is more straightforward but equally consequential: each additional bit of depth per color sample directly, proportionally increases bandwidth, since bit depth determines how many discrete brightness/color levels each sample can represent — 10-bit video (1,024 levels per sample) requires 25% more bandwidth than 8-bit video (256 levels) at otherwise identical settings, and 12-bit (4,096 levels) requires 50% more than 8-bit, precisely because more bits are needed to represent each individual sample's value. 10-bit has become the broadcast-standard default specifically because it provides meaningfully finer gradation than 8-bit — genuinely reducing visible banding artifacts in smooth color gradients and providing meaningfully more headroom for color grading and post-production work — without the further bandwidth cost that full 12-bit (more commonly reserved for high-end digital cinema and specialized production workflows) would add.

The network overhead this calculator adds on top of the raw pixel payload calculation accounts for the genuinely real cost of RTP (Real-time Transport Protocol) packet headers, UDP/IP headers, and ST 2110-20's own specific packetization scheme (which organizes raw pixel data into a defined sequence of network packets according to the standard's own timing and structure rules) — none of which carry any actual pixel data themselves, but all of which consume real network bandwidth on top of the pure payload. This overhead, combined with the practical need for meaningful additional headroom beyond even the overhead-inclusive calculated minimum (since ST 2110 traffic is notably sensitive to packet loss, jitter, and network congestion in ways that traditional data traffic often tolerates far more gracefully), is exactly why real-world broadcast facilities consistently provision network links to a standard tier meaningfully above the pure calculated requirement — a 1080p59.94 4:2:2 10-bit flow's roughly 2.7 Gbps calculated requirement is routinely carried on infrastructure sized for a full 3G-class allocation, and a UHD flow's roughly 11 Gbps requirement is commonly provisioned on 25G-class links or bonded/aggregated link groups, precisely to maintain the substantial operational margin that reliable, broadcast-grade IP video delivery genuinely requires.