The length is never the length
The first mistake in a cable run is measuring the straight line. A cable from a rack in the corner to a display on the opposite wall does not fly across the room. It leaves the back of the rack, drops behind it, runs to a wall plate, goes up inside the stud bay, crosses a ceiling along a joist, comes back down the other side and emerges behind the panel. Every one of those turns adds length, and every one of them adds bend radius that the tape measure on the floor knew nothing about.
Add slack to what you measured, add a service loop at each end so the rack can be pulled out without unplugging everything, and round up. Cable is inexpensive relative to the cost of pulling a run twice, and a run that arrives eight inches short is not a run at all. The one thing not to do is coil a large excess tightly behind the rack: a tight coil of any cable is an inductor, and on a long mains-adjacent run it is also a good antenna. Figure-eight the excess loosely instead.
Why HDMI length depends on what you are sending
An HDMI cable is copper carrying very fast electrical transitions, and copper attenuates high frequencies more than low ones. That is the entire story. A signal at four and a half gigabits per second survives further down the same cable than one at forty, because the higher rate has less margin before the receiver can no longer distinguish one symbol from the next.
| Format | Approximate bandwidth | Relative to 1080p60 |
|---|---|---|
| 1080p60 8-bit | 4.5 Gbps | 1x |
| 4K30 8-bit | 8.9 Gbps | 2x |
| 4K60 8-bit 4:4:4 | 17.8 Gbps | 4x |
| 4K60 10-bit 4:4:4 | 22.3 Gbps | 5x |
| 4K120 10-bit | about 40 Gbps | 9x |
| 8K60 10-bit | about 80 Gbps | 18x |
Those figures are approximate uncompressed rates and they move depending on colour subsampling, bit depth and whether stream compression is in use, which can substantially reduce the rate on the wire. The point of the table is the ordering, not the third digit. A cable installed years ago that carried everything you owned may simply not carry a new source, and it did not degrade — the demand went up.
This is why the calculator asks you for the passive limit rather than asserting one. Quoted lengths vary with conductor gauge, construction and the certification the cable was tested to, and a number stated as fact on a web page would be wrong for a good proportion of readers. Take it from the documentation for the cable you are actually buying, for the format you are actually sending.
Active, optical, and over twisted pair
When passive copper runs out, there are three common routes onward. An active copper cable includes a small amplifier at the display end, powered from the connector, which reconstructs the signal after the loss. An active optical cable converts the electrical signal to light at the source end, carries it as light, and converts back at the display; the run itself is glass or plastic fibre and the length limits become very generous. Both are directional — the ends are marked, and connecting one backwards produces exactly nothing, silently.
The third route sends the signal over twisted pair cable using an extender pair, which is how long runs in structured buildings are usually done. The length the extender supports is set by the extender, and typically shortens for higher formats in the same way passive copper does. That figure comes from the extender documentation, not from the cabling standard, and the 100 metre structured cabling channel is an upper bound rather than a promise.
Whichever route, pull a spare and leave a pull string. Every installer who has opened a finished wall to add one cable has a strong view about this.
Speaker wire, where the arithmetic is old and settled
Speaker cable is the one AV run where the physics is completely undramatic. The cable has resistance, that resistance is in series with the speaker, and it forms a divider with the speaker impedance. Resistance is two times the run length times the resistance per unit length, because the current goes out on one conductor and back on the other.
| Gauge | Ohms per 1000 ft | Loop R at 50 ft | % of 8 ohm | % of 4 ohm |
|---|---|---|---|---|
| 18 AWG | 6.385 | 0.639 | 8.0% | 16.0% |
| 16 AWG | 4.016 | 0.402 | 5.0% | 10.0% |
| 14 AWG | 2.525 | 0.253 | 3.2% | 6.3% |
| 12 AWG | 1.588 | 0.159 | 2.0% | 4.0% |
| 10 AWG | 0.999 | 0.100 | 1.2% | 2.5% |
A widely used working target is keeping the loop resistance under about 5 percent of the nominal speaker impedance, and under 2 percent if you want the question closed. Notice how much harder a 4 ohm load is than an 8 ohm one: the same cable is twice the fraction. For a 50 foot run to an 8 ohm speaker, 14 gauge is comfortable. For the same run to a 4 ohm speaker, 12 is the sensible choice. None of this is about exotic materials — it is a resistance calculation that has not changed in a century.
Planning the run before the wall closes
The best time to solve a cable problem is before there is drywall in front of it. If a wall is open, run conduit rather than cable, or run cable plus a pull string, and terminate at plates rather than leaving cable hanging out of a hole. Cable inside a cavity generally needs a jacket rated for that use, and which rating applies is a question for the manufacturer documentation and the local authority having jurisdiction rather than for a formula.
Everything downstream of that decision is easier: the rack power and heat calculator tells you what is going to live at one end of the run, the projector throw distance calculator tells you where the other end has to be, and the Wi-Fi improvement guide covers the case where the honest answer is that the signal should never have been on a cable at all. For the mains side of any of it, the wire size calculator and the voltage drop calculator handle the arithmetic, but the work itself belongs to a licensed electrician.
Questions people ask
How long can an HDMI cable be before it stops working?
There is no single answer, and anyone who gives you one without asking what format you are sending is guessing. The workable passive copper length falls as the data rate rises, so the same cable that carries 1080p60 across a long room may fail at 4K120 over half that distance, because 4K120 is roughly nine times the data rate. It also depends on the cable itself — conductor gauge and construction vary enormously between products that look identical. The reliable approach is to take the length figure from the documentation for the specific cable at the specific format, treat that as an upper bound rather than a target, and move to an active or optical cable well before you reach it if the run is going somewhere you cannot easily get back to.
What does it look like when an HDMI cable is too long?
Not like an analogue signal fading. Digital links tend to work or not work, and the failure modes are distinctive: sparkles or glittering pixels across the image, brief black frames that come and go, a handshake that takes ten or twenty seconds every time the source changes, audio dropping out while the picture stays, or the display simply reporting no signal at a high format while working at a lower one. That last symptom is the useful diagnostic — if dropping the resolution or refresh rate makes the problem disappear, the run is the marginal element. Because it is intermittent, the fault frequently gets blamed on the source, the display, or a firmware update instead.
Do I need in-wall rated cable?
Cable installed inside a wall, ceiling or floor cavity generally has to carry a rating appropriate to that use, and that rating is about how the jacket behaves in a fire rather than anything to do with signal quality. What it does not do is make the picture better. Which specific rating applies to your cavity depends on the construction and on local rules, and it is settled by the cable manufacturer documentation together with your local authority having jurisdiction — not by a calculator and not by an internet consensus. Ask before you pull, because the cost difference is small and the cost of pulling it out again is not.
Does thicker speaker wire actually sound better?
It reduces series resistance, and below a certain point that stops being audible. The mechanism is straightforward: cable resistance sits in series with the speaker, forming a divider that loses a small amount of level and, more importantly, reduces how tightly the amplifier can control the driver. Once the loop resistance is under about 2 percent of the speaker impedance, further reduction is not something you can hear. What that means in practice is that gauge matters for long runs and for low impedance speakers, and matters very little for a 10 foot run to an 8 ohm speaker. Spending more per foot beyond that buys jacket durability, better connectors and easier handling, which are worth having for their own reasons.
Can I run speaker cable and mains cable in the same hole?
Keeping them separated is the right habit, and where a route forces them together they should cross at right angles rather than run alongside each other, because parallel runs couple much more than crossing ones. Beyond the interference question there is a separation requirement in most electrical rules between mains conductors and low voltage cabling, and the details vary by jurisdiction and by how each is enclosed. That is a question for the electrician doing the mains side, and it is one of the reasons the low voltage path is worth planning at the same time as the mains rather than afterwards. Speaker wire is the least sensitive of the AV cables to induced noise because it is a low impedance circuit, but line level interconnects are among the most sensitive.