Enough Speed, Still Pixelating: Wi-Fi, Ethernet and the Number That Actually Matters

Why your speed test does not predict IPTV

A speed test measures peak throughput to a nearby server for a few seconds. Live IPTV needs something different: a modest but completely uninterrupted rate, for hours. Those are not the same requirement, and one does not prove the other.

What breaks playback is jitter — packets arriving unevenly — and packet loss. Both can be bad enough to stall a stream while leaving your average speed untouched, because a test that runs for a few seconds simply does not see them. That is how "but I have 300 Mbps" and "it keeps buffering" end up being true at the same time.

If you have not yet run the two phone runs in the buffering guide, do that first. This guide is for people who already know the fault is inside their house.

Bitrate versus resolution

This is the fact that changes how you read everything else.

Resolution is the pixel count — 1080p, 2160p. Bitrate is data per second, and bitrate is the only one your connection has to carry. The same 1080p channel can arrive at 3 Mbps or at 12 Mbps depending on how it was encoded, and those two look very different on screen.

So "4K" in a channel name is a resolution label and nothing more. It tells you nothing about bitrate or source quality, and an upscaled 4K stream at a low bitrate looks worse than a well-encoded 1080p one. Judging a service by how many channels have 4K in the name is judging the label, not the picture.

Codec matters as much. H.265/HEVC reaches comparable quality at roughly half the bitrate of H.264 — but 2160p carries four times the pixels of 1080p, so halving it still leaves you above 1080p, not level with it. A 4K HEVC stream costs roughly two to three times a 1080p H.264 one rather than four times. The ranges below say the same thing in numbers.

Typical live ranges, as ranges: SD around 1–3 Mbps, 1080p H.264 around 4–10 Mbps, 4K HEVC around 15–25 Mbps. For an outside yardstick, Netflix publishes its own recommended connection speeds — 3 Mbps for 720p HD, 5 Mbps for 1080p, and 15 Mbps for 4K. Rule of thumb for live IPTV: aim for roughly double the stream's bitrate in stable headroom, because live has no buffer to fall back on.

Find the bitrate you are actually receiving

In VLC, open Tools > Media Information > Statistics while a channel is playing and read the input bitrate. The Codec tab on the same window gives you the resolution and the codec. Several TV players have a stream-info overlay that shows the same thing.

Do this once and "it feels slow" becomes a number you can compare against your link. A 4K stream at 20 Mbps arriving over a Wi-Fi path that manages 25 Mbps on a good day is not a mystery — it is arithmetic.

2.4 GHz, 5 GHz, and why full bars mean nothing

2.4 GHz reaches further and passes through walls better. 5 GHz carries far more data and is much less crowded, over a shorter range.

The problem with 2.4 GHz is not strength, it is space. The band is only wide enough for a handful of non-overlapping 20 MHz channels, and how many depends on where you live: 1, 6 and 11 in North America, and 1, 5, 9 and 13 in regions that allow channel 13, which includes the UK and most of Europe. Whichever set applies, it is shared with microwave ovens, Bluetooth, wireless doorbells and cheap baby monitors. In a block of flats, every neighbour is in those same few channels. The band can be effectively unusable in the evening while your TV still shows four bars, because bars measure signal strength, not free airtime.

Which band is your TV actually on?

If your router broadcasts one merged SSID for both bands, band steering decides for you — and it can park a TV on 2.4 GHz and leave it there for days, including through the evening when that band is worst.

Check what the TV joined in the device's own network info screen, or in the router's client list, which usually shows the band per client. Then pin it: give the two bands separate SSID names (for example Home and Home-5G) and join the 5 GHz one explicitly, or set a per-device band preference if your router has one.

Placement and interference

Wi-Fi is stopped by things, not by distance alone. A box sitting on the floor, inside a closed cabinet, or tucked behind a wall-mounted screen is being shielded. Metal shelving, mirrors and foil-backed insulation are worse than plasterboard. Two walls at four metres is harder than open air at fifteen.

Test it properly: take the phone to exactly where the TV box sits, hold it there, and stream. If the phone struggles in that spot too, it is the spot, not the box.

Ethernet

One cable beats every tweak in this article. It removes interference, band choice, placement and neighbours in a single step.

The detail that stops people unnecessarily: plenty of TV boxes only have a 100 Mbps port. In practice that carries about 94 Mbps of real traffic — three to four times the heaviest 4K stream in the ranges above, and around ten times a typical 1080p one. "Only Fast Ethernet" is not a reason to stay on Wi-Fi. Cat5e is enough for anything you will do here.

If a cable is genuinely impossible, powerline adapters or MoCA are the compromise. Be honest about powerline: it depends entirely on your building's wiring, it hates extension leads and separate circuits, and it can easily come out worse than 5 GHz. Test it, keep the receipt.

Mesh and repeaters

A single-radio repeater roughly halves throughput, because it receives and retransmits on the same radio. A mesh node with a wired or dedicated backhaul does not. That difference is why one person swears by mesh and another says it made everything worse.

Place a node roughly halfway to the dead spot, not in it. A node that can barely hear the router has nothing good to pass on.

Router settings: real, and placebo

Picking a clear channel and a sane channel width — real. On 2.4 GHz, choose one of the non-overlapping channels for your region — 1, 6 or 11, or 1, 5, 9 or 13 where channel 13 is allowed — and stay at 20 MHz. On 5 GHz, a wider channel helps if the band is quiet.
QoS — partly real. It helps when a busy upload is starving everything else. It does nothing about a saturated Wi-Fi band, because the congestion is in the air, not in the router's queue.
Turning off "smart" client steering — sometimes real. If your TV is being re-steered onto 2.4 GHz mid-evening, disabling that or splitting the SSIDs fixes it.
Rebooting the router nightly, changing MTU, generic "optimiser" apps — placebo, unless you have a specific reason and a measurement to match.

Verify the same way every time

Same channel, same two minutes, same time of evening, one change between runs. A fix you cannot measure is a fix you will be making again next week. If the stall survives every change here, go back to the buffering guide and re-run the two phone tests — the answer may have moved.