World Icon

The film has just started and your TV may be sampling the screen and matching it against a database. To turn off ACR you hunt through a privacy menu where each brand uses a different name, and whether it goes quiet depends on whose TV it is.

Should I turn off ACR on my smart TV?

Yes, for most households: automatic content recognition (ACR) samples what your screen shows and matches it against a reference library, brands use the result for advertising, and the opt-out costs a few minutes in the menus. Vizio's Viewing Data page says its OS has built-in ACR and licenses the result for targeted advertising (undated page). Standby draw, priced in our TV electricity cost piece, is another thing a TV does unprompted.

A 2024 Internet Measurement Conference paper from a UCL-led team read vendor documents as a capture every 10 ms on LG and every 500 ms on Samsung: stated intent, not measurement. The authors infer both brands batch samples into periodic uploads, so a quiet network graph proves little, and the volume is modest beside video (see 4K streaming data usage against a home data cap).

Four figures decide most of this: upload rhythm and volume from that paper, the platform count from Northeastern's 2026 study (5,943 experiments on 12 TVs, so one flaky set cannot explain a pattern), and the penalty from TechSpot's report of Texas's December 2025 suits.

Batching is why ACR is hard to catch from the sofa. Dividing upload gap by capture gap gives roughly 1,500 sampled frames per LG upload and about 120 per Samsung upload, our arithmetic, not a measurement. Check the setting; do not trust the router.

Where is the ACR setting on a Samsung, LG or Vizio TV?

It sits in a privacy or system menu under a different name on each brand: Samsung's Viewing Information Services, LG's Live Plus, Vizio's Viewing Data, Sony's Samba Interactive TV, and Roku's Use info from TV inputs. Paths are as reported by Malwarebytes (2 March 2026) and Tom's Guide (30 January 2026), Vizio's from its own page; names shift by model year, so use settings search. Hisense and TCL sets use their platform's path (Roku OS or Google TV).

• Samsung: Settings > Support > Terms & Privacy > Privacy Choices, Viewing Information Services off.
• LG: Settings > All Settings > General > System > Additional Settings, Live Plus off.
• Vizio: All Settings > Privacy & Legal > Viewing Data (older sets: Admin & Privacy).
• Sony: Settings > Privacy (or System), Samba Interactive TV off.
• Roku: Settings > Privacy > Smart TV Experience > Use info from TV inputs off.

Finding the switch is half the job; this side by side of two test programmes is the other.

Read the middle column as evidence about the sets tested, not a guarantee about yours: most ACR traffic is encrypted, so researchers infer from timing and volume.

This settles which sources still send traffic from a Samsung with the switch off: mostly antenna and HDMI, little on free channels, none on Netflix or YouTube. Percentages divide Northeastern's run counts (41 of 60, 18 of 30, 6 of 30).

Worked example: one LG TV, three hours a day of antenna or free-channel viewing, every day (our assumptions). At LG's documented 10 ms capture that is 3 × 3,600 × 100 = 1,080,000 captures a day, about 394 million a year. The upload: 3 h × 4.58 MB × 365 days is 5.0 GB a year. Samsung's documented 500 ms and 0.45 MB an hour give about 8 million captures and 0.5 GB. Swap in your own hours below; estimates, not measurements.

How much does your TV sample and send?Set your evening hours, your TV count and your brand. Upload uses the paper's measured rates; captures use each brand's documented rate.

Your numbers

Antenna or free-channel hours

3.0 h

0.5 h8.0 hTVs in the house

1

13

TV brand

LGSamsung

LG-style upload a year

5.0 GB

Samsung-style upload a year

0.5 GB

Captures a year, your brand

394 million

Upload is 5.0 GB on LG-style or 0.5 GB on Samsung-style: too small for a bill to notice. The 394 million captures behind it are the part to weigh.Reset to the article's exampleSliders need JavaScript. Figures shown are the article's example.Upload: IMC 2024 US antenna readings, LG 4.58 MB an hour, Samsung 0.45 MB (ACR domains summed). Capture rates are vendor documents as read by IMC 2024. Estimates as of November 2026.

Does turning off ACR actually stop the tracking?

On some brands, yes: two independent test programmes saw LG's traffic stop, while Samsung and Vizio kept sending after opt-out, and the studies cannot prove from encrypted traffic what was stored or used.

Here is the dated before and after. In 2024 the paper saw complete silence to every identified ACR domain once Samsung or LG was opted out; in 2026 Northeastern saw Samsung traffic consistent with ACR still flowing. Different sets and methods, so no proof of regression.

Texas's December 2025 suits against five TV makers, per TechSpot, call Samsung's opt-out four or more menus and over 15 clicks against one click to opt in. The National Law Review reported on 11 May 2026 that LG agreed to a clear opt-out after Samsung's earlier deal; Sony, Hisense and TCL were pending. Allegations and settlements, not findings; none says a TV in another state got the new prompt.

Our call: on a Samsung or Vizio, move viewing to a different box, as Northeastern advises; our piece on smart TV software updates versus a streaming stick covers that trade. The old advice to flip the switch and relax is out of date for two big names, and a stick alone is not enough: the TV itself must stay offline.

• Vizio's page says switching Viewing Data off deletes that TV's historical logs; other data, such as Activity Data, may still be collected.
• Even with ACR off, researchers saw Vizio's WatchFree+ expose program titles in unencrypted lookups; no switch covered it.

The decision is per brand: on LG, Sony and Fire TV the switch is enough, on Samsung and Vizio it is not. This week, switch off the setting above, then take a Samsung or Vizio off your Wi-Fi and watch through a stick. That is wrong if you need the TV's own apps; check first that the switch reads off. While you are in settings, motion smoothing for movies versus sports and games is the next call.

The match looked sharp, the ball easy to follow. Then the film began, and a candlelit costume drama suddenly looked like a daytime soap. Same TV, same setting. That setting is motion smoothing, and whether it is a friend or a villain depends on what is on screen.

Why does motion smoothing wreck films but sometimes help sport?

Motion smoothing invents extra frames between real ones, which is why it can make a 24 fps film look like a live broadcast while making fast live action easier to follow. RTINGS, whose motion-interpolation page was last modified in March 2026, calls it an optional feature TVs use to increase the frame rate of content. A guess between two real pictures helps when a camera whips across a pitch, and hurts when a director chose 24 frames a second on purpose.

Samsung's support page (dated July 2025 in its metadata) says smoothing suits live content such as sports, which it says is shot at 30 fps, while 24 fps movies are not meant for it and can pick up the soap opera effect. We trust the film half. The sport half is a vendor's preference worth testing.

Filmmaker Mode is the usual shortcut. Audioholics reported the UHD Alliance programme in November 2019 as switching off post-processing, including smoothing and noise reduction, and keeping the filmmaker's frame rate; Sony said it sets its own standards. That was 2019, so check your set. If it looks dim, see Filmmaker Mode looking too dark, and note that picture mode also sets what the set draws from the wall.

Four numbers decide how smoothing behaves in your room; one rests on an assumed delay you should replace with your own.

The ratio is the one a spec sheet never shows you. A 60 Hz panel cannot divide a film's frames evenly, so it alternates long and short holds, and that alternation, not smoothing, is the stutter on slow pans.

Which setting fits which screen time?

The right setting depends on what is on screen and whether your thumbs are on a controller, so the rule is per content type, not per TV. Try this sum: divide your panel's refresh rate by 24, the frame rate of film. A whole answer (120 Hz gives 5) means even holds; a trailing .5 means alternating long and short ones.

Worked example: a 60 Hz TV showing a 24 fps film, using the 3:2 pull-down on LG's support page (June 2024). One refresh lasts 1000 / 60 = 16.7 ms. Frames held three refreshes stay up 50 ms, those held two stay up 33.3 ms, and an even rhythm would give 41.7 ms (1000 / 24). That 16.7 ms gap between neighbours is the stutter, and smoothing off does not remove it. For a console, assume 50 ms of added delay at 60 fps (our placeholder, not a measurement): 50 x 60 / 1000 = 3.0 frames late. By our reasoning the delay cannot fall below one source frame, about 16.7 ms at 60 fps, since the TV needs the next frame before drawing the one between.

Cadence and lag on your TVPick your panel, the content and your game's frame rate, then set the lag to your own TV's number.

Your numbers

Panel refresh

60 Hz120 Hz144 Hz

Content frame rate

24 fps film25 fps30 fps60 fps

Added processing lag, ms

50

0100

Game frame rate

30 fps60 fps120 fps

Even hold per frame

41.7 ms

Longest hold

50.0 ms

Shortest hold

33.3 ms

Lag in frames

3.0 frames

Frames alternate between 50.0 ms and 33.3 ms, the uneven rhythm smoothing tries to hide. At 3.0 frames of lag, keep it off for games.Reset to the article's exampleSliders need JavaScript. Figures shown are the article's example.Hold times use refresh divided by frame rate and the 3:2 pull-down in LG's support page (June 2024). The 50 ms lag is our placeholder, not a lab figure.

The table turns the arithmetic and the vendor pages into five situations, each action with its reason beside it.

Row three's lag is a placeholder until you set the calculator to your own set. Whether cleaner pans are worth a faintly artificial ball is taste, and no support page settles it.

What goes wrong when you change the setting?

Switching smoothing off is not a cure-all: it leaves the 60 Hz stutter in place, an update can undo it, and every brand hides and names the control differently.

The update risk is documented. In June 2024, TechRadar reported that Roku OS 13 had turned on Roku's Action Smoothing for some users, including owners of TCL sets, and that some could not switch it off. One poster on Roku's community forum wrote: "We need the option to roll back to the software version before the Smart Picture update." That is one undated forum voice, not a statistic. A stick is one answer, weighed in smart TV updates versus a streaming stick.

Names are the other trap. Roku's support page, updated 24 June 2026, advises Low or Off for movies and warns that High can cause odd motion or ghosting, so a set shipped on High needs lowering, not returning. The control sits in Expert settings and may be named differently on partner TVs.

As of November 2026, the vendor pages put the control here:

• Samsung: Settings, Picture, Expert Settings, Auto Motion Plus Settings; Auto, Off or Custom (page dated July 2025).
• LG: Picture Options, TruMotion; Smooth, Clear, Clear Plus or User; options differ by picture mode (page dated June 2024).
• Sony: Motionflow, Off or Custom; the path varies by software version and not every model adjusts it (page modified June 2026).
• Roku: Star button, Picture settings, Customize Picture, Expert settings, Action Smoothing (page updated June 2026).

The decision is per content type, not per TV: films and games off, live sport on the lowest setting for a trial. This week, watch the same ten minutes of a match twice, once on each setting, and keep whichever the room prefers.

Fourth night of a new series in 4K, lights off, and a phone on the sofa arm lights up: Xfinity says the house is near its monthly data limit. Nobody downloaded anything unusual. The TV did it, one evening at a time, and that is the side of 4K streaming data usage no spec sheet mentions. The picture costs nothing to look at. The month it eats does.

Why does 4K streaming data usage hit a cap so fast?

4K hits a cap fast because each stream draws several gigabytes every hour it plays, and a monthly cap counts every one of those hours, so an ordinary evening habit nears a terabyte.

Netflix's help page (read 30 October 2026) puts HD at up to 3 GB an hour and standard definition at 1 GB, per device. Two TVs in 4K are two meters running, not one stream shared.

Xfinity's page, updated 31 August 2026, says the 1.2 TB limit now covers only older, legacy plans; newer plans include unlimited data. So first ask whether you are metered at all.

It is the per-hour sum we ran on what each picture mode and standby add to a TV's yearly electricity bill, with a different meter. If your 4K comes through a stick, quality is set inside each app, which matters more once smart TV software updates stop and a streaming stick takes over.

Four numbers decide whether a capped house copes. Netflix's help pages supply the hourly volume and the speed it asks for 4K; the time and cost cards are our division of those against Xfinity's overage terms.

The speed card is the one people misread. An hour at that rate moves about 6.75 GB, close to Netflix's hourly figure. A faster tier lets more screens play at once; it does nothing for the month.

How many hours of 4K fit in a 1.2 TB month?

On its own, one 4K TV running Netflix gets about 171 hours from a 1.2 TB month, but every phone and second screen in the house draws from the same pool, so your real figure is lower.

Worked example: a legacy-plan house runs one 4K TV four hours a night, a bedroom TV in HD two hours a day, and 250 GB a month of phones and laptops (assumed; your meter has the real figure). The 4K TV uses 4 × 7 × 30 = 840 GB. The bedroom set uses 2 × 3 × 30 = 180 GB. Add the 250 GB and the month ends at 1,270 GB, which is 70 GB over. Xfinity bills by 50 GB block, and assuming a part block counts whole, that is two blocks: $20, or nothing in the courtesy month.

Will your house go over 1.2 TB?Set your 4K hours, the second screen and everything else the house uses. The figures show your month and any overage.

Your numbers

4K hours a day

4.0

0.08.04K data rate, GB an hour

7.0

3.015.0Second screen hours a day

2.0

0.08.0Everything else, GB a month

250 GB

50 GB600 GBSecond screen on Medium

Your month

1,270 GB

Your cap

1,200 GB

Overage after courtesy month

$20

4K hours left under cap

0

You would pay $20 in a month that is not your courtesy month. Try the second screen on Medium first.Reset to the article's exampleSliders need JavaScript. Figures shown are the article's example.Rates from Netflix Help (4K up to 7 GB, HD up to 3 GB, Medium up to 0.7 GB an hour per device, read October 2026); 1.2 TB cap, $10 per 50 GB and $100 maximum from Xfinity, August 2026. Part blocks rounded up (our assumption).

Find your nightly 4K hours (one 4K TV on Netflix, 30 days), then read across at what the rest of the house uses a month. Each cell is the overage outside the courtesy month.

The tinted cell is the worked example, bedroom TV folded into the 430 GB column. One hour less of 4K takes that house to $0; one hour more triples the bill to $60. The late hours cost money, not the first.

What happens if I go over my data cap?

On a capped Xfinity plan, the first month over in any 12 is free under Xfinity's current policy, and after that each 50 GB block costs $10 until the monthly charge reaches $100.

The courtesy month is generous. Well, generous once, and Xfinity says it is subject to change.

The ceiling has an odd consequence. By our arithmetic ten blocks hit it at 1,700 GB, and anything beyond is free that month. Land there twice a year and you are paying for unlimited data without getting it; we would price the unlimited option then.

Overage often comes from screens nobody is watching. A household on Xfinity's community forum reported 1,255 GB in May 2024, writing "We don't have any significant changes to our online habits." Staff pointed at consoles stuck downloading and apps streaming in the background.

Plenty of homes are not capped: Xfinity excludes new plans of 300 Mbps or more, several budget plans, and legacy plans across most of the Northeast, the Mid-Atlantic and DC. With no limit, cutting quality saves nothing, so check your meter or Xfinity's Terabyte Data Usage Plan page first.

If you are capped, the cheapest lever is the second screen. Netflix's Medium setting uses up to 0.7 GB an hour, about 90% less than 4K, and Low 0.3 GB. In the worked example, the bedroom TV on Medium drops from 180 GB to 42 GB and the month ends at 1,132 GB, under the line. Leave the main TV in 4K. If it looks flat, our guide to Filmmaker Mode looking too dark names the two settings to blame.

• Consoles and PCs pull big updates overnight; check the meter's daily view.
• Netflix's Auto setting follows connection speed, so a fast line keeps every screen at top quality.
• The courtesy month returns only a full year after you last used it.

Our call: for most capped houses the fix is a setting, not a new plan. This week, open your Xfinity data meter. No limit: keep every screen in 4K. A 1.2 TB limit: put secondary screens on Medium, and if the meter still ends past the line two months running, price unlimited data before paying another block.