True 4K vs upscaled 4K: reading the sensor spec

The short answer

A true 4K dash cam has an image sensor that natively captures 3840 x 2160, which is about 8 megapixels. An upscaled 4K dash cam has a smaller sensor, commonly around 5 megapixels, and the processor enlarges the picture to fill a 4K-sized file. The file is labeled 4K and plays as 4K, but the extra pixels are interpolated, not captured, so they add file size without adding real detail. The practical test is the sensor model number: Sony IMX415 is an 8MP sensor used for native 4K, while Sony IMX335 is a 5MP sensor that is often sold as 4K through upscaling.

Why the label can mislead

There is no single enforced definition of 4K for a consumer dash cam. A manufacturer can state the output resolution of the video file and remain technically accurate, even when the sensor was never able to capture that many pixels. Online forum discussions of specific boards have noted exactly this pattern with the 5MP part. The honest summary for a buyer is simple: the output size and the captured size are two different numbers, and a spec sheet that gives only the first is incomplete. The pattern spreads because hardware is shared: the same board appears under many labels across marketplaces, so one upscaling decision is repeated across dozens of listings. This is also why verification scales, because a component confirmed once, in writing, serves every label built on that platform.

How the two compare

Sensor and chipset are separate choices. A chipset able to encode 4K can be fitted behind either sensor, so the processor name alone does not prove native capture.

Notice the file size line. Both files occupy similar space at a 4K bitrate, so storage planning cannot separate them; only a zoomed comparison does. A customer who never zooms may never notice, but a customer who does will return the product.

What an upscaled model can still be good for

Upscaling is not a defect, and a 5MP sensor is a sound part. A well-tuned 5MP camera, sold as what it is, can be a sensible mid-range product with lower heat and a lower data rate. The problem is the claim, not the part. A listing that promises 4K plate detail from a 5MP sensor sets up a return when the customer zooms into the footage. If you stock such a model, describe it by its real recorded resolution or its sensor megapixels and let the tier name follow honestly. It also leaves room for a cleaner range story: an honestly labeled 5MP model undercuts the flagship on heat and data load, and both listings can point to the same plain explanation of what each records.

How to verify before you order

1. Ask for the sensor model and its datasheet megapixel count, in writing, on the specification for your program.
2. Request a raw sample clip from the actual unit, recorded at default settings, in daylight with a fixed target such as a printed resolution chart or readable signs at known distances.
3. Compare two units on the same scene: a known native model and the candidate. Pause and zoom to 100 percent on a plate-sized target.
4. Check the file: the container reports the encoded size, but only a zoomed comparison shows whether detail actually exists at that size.
5. Record the result in the sample approval note so the production spec cannot drift.

This is the same sample discipline described in the pre-bulk sample checklist, and Roadtrace runs it with partner factories as part of the quality control routine.

The same discipline protects you mid-program. Component supply shifts over a production run, so the specification should name the sensor and require written approval, with a new sample, for any substitution. A kept golden sample from the approved batch gives the inspector a reference when a later shipment arrives.

The wider sensor picture

Resolution is only one reason sensors differ. IMX675, a Sony STARVIS 2 part used in newer models, is chosen mainly for better low-light behavior, and rear cameras often use a 2MP 1080P part such as the GalaxyCore GC2053. These are component facts, not endorsements: the value of any sensor depends on the lens, the tuning and the recorder around it. The image sensor note goes through the model numbers a buyer is likely to meet, and the resolution guide covers the tiers. When you brief a factory, the useful question is not which sensor is newest but which pairing the platform is tuned around, since exposure and noise processing are optimized for a family of parts rather than one model.

Listings, wording and the sourcing brief

Whichever way you buy, write the listing so the customer cannot feel misled. Name the recorded resolution, mention the sensor class where your platform allows it, and avoid promising plate capture at distance. Sellers who label a native model clearly can also charge for that clarity in a range with both: a true 4K flagship and an honestly described 5MP value model reads as a considered line rather than two near-identical boxes. Retail buyers do read the details, and a return caused by a stretched image costs more than the label it saved.

Bringing it into a sourcing brief

When you brief a supplier, state the requirement in sensor terms: native 3840 x 2160, 8MP class sensor, with the model named. Commercially, a native 4K line is bought like any other model on this platform: typical published terms from Shenzhen OEM suppliers put private-label minimums from about 100 pcs per model, with firmware or app branding carrying higher minimums, stock samples shipping in days, custom samples in about 7 to 15 working days, and stock-based production in about 15 working days against about 25 to 45 days for custom programs. Treat every figure as a typical reference range to be confirmed in writing for each program; the MOQ and lead time page and the MOQ explained note give the working detail. For a sourced shortlist of native 4K lines, see the 4K dash cam line and send your brief through the quote page.