The real economics of memory cards.
Memory cards are one of the few photography purchases where buying cheap actively costs more — in lost frames, dropped video, and the unique horror of a corrupted card after a wedding.
The market has bifurcated in the last few years. On one side, you have a flood of generic SD cards from no-name brands on auction sites and marketplaces — often counterfeit, frequently failing, sometimes with inflated capacity claims that erase data the moment you exceed the real limit. On the other side, a small group of trusted manufacturers — SanDisk, Lexar, ProGrade, Sony, Delkin — produce cards that survive professional abuse year after year.
The price gap between these two tiers is real, often two to three times. The reliability gap is much larger.
What changes with camera resolution
The math behind card capacity is simple but its consequences are not. A 24-megapixel camera shooting RAW produces roughly 25 to 30 megabyte files. A 45-megapixel camera produces 50 to 60 megabyte files. The newer 60 and 100-megapixel sensors push individual frames toward 80 megabytes and beyond.
This means that the 128GB card that comfortably held a thousand RAW frames on your old camera will hold around four hundred on the new one. The upgrade math nobody warns you about: doubling sensor resolution roughly halves your usable card capacity.
Why bitrate destroys video budgets
Video is even more aggressive on storage. A modern mirrorless camera shooting 4K at 100 Mbps consumes about 750 megabytes per minute. At 400 Mbps All-Intra, the same minute eats roughly 3 gigabytes. The 8K modes on flagship cameras can hit 800 Mbps and beyond — which is why CFexpress Type B became necessary and why a single hour of footage may need a 1TB card.
"A 128GB CFexpress card can hold under twenty minutes of 8K RAW video. That is the new reality of professional shooting."
The format compromise
Most working photographers settle into one of three storage strategies. Each has trade-offs:
The speed ratings that actually matter
The packaging on a typical memory card is covered in numbers and symbols that mean different things and sometimes contradict each other. Three matter most:
- UHS Speed Class — UHS-I caps around 104 MB/s. UHS-II adds a second row of pins and reaches up to 312 MB/s in compatible cameras and readers.
- Video Speed Class (V30 / V60 / V90) — The minimum sustained write speed in megabytes per second. V30 is the floor for 4K video. V60 and V90 are required for high-bitrate 4K and 8K respectively.
- Sustained vs burst speeds — Marketing speeds describe peak burst performance. What matters for video and rapid-fire stills is the sustained write speed, often half of the advertised number.
The number you see on the front of the card — typically 170 MB/s or 200 MB/s — describes read speed, which matters for offload but not for capture. A card that reads at 200 MB/s but writes at 90 MB/s will struggle with high-bitrate video despite the impressive sticker.
Counterfeit cards and where they come from
The counterfeit problem is real and well-documented. The cards are convincing physically — the printing matches, the branding looks correct, sometimes the controller is even close to right. What fails is the underlying NAND flash. A fake card advertised as 256GB might have only 32GB of actual usable storage with controller tricks that allow apparent writing to higher addresses, silently dropping data.
The defense is simple and unglamorous: buy from authorized US retailers. B&H Photo Video, Adorama, Best Buy, Amazon Direct from Manufacturer, and Crutchfield all source directly from manufacturers and stand behind warranties. Marketplaces with mixed third-party sellers — including some on Amazon under third-party listings — are where counterfeit cards proliferate. A 30-second check with H2testw or F3 after purchase verifies capacity.
What this calculator does
The estimator above takes your camera model and combines it with average file sizes from real-world testing data published by manufacturers and reviewed by professional photographers. It then divides your selected card capacity by those averages, with a 10% overhead reserved for file system and metadata. The result is a directional estimate accurate to within 15 percent for most real-world shooting.
It is not a guarantee. Scene complexity, ISO setting, color depth, and individual camera firmware all affect actual file sizes. The estimates assume average mixed-content shooting at base ISO. High-ISO shooting in low light typically produces larger RAW files; flat color scenes produce smaller ones.