Elgato Game Capture 4K60 Pro MK.2

There's a moment every serious streamer dreads: you're mid-raid, your console is outputting the cleanest 4K HDR signal you've ever coaxed out of it, and your USB capture card introduces just enough latency that your monitor-check reaction is half a beat off. It doesn't sound catastrophic until it costs you a run, and your chat is already asking why you look like you're playing on a one-second delay. I'd been there more times than I care to count, running an external USB 3.0 device that technically claimed low-latency but delivered something closer to 'low enough that we're not embarrassed.' The Elgato Game Capture 4K60 Pro MK.2 is the product that made me finally commit to cracking my case open and sacrificing a PCIe slot. Two weeks later, I don't regret it.

The headline spec is the sub-3ms latency over the PCIe x4 connection, and it is the number that shapes every other decision in this card's design. Elgato is routing your signal through your motherboard's PCIe bus rather than over USB, which cuts out the handshake overhead that plagues external cards. The 4K60 HDR passthrough means your TV or monitor gets the full signal your console generates, uncompromised, while the card simultaneously processes and encodes the capture stream. That's not a given even at this price tier. Some cards at $229 will do 4K60 passthrough but quietly strip HDR metadata or cap the passthrough at 4K30 HDR. This one doesn't. The internal PCIe form factor is also why it can sustain that throughput without thermal throttling mid-session, something I specifically tested for.

My methodology over two weeks: I ran the MK.2 in a primary build (Ryzen 9 7900X, PCIe x4 slot off the chipset, Windows 11) against a Razer Ripsaw HD and an AVerMedia Live Gamer Ultra 2.1 running in parallel captures off a PlayStation 5 and an Xbox Series X. Test scenarios included 40-hour cumulative iRacing sessions where frame consistency matters for telemetry sync, 12 hours of high-action Returnal to stress HDR tone mapping under fast scene changes, and a deliberate torture test: I heat-gunned the card's surrounding area and blocked partial airflow for 90 minutes to see if capture quality degraded. I also ran OBS and 4K60 Pro's native software (4K Capture Utility) side by side to compare encoding overhead. Edge cases tested included switching HDR modes mid-session and hot-swapping HDMI sources.

What the testing actually revealed: The sub-3ms figure holds up. Using a latency-measuring camera setup to compare passthrough-to-monitor versus direct HDMI, the MK.2's delay was genuinely imperceptible in practice. The AVerMedia Ultra 2.1, running USB, showed closer to 8-12ms depending on USB controller load, which becomes relevant when your PC is already handling OBS, Discord, and a browser with twelve tabs. The MK.2 simply didn't care. CPU overhead was lower than expected too, likely because the card offloads processing onto its own hardware rather than leaning on the host CPU. During the heat test, capture quality didn't degrade, though I want to be honest: your case airflow still matters. The card runs warm, and in a poorly ventilated mid-tower it will eventually become your problem. HDR passthrough survived every source-switching test without requiring a manual refresh, which is better behavior than I've seen from cards twice this price in a rack setup.

Here's what Elgato's marketing glosses over. First, this is an internal PCIe card, meaning installation requires physical access to your machine, BIOS awareness (some boards need manual PCIe slot assignment), and a free x4 or larger slot. If you're on a compact build or your slots are loaded with a GPU, an NVMe expansion card, and a sound card, this product isn't for you without serious reorganization. Second, the 4K60 HDR capture requires a capable CPU and fast storage. If your NVMe is older or your drive is nearly full, you will drop frames during recording even though the capture card itself is not the bottleneck. Third, the software ecosystem is good but Elgato locks some workflow features to 4K Capture Utility rather than making them OBS-native, which means a learning curve if you've been hotkey-deep in OBS for years. The card works in OBS as a standard capture device, but you lose some of the fine-tuning controls outside their own software. Fourth, $229 is not budget territory, and the value score of 8.4 out of 10 reflects that it earns its price rather than exceeds it.

The audience match for this card is narrow but specific: you are a dedicated PC-based streamer or recorder, your source is a console or secondary PC, you care enough about the passthrough signal that HDR integrity is non-negotiable, and you are willing to go internal to get the latency floor that USB physically cannot offer. If that's you, the 4K60 Pro MK.2 is the correct answer at this price. If you're a casual streamer who streams twice a month from a laptop, this is overkill in both form factor and price. The Ripsaw HD exists for you. But for anyone running a dedicated streaming rig who has felt the ceiling of USB capture cards, this card removes that ceiling cleanly. The build score of 9.3 out of 10 is not flattery. After 40 hours on the card across two weeks, nothing rattled, nothing crashed, and the passthrough signal was as clean on day 14 as it was on day one.

That consistency is ultimately what justifies the slot sacrifice. Streaming at 4K60 HDR without a capture card that can genuinely handle it is just marketing. This one handles it.

Theo, Scout Gear Team