GPU Mirrored Output to a Capture Card Shows Black or Records Nothing? Duplicate Displays, EDID, HDCP & Passthrough Troubleshooting

2026-10-25 ｜ GPU Repair ｜ About 21 minutes ｜ HK GPU Fix Editorial Team

重點摘要

When a duplicated display sent to a capture card shows black or records nothing, the cause is usually not recording performance. Duplicate mode forces one common resolution and refresh rate, the capture card's EDID shapes what the source sends, and HDCP affects only protected content. This guide separates those causes step by step.

The Short Answer First

This guide deals with one specific arrangement: a picture is duplicated, or cloned, to a capture card, or passed through a capture card to a monitor, and the result is a black screen, no signal, or a recording that contains nothing. The first step in every such case is to say exactly which side is black, because the two sides can fail independently. A monitor can show a picture while the capture software shows black, or the capture software can show a picture while the monitor shows black, or both sides can be black. Each pattern points to different checks, and treating all black screens as one problem is the fastest route to replacing the wrong part.

Three ideas organise the whole subject. The first is duplicate mode itself: when the same picture is sent to two destinations, the system must choose one resolution and refresh rate that both destinations support, and that common mode can be lower than either destination's best mode. The second is EDID: the information a display or a capture card sends to the source about what it can accept, which shapes the mode the source chooses. The third is the distinction between the capture card's passthrough specification and its capture specification, which are different numbers with different limits. High-definition copy protection, known as HDCP, is a fourth factor, but only for protected content, and it must not be used as a blanket explanation for every black screen.

Two boundaries must be stated at the outset, because they prevent the two most common misdiagnoses. This guide is not about encoder overload. Encoder overload, covered in a separate guide on NVENC encoder load, produces a recording that stutters, drops frames, or triggers encoder warnings while input is present. A black input, where no picture reaches the capture card or the monitor at all, is a different symptom with different causes. This guide is also not a general guide to HDCP handshakes on an ordinary display chain. A separate guide covers HDCP handshakes between a source, a cable, and a display. This guide deals with HDCP only in the specific chain that includes a duplicated output and a capture card, with its EDID modes and its passthrough path. Keeping those boundaries clear is what makes the diagnosis in this guide work.

Duplicate Mode Versus Extend Mode: The Common Mode Rule

Windows offers two fundamentally different ways to use two displays, and the difference between them is the single most misunderstood point in capture troubleshooting. In extend mode, each display is independent. Each display can use its own native resolution and refresh rate, because each display receives its own picture. In duplicate mode, also called clone mode, the same picture is sent to two or more destinations. The system must therefore find one mode that every destination supports: one resolution, one refresh rate, one set of timing parameters that all destinations can accept.

An official AMD explanation illustrates the rule with a concrete pair of displays. One display has a native resolution of 1920 by 1080 at 75Hz, and the other has a native resolution of 1280 by 1024 at 60Hz. In duplicate mode, both displays run at 1280 by 1024 at 60Hz, because that is the highest mode they share. That result is not a fault, a driver bug, or a sign that one display is slowing the other down. It is the definition of duplication: one picture, one mode, shared by all destinations.

The consequence for capture cards is immediate. Suppose the main monitor is a 1440p display running at 144Hz, or a 4K display at 144Hz, while the capture card only supports capturing at 1080p at 60Hz. Duplicating to that capture card forces a compromise. Either the main monitor is pulled down to a mode the capture card also supports, or the capture side fails because the chosen mode exceeds what the capture card can accept. Users often report this as a sudden drop in the main monitor's refresh rate when duplication is switched on, and they interpret it as the graphics card throttling or the capture card slowing the main display. In the usual case, neither interpretation is correct. The refresh rate drop is the common mode rule doing exactly what it is defined to do.

A related expectation must also be corrected. In duplicate mode, the main monitor cannot keep its highest refresh rate while the capture card independently runs at a lower refresh rate. That expectation contradicts the definition of duplication. A user who insists on keeping the main monitor at its highest refresh rate while also capturing needs a different arrangement, such as extend mode with a specific window captured, or a passthrough arrangement, each with its own trade-offs described below. Duplication cannot deliver two independent best modes at once, and no setting, cable, or driver update changes that definition.

In Windows display settings, duplicate mode also constrains the refresh rate options offered for individual displays, because the options must remain consistent across the duplicated destinations. A refresh rate that is available for a display on its own may disappear, or may be shared, once duplication is active. That behaviour is another expression of the same rule, not a separate fault to chase.

Two Wiring Arrangements, Kept Separate

Capture troubleshooting goes wrong when two different wiring arrangements are treated as one. This guide keeps them separate throughout.

The first arrangement is Windows duplication to a capture card. The graphics card output connects to the capture card input, and the capture card connects by USB or PCIe to a second computer, in a dual-computer setup, or to recording software on the same computer. In this arrangement, the capture card appears in Windows as a display, with its own EDID. The duplicate mode common mode rule applies directly, because the capture card is one of the duplicated destinations. A typical use is duplicating a game console or a main display to a streaming computer for recording.

The second arrangement is capture card passthrough, also called loop-out. The source connects to the capture card input, the capture card output connects to a monitor, and the capture card also sends a copy of the signal to recording software. The user watches the monitor through the passthrough path while the software records the captured copy. The key point of this arrangement is that passthrough and capture are different paths with different specifications. Passthrough does not require encoding, so it can support a higher resolution or refresh rate than capture. Capture requires compression and transmission over USB or PCIe, so its ceiling is usually lower, limited by connection bandwidth, encoding hardware, and software overhead. An official Elgato explanation states the same principle: passthrough can exceed capture precisely because it does not need to be encoded.

Three numbers must therefore be read separately for any capture card, each as stated in that model's specification table: the input limit, the passthrough limit, and the capture limit. A card advertised with a high passthrough figure does not thereby promise the same figure for capture, and a card advertised with a particular capture figure does not thereby limit its passthrough to that figure. Readers who check only one number, usually the largest one in a product listing, will be surprised in exactly the situations this guide addresses.

EDID: How the Source Learns What the Other End Accepts

EDID is the data by which a display or a capture card tells the source what it can accept: which resolutions, refresh rates, colour depths, and audio formats it supports. The source reads that data and chooses an output mode within it. If the EDID data does not match the actual hardware capability, the source may choose a mode the hardware cannot display or capture, producing a black screen, no signal, or a wrong mode, even though every cable is intact.

Capture cards commonly offer several EDID modes, and the Elgato 4K Capture Utility provides a clear example, with the caution that mode names and defaults vary by manufacturer and model, so the names here should not be treated as universal across all brands. In that example, the first mode is called Merged, and it is usually the default. It combines the EDID of the display connected to the passthrough output with the capture card's own capabilities, producing an optimised EDID for a typical arrangement where a console or computer connects through the capture card to a television or monitor. The second mode is called Internal. It ignores the passthrough display and uses an EDID stored inside the capture card, for example an EDID corresponding to a 1080p or 1440p main display. It suits dual-computer capture, or an arrangement where no display is connected to the passthrough output. Its warning matters: if the internal EDID defines a mode beyond the connected display's capability, that display may show black or no signal. The third mode is called Display. It copies the passthrough display's EDID directly to the capture card, and it suits use as a control test when the Merged mode causes trouble. Its warning is the mirror image: if the display's EDID defines a mode beyond the capture card's capability, the captured signal may be wrong or entirely absent.

Those three modes explain why the same wiring can produce opposite symptoms. Depending on which EDID mode is active and which destination exceeds its capability, the result can be a monitor with a picture and black capture software, or capture software with a picture and a black monitor. Changing the EDID mode is therefore not a matter of trying modes at random in the hope that one works. It is a logical isolation step: if Merged causes trouble, try Display as a control; in a dual-computer arrangement with no passthrough display, try Internal matched to the main display's resolution. Each change should alter only one variable, and the result for each side should be noted separately.

A capture card with no display connected, sometimes called a headless arrangement, has no physical EDID source of its own. In that case, an internal EDID is the only basis on which the source can choose a mode. That idea is related to the dummy plug concept used for monitorless graphics cards, where a small device presents an EDID so that a source has a display target, but this guide does not expand into dummy plug instructions, because the subject here is the capture card's own EDID modes, not monitorless operation in general.

The HDCP Chain, and Why It Does Not Explain Every Black Screen

HDCP is a content protection encryption system. When a source outputs protected content, it encrypts the signal, and each sink or repeater in the chain, including a capture card, a splitter, or a passthrough path, must authenticate before it can decrypt and display or forward the content. Most capture cards are designed not to accept HDCP-encrypted input, which means they will not record protected content. When a source outputs protected content to such a card, the recording software shows black or no signal. That outcome is a design limit, not evidence that the capture card is broken.

The essential distinction in this guide is between protected and unprotected content. Unprotected content, such as an ordinary desktop, an ordinary game picture, or the user's own unprotected material, is not normally blocked by HDCP. If unprotected content also shows black, the priority should be EDID, the duplicate common mode, the passthrough and capture specification limits, and the connection, not HDCP. Protected content, such as some streaming services, protected video, and some console content, can show black even when the wiring, EDID, and resolution are all correct, because the HDCP chain itself blocks it. In that situation, the correct diagnostic step is a controlled comparison: test with unprotected content, then test with protected content. Unprotected content with a picture and protected content black points toward the HDCP chain. Both kinds of content black points elsewhere, and HDCP should not be blamed.

Each additional device in the middle of the chain — a splitter, a converter, a capture card, a keyboard-video-mouse switch — adds another authentication step, and therefore another place where the chain can fail. Diagnosis should bypass middle devices one at a time, as described in the diagnostic sequence below, rather than assuming that the first device in the chain is at fault.

A firm boundary applies to this whole section, and it must be stated plainly. This guide does not teach, recommend, or point toward any method of stripping or bypassing HDCP, and it does not recommend any device claimed to decrypt protected content. The legitimate directions are threefold: use unprotected content for testing and recording; where the source itself provides an HDCP setting, as some consoles do in their system settings, check that setting only as described in that source manufacturer's own documentation, noting that changing it may affect playback of other protected services; and use any HDCP-compliant recording path the manufacturer provides, where one exists. No other direction is offered here.

Other Specification Limits: DSC, VRR, HDR, Colour, Connection, Software

Beyond the common mode rule, EDID, and HDCP, several specification limits can independently produce a black screen, and each must be checked separately rather than lumped together as a general capture failure.

Display Stream Compression is one such limit. Some capture cards explicitly do not support DSC signals for passthrough or capture, and the Elgato Game Capture 4K Pro specification is one documented example. If a source outputs a DSC signal in order to reach a high resolution or refresh rate, such a capture card may show no signal at all. In that situation, a direct connection from the source to a monitor showing a picture, while the same signal through the capture card shows nothing, is itself evidence pointing toward DSC or the capture card's supported range, not toward a broken cable. The cable has already proven itself by carrying the direct signal.

Variable refresh rate, high dynamic range, and auto low latency mode form another group whose support can differ completely between passthrough and capture. One model may offer high refresh rate passthrough with HDR and variable refresh rate, while its HDR capture is limited to a lower resolution or refresh rate. Another model may support variable refresh rate passthrough only within a particular resolution, refresh rate, and chroma subsampling range, for example requiring a particular subsampling format where another format is not supported for variable refresh rate. Those examples are given only to show the shape of the problem, not as universal figures, and every such claim must be checked against the individual model's specification table.

Chroma subsampling and colour depth belong in the same check. The figures 4:4:4, 4:2:2, and 4:2:0, together with 8-bit and 10-bit colour, affect both the bandwidth a mode requires and the range a capture card accepts. A specification that reads only as a resolution and refresh rate, such as 4K at 60Hz, is incomplete without those additional parameters and the HDR setting. Two modes that share the same resolution and refresh rate can differ in whether a capture card accepts them, and reading only the headline figure hides that difference.

The physical connection also sets a capture ceiling that does not apply to passthrough. A USB capture card's capture limit depends on the USB version of the port in use, such as USB 2.0, USB 3.0, or USB 3.2 generations, and on the power available from that port. A PCIe card's limit depends on its PCIe lanes. Passthrough, which does not travel over USB, is not limited by USB bandwidth, which is why a monitor can show a picture through passthrough while recording simultaneously fails or shows black. A USB extension cable, a poorly powered hub, or a USB 2.0 port can therefore produce intermittent black screens that look like EDID or HDCP failures but are in fact connection failures, and they are isolated by changing the port and using a short direct cable.

Finally, the software layer can produce a false black screen. A capture card can usually be used by only one program at a time, so a second program opening the same card may show black even though the card and the signal are fine. In recording software such as OBS, the source must be added as a video capture device, and the correct device name must be selected; a wrong source selection produces a black preview that looks like a hardware failure but is only a selection error. Those software checks belong early in the sequence, before hardware is suspected, because they cost nothing and they rule out a whole class of false alarms.

Diagnosis in Order: Which Side Is Black, Then Narrow Down

The diagnostic principle for this subject is to define the symptom precisely, establish a direct baseline, reduce to a common mode both sides certainly support, and then add middle devices back one at a time. Changing several settings at once destroys the evidence, because a recovery after several changes does not reveal which change mattered.

First, classify the symptom into one of four patterns, and keep that classification in mind throughout. Pattern one: the monitor shows a picture, but the capture software shows black or no signal. Pattern two: the monitor shows black, but the capture software shows a picture. Pattern three: both sides show black or no signal. Pattern four: the capture software preview shows a picture, but the recorded file is black or empty. Pattern four points first toward recording software source or encoding settings, and this guide mentions it only to assign it correctly; the detailed encoding discussion belongs to the encoder overload guide, and it is not expanded here. The main focus of this guide is patterns one to three, where input itself is missing on one or both sides.

Second, establish a direct baseline by bypassing every middle device. Connect the source directly to the monitor, without the capture card in the path, and confirm that the source outputs normally and that the monitor shows a picture at the target mode. If the direct connection is also black, the problem is not a capture card problem, and capture troubleshooting should stop until the direct path works.

Third, connect the source directly to the capture card, with no passthrough monitor attached, and observe whether the capture software shows a picture. This step separates the capture input, EDID, and HDCP questions from the passthrough output question. A failure here points toward the input side; a success here, followed by a failure when the passthrough monitor is added, points toward the passthrough side or the EDID mode that combines the two.

Fourth, reduce to a common mode that both sides certainly support. In duplicate mode, set a low combination such as 1080p at 60Hz, with HDR off, variable refresh rate off, and DSC off, where the individual model's settings allow those controls. If a picture appears at the common mode, the evidence points toward a common mode or specification limit, not toward broken hardware. If no picture appears even at the common mode, continue to the EDID, HDCP, and connection checks below, because the cause lies deeper than the mode choice.

Fifth, perform the HDCP comparison using only legitimate methods. Test with unprotected content, such as a desktop or an ordinary game picture, and then test with protected content. Unprotected content with a picture and protected content black points toward the HDCP chain. Both kinds of content black points toward EDID, mode, or connection, and HDCP should not be blamed in that case.

Sixth, isolate the EDID mode. Using the capture card manufacturer's software, try the available EDID modes one at a time, in the order that manufacturer suggests where one is given, changing only that one setting and noting which side gains or loses a picture each time. Where an internal EDID is used, it must match the main display's resolution; for example, a 1440p main display calls for a 1440p internal EDID, not an unrelated mode chosen at random. A mode that restores one side while breaking the other has still provided useful information, because it shows which destination's capability is being exceeded.

Seventh, isolate the passthrough path. In the source to capture card to monitor arrangement, change one segment at a time: the cable, the passthrough output port, and the monitor input port. A black passthrough monitor with working capture points toward the passthrough output or the monitor mode. Black capture with a working passthrough monitor points toward the capture input, EDID, HDCP, or the capture specification limit. Each side's independence is the diagnostic tool, not a mystery to be puzzled over.

Eighth, isolate DSC, variable refresh rate, and HDR one at a time. Switch each feature off in turn and retest, noting which switch restores a picture on which side. Switching all three off at once and then declaring the problem solved leaves the cause unknown, and the problem will return as soon as the features are switched back on without understanding which one mattered.

Ninth, check the connection and software layer. Use a USB 3.0 or higher port where the model requires it, avoid USB extension cables and poorly powered hubs, close any other program that may be holding the capture card exclusively, re-add the video capture device source in the recording software and confirm the device name, and update the capture card driver and firmware where the manufacturer provides an update. Each of those steps is cheap, and each rules out a distinct cause.

Tenth, compare duplicate mode with extend mode. Change Windows from duplicate to extend, and observe whether the capture card side can then reach its own native mode independently. If extend works and duplicate shows black, the evidence points toward the duplicate common mode limit, not toward capture card hardware. That single comparison often resolves the whole case, because it changes only the mode rule while leaving the hardware untouched.

Eleventh, and only after the direct baseline, the common mode test, the EDID isolation, the HDCP comparison, the passthrough isolation, and the software layer have all been completed, consider graphics card hardware. Suspicion of a graphics card output or interface is warranted when a particular output still produces no signal at all after every other step, and at that point a hardware inspection is sensible, with the fee confirmed after inspection as described at the end of this guide.

What the Symptoms Suggest, and What They Do Not Prove

As with any symptom table, the rows below are priorities for investigation, not diagnoses. A symptom suggests where to look first, and the final column names the conclusion that must not be jumped to before those checks are done.

The contrast with encoder overload should be kept in view throughout that table. Encoder overload produces a recording that stutters, drops frames, or triggers encoder warnings while input is present. A black input produces no picture to encode at all. The two symptoms may appear in the same recording session, but they have different causes and different fixes, and conflating them leads to encoder settings being adjusted for a problem that lies in EDID or in the duplicate mode rule.

Boundary Cases

Several boundary cases change the answer, and each deserves a plain statement rather than a guess.

EDID emulation is a tool, not a universal remedy. Some capture cards and EDID emulators can present a fixed EDID to the source, so that the source outputs a fixed mode. That is useful in dual-computer setups, in arrangements with no passthrough display, or where a display is frequently switched off. But an emulated EDID that defines a mode the display or the capture card does not support will itself cause a black screen, in the opposite direction from the problem it was meant to solve. Every change to an emulated EDID must be matched to the actual capability of the destination it serves.

The protected versus unprotected comparison is a diagnostic standard, not a random fault. The same wiring, the same cable, and the same card can show a picture for unprotected content and black for protected content, entirely consistently. That consistency is the point: it separates a protection chain effect from an intermittent hardware fault, which would not respect the content boundary so neatly.

Passthrough is not a guarantee of zero delay or zero loss. Some models advertise low-delay passthrough, but delay and the supported ranges for HDR, variable refresh rate, and other features remain as stated in each model's specification table. A claim that passthrough is always delay-free would go beyond what any single specification can support, and this guide does not make it.

Dual-computer and single-computer arrangements call for different EDID choices. In a dual-computer arrangement, where a gaming computer feeds a streaming computer's capture card, the EDID mode that suits the setup differs from the mode that suits a single computer duplicating to its own capture card, because the passthrough display's role differs between the two. Asking which arrangement is in use, before recommending an EDID mode, is therefore essential, and one set of EDID advice should not be applied to both arrangements unchanged.

A console HDCP setting, where the console manufacturer provides one, illustrates the legitimate direction for HDCP questions. Some consoles provide a system setting to enable or disable HDCP for HDMI output, and disabling it may affect playback of some protected services, as that console manufacturer's documentation states. This guide mentions such a setting only in that conditional form: if the manufacturer provides it, check it as that manufacturer describes. It does not claim that every console has such a setting, and it does not present disabling HDCP as a general step for all sources.

High-resolution and high-refresh-rate capture each have independent ceilings. A 4K capture ceiling, a 144Hz capture ceiling, and an HDR capture ceiling are separate figures, and each must be read together with its chroma subsampling, colour depth, and encoding format where the specification states them. A product claim that mentions only 4K, without those companions, does not establish that every 4K mode is captured.

USB power and port choice can mimic EDID and HDCP failures. A USB capture card on an underpowered port, a USB 2.0 port, or a long extension cable can show intermittent black screens or no signal, with symptoms that look identical to an EDID or protection failure. The isolation for that cause is simple: change to a short direct cable on a suitable port, and observe whether the symptom changes. Symptom alone cannot distinguish the causes, but that one change can.

Software exclusivity is a similar mimic. Two recording programs competing for the same card will leave the second program black, even though the card is working correctly. Closing all other programs and retesting with a single program is a required software step, not an optional tidiness measure.

Finally, the combination of Windows duplication, a capture card, and a high-refresh-rate main monitor is the arrangement most likely to collide with the common mode limit. If the user insists on keeping the main monitor at its highest refresh rate, the honest direction is to consider extend mode with a specific window captured, or a passthrough arrangement, and to accept the trade-off that each alternative carries, rather than promising that duplication can somehow keep both sides at their best modes at once. Duplication, by definition, cannot do that, and no product claim changes the definition.

Misconceptions, Corrected in One Place

The first misconception is that a black capture preview means a faulty graphics card or an overloaded encoder. A black preview can be caused by EDID, by the duplicate common mode, by HDCP for protected content, by a passthrough or capture specification limit, or by a wrong software source selection. Encoder overload, by contrast, shows stutter, dropped frames, and encoder warnings while input is present. The two symptoms are different, and they point to different guides.

The second misconception is that every capture card black screen is HDCP. If unprotected content is also black, HDCP is not the priority. EDID, mode, and connection checks come first in that case, and HDCP should only be blamed where the protected versus unprotected comparison supports it.

The third misconception is that the passthrough specification and the capture specification are the same number. They are not. Passthrough needs no encoding and can exceed capture, while capture needs compression and transmission and is therefore usually lower. Reading only the larger figure misleads in both directions.

The fourth misconception is that in duplicate mode the main monitor can keep 144Hz while the capture card independently runs at 60Hz. Duplication requires a common mode, so the two sides cannot independently run at their best modes. That expectation must be abandoned before any other diagnosis can make sense.

The fifth misconception is that a capture card is the same as an HDMI splitter. A splitter only copies a signal to several displays. A capture card also handles EDID, participates in the HDCP chain, encodes, and transmits over USB or PCIe, with limits that a splitter does not have.

The sixth misconception is that a more expensive cable solves an EDID or HDCP problem. A cable is one segment of the chain. EDID mode, the HDCP chain, and the common mode limit are not changed by replacing a cable, and a cable replacement that appears to help should be checked against the possibility that reseating, rather than the new cable itself, changed the link negotiation.

The seventh misconception is that a card advertised as supporting 4K thereby records every 4K mode. Input, passthrough, and capture ceilings are separate, and each must be read with chroma subsampling, colour depth, HDR, variable refresh rate, and DSC support for that model.

The eighth misconception concerns any device claimed to bypass protection. This guide does not recommend such a device, and it does not teach any bypass method. The legitimate directions are those stated in the HDCP section: test and record with unprotected content, check a source manufacturer's own HDCP setting where one is provided, and use a manufacturer-compliant recording path where one exists.

The ninth misconception is that a working monitor proves the capture card is fine. A working passthrough monitor proves only that the passthrough path works. The capture input, EDID, and HDCP paths can fail independently of passthrough.

The tenth misconception is the mirror image: that a working capture preview proves the monitor is fine. A working capture preview proves only that the capture path works. The passthrough output and the monitor mode can fail independently of capture.

The eleventh misconception is that switching DSC, variable refresh rate, and HDR on is always better. Each feature has a supported range on each capture card, and switching one on can itself cause the capture card to show no signal. Each feature should be isolated on and off in turn, so that the feature responsible is identified.

When Inspection Is Warranted, and When It Is Not

A hardware inspection is warranted when a graphics card output shows a fault on its own, independent of the capture arrangement. That includes an output that produces no signal when connected directly to a single known-good monitor, or an output that fails in the same way regardless of which monitor, cable, capture card, or EDID mode is used. Those findings point toward the output or the card itself, and they are worth investigating as hardware questions.

An inspection is not warranted as a remedy for a duplicate mode limit, an EDID mismatch, a protection chain effect on protected content, or a passthrough versus capture specification difference. If a direct connection works, a lower common mode works, an EDID change restores one side, or unprotected content works while protected content does not, no inspection will change the mode rule, the EDID, or the protection chain. The time is better spent on the sequence above, which identifies the actual limiting factor.

If inspection is needed, the workshop is at Shek Mun Dali Plaza Room 1209, reachable by phone on 2186 8010 or by WhatsApp on 6558 6806. Diagnostic and repair fees follow the published three-tier range of HK$400–2,800, with the actual figure confirmed after inspection. There is no Mong Kok shop and no on-site service.

Frequently Asked Questions

Why does my capture software show black while my monitor is fine?

The two paths are independent. A working passthrough monitor proves only that the passthrough path works. Black capture software points toward the capture input mode, the EDID mode, the capture specification limit, HDCP for protected content, or a wrong software source selection. Check the capture side separately, starting with a low common mode and an EDID mode change, one variable at a time.

Why does my main monitor's refresh rate drop when I duplicate to a capture card?

Duplicate mode requires one mode that all destinations share. If the capture card supports only a lower resolution or refresh rate than the main monitor, the shared mode is the lower one, and the main monitor follows it. That is the definition of duplication, not throttling and not a capture card slowing the main display. Keeping the main monitor at its highest refresh rate while duplicating is not possible under duplicate mode.

Is every capture black screen caused by HDCP?

No. Test with unprotected content first. If unprotected content such as a desktop or an ordinary game picture is also black, the cause lies in EDID, the duplicate common mode, a specification limit, or the connection, not in HDCP. HDCP is supported as an explanation only where unprotected content works and protected content does not.

What is the difference between passthrough and capture specifications?

Passthrough forwards the signal to a monitor without encoding, so it can support a higher resolution or refresh rate. Capture compresses the signal and transmits it over USB or PCIe, so its ceiling is usually lower. A card advertised with a high passthrough figure does not promise the same figure for capture, and the input, passthrough, and capture limits must be read separately from that model's specification table.

Which EDID mode should I use on my capture card?

It depends on the arrangement, and mode names vary by manufacturer, so the Elgato example of Merged, Internal, and Display should not be treated as universal. As a guide, a typical arrangement with a passthrough monitor often starts with a merged mode that combines the monitor and the card. A dual-computer arrangement with no passthrough monitor often suits an internal EDID matched to the main display's resolution. A display mode that copies the monitor's EDID is useful as a control test when the merged mode fails. Change one mode at a time and note which side gains or loses a picture.

My preview shows a picture but the recorded file is black. Is the capture card broken?

Not necessarily, and not on that evidence alone. A preview with a picture shows that input is reaching the software. A black recorded file points first toward the recording software's source or encoding settings, which belong to the encoder discussion rather than to input diagnosis. Check the software source selection and encoding settings before suspecting capture hardware.

Can I record protected content by using a different device or cable?

No method of bypassing content protection is taught or recommended here. The legitimate directions are to test and record with unprotected content, to check a source manufacturer's own HDCP setting where one is provided and as that manufacturer describes it, and to use a manufacturer-compliant recording path where one exists.

Why does 1080p at 60Hz work but 4K or a high refresh rate shows black?

A higher mode may exceed the capture card's input, passthrough, or capture ceiling, or may involve DSC, variable refresh rate, HDR, chroma subsampling, or colour depth that the card does not support for that path, or may exceed the USB connection's capacity for capture. Reduce to the known working mode, then reintroduce one feature at a time to identify which limit is reached.