Ezcap 284 HDMI, component, and composite video capture device

I used to need a video input on my computer fairly often, for example, to check satellite receivers, video cameras, and so on. I used a TV tuner for this, but after replacing the entire computer, I had to figure out how to do the same thing with the new one. I couldn't install a regular card anymore, and my priorities had changed a little, too. An HDMI input would come in handy now, but not many devices have one.
So when I came across a device like this, I decided to order it. In the end, though, things didn't turn out as well as I'd initially hoped—but more on that in the testing section.

First, a quick look at the packaging and unboxing.
The kit comes in a small black box made of thick cardboard.


The bottom lists the basic specifications; I'll repeat them later in more detail. There's nothing particularly interesting on the other sides of the box, except for a note asking you not to use the device to copy copyrighted material :)


Even during unboxing, the device makes a good impression: the accessories come in a separate little box underneath. That's quite convenient, though I think a pouch to keep everything in would be even better.


What's included:
1. Video capture device
2. Power supply
3. USB-to-mini-USB cable
4. Cable for connecting component, composite, and audio signals.
5. Two instruction manuals.
6. CD with software and drivers.
7. IR remote control.

It's very strange that there's no HDMI cable in the box. Since the device connects in line with the monitor, including an extra cable would have made sense.


I wasn't kidding when I said there were two manuals. At first, I thought they were the same manual in two languages, but it turned out to be much simpler: both manuals are indeed bilingual, with Chinese on one side and English on the other. One manual covers the device itself—its connectors, how to connect it, and how to install the drivers and software—while the other is for the software.
But as it turned out, they didn't help me.


The manual also lists the specifications, which say the device has:
Recording inputs: HDMI, component, composite, analog audio.
Outputs: HDMI, analog audio.
H.264 hardware compression.
Recording to a USB flash drive or SD memory card.
ExFAT format is not supported.
A long list of supported resolutions.
Maximum recording bitrate: 18 Mbps.
Maximum file size: 2 GB.
Approximate output video size: 140 MB per minute, or 8.4 GB per hour.


A little more about what's included.
1. CD. It contains the drivers, a copy of the manual in PDF format, and ARCsoft software.
2. Cable for connecting to the power supply or computer for data transfer.
3. Cable with RCA connectors.
4. The other end of the cable has a connector for the capture device. It's a somewhat unusual connector, and I doubt it would be easy to find a replacement cable if you lost this one, so take care.


Since I've mentioned the analog inputs, I'll digress briefly to explain them.
People often get confused about which connection type is better, especially component versus composite, since the names are so similar. So I thought I'd explain the difference.

Going back to the days of analog TV—all these interfaces date back to that era—there were at least five types of signal you could use to transmit a signal from a source to a receiver. I'll list them in order of increasing quality.

First is RF. It only makes the list indirectly, but since it was used with all sorts of Dendy consoles and the like, why not? Before DVB-T (T2) arrived, it was the worst connection option.

The second method is much better: composite video. For some reason, it's often called RCA, which is quite incorrect, since RCA is the type of connector used to carry it. To be precise, RCA is actually the name of a company (Radio Corporation of America). But this connector type is also used to carry other signals.
Formally, this signal type is called a "composite color television signal"; in other words, all the necessary information is carried over a single line.
The connector is yellow. Don't confuse it with the similar single orange connector: orange is S/PDIF.


The third type is also familiar to many: S-video. Unlike composite, it uses two wires. Technically, this is also a component signal, but that name never caught on, so people usually just call it S-video. It carries two signals: synchronization (using the composite video signal) plus brightness and color information. The quality of this connection is better than composite.
Plug on the left, socket on the right.
1, 2. Ground
3. Brightness and sync signal
4. Chroma signal.


Adapters that convert composite video to S-video are sometimes sold. The circuit is extremely simple, though for obvious reasons the quality will still be limited to that of a composite signal.


The fourth type is significantly better and is called YPbPr, though it is more commonly known as component video.
It uses three cables, carrying the following signals:
Y - brightness and sync signal.
Pb - blue color-difference signal
Pr - red color-difference signal

It connects via three color-coded plugs: red, blue, and green. But don't confuse it with RGB; that's something different.
This connection type can transmit a 576p signal without any problems, which is practically the maximum for CRT TVs.


The fifth signal type is not Scart, but RGB. The connector is included here for a reason, though. RCA connectors (the familiar yellow, red, and white plugs) are rarely used for RGB, since it requires four lines. And if you consider that every connection type except RF also needs a pair of audio connectors, that makes six cables in total.
In fact, the Scart connector also carries at least audio in and out, a control signal (for example, I once set it up to turn on an amplifier automatically when the DVD player was switched on), and other signals. Scart can carry RGB, composite, S-video, and YPbPr signals.

An RGB connection carries:
1. R - red
2. G - green
3. B - blue
4. Sync signal, usually provided by the composite video signal.


I've listed the four main signal types above, but there's actually at least one more, derived from RGB. It's derived from RGB because RGB is a “pure” signal, meaning it hasn't undergone any processing. The other types require a decoder to extract the RGB signal. That's why RGB is considered the highest-quality option. That's also why this connection made its way into the world of personal computers, though under the name VGA. Don't confuse the two: VGA can refer to a connection type or to an image resolution (640x480).
A VGA connection uses a 15-pin D-sub connector, which I'm sure many of you have seen.


Here's a table of image resolutions. You can see that VGA is 640x480 pixels. But this connection type had plenty of headroom, so it was also used at higher resolutions, with the name changing slightly.


Anyway, I've gone off on a tangent. The reason I brought up VGA is that professional monitors used five separate cables:
1. R - red
2. G - green
3. B - blue
4. Horizontal sync signal
5. Vertical sync signal.

So it's basically the same RGB, but with separate sync signals. The connectors are also different: instead of RCA, it uses BNC, which is better suited to high-frequency signals.


It's easy to remember:
1. Composite - one signal
2. S-video - two signals
3. YPbPr - three signals
4. RGB - four signals
5. VGA - five signals

As you can see, only composite can really be called a full signal; the others are component signals. But the term “component” has become firmly associated with YPbPr, and that's usually what people mean by it.

RCA connectors are also used for audio signals, with white and red plugs. By the way, if the connectors aren't labeled, it's easy to remember which color corresponds to which channel (at least, this is how I remember it): Red means Right.
You may also see a separate RCA connector on satellite receivers, but that carries an auxiliary control signal.

That's it. Sorry if that was a lot to take in; I'll get back to the review.

1. A small remote with buttons for recording video and taking screenshots, selecting a recording resolution of 720 or 1080p, and choosing the video input. By default, the device starts up on the HDMI input.
2. The remote is powered by a 2025 battery.
3, 4. Power supply with a European plug. It's basically a standard power supply with a USB connector, a 5 V output, and a current of up to 2 A.
If you want—or need to, when recording to a computer—you can power the device from the computer's USB port. I also tried powering it from a tablet. It struggled, but it started up.


I opened up the power supply to inspect it, especially since it was easy to do.


Externally, the board is the "economy" version, with no input or output filters.


1. There is a fuse at the input, and the input capacitor is 10 µF, which is fine for the stated 10 W.
2. The SD6830 PWM controller IC is rated for up to 12 W, according to its datasheet, so no issues there either.
3. The interwinding noise-suppression capacitor is the correct type.
4. The output has a diode rated for a maximum current of 3 A and a 1000 µF capacitor. That seems fine, although for a current of 2 A it would be a good idea to use a pair of these diodes.


The board hasn't been cleaned, although the soldering is relatively neat. You can see resistor footprints under the USB connector, which tells us that this power supply is a standard USB "charger".


Now I've finally reached the subject of this review: the video capture device itself. I don't want to start by praising it, but I have to mention the matte case. To be honest, I'm a little tired of glossy finishes that get dirty instantly. This case is practically fingerprint-proof, and I wanted to point that out.


It's about the size of a small TV box, maybe a little smaller, but overall very similar.


There's an infrared receiver for the remote control on the front. The record button and operating-mode indicator are on top. Since most of the case is made of translucent plastic, you can't see the receiver or the LEDs.


The power supply connector is on the left. It's also used to connect the device to a computer, in which case it handles both power and data transfer.
On the right are a USB port for a flash drive and a memory card slot. The slot has a push-to-eject mechanism, so the card sits completely flush and is barely noticeable when inserted.


There are more connectors on the back.
1. HDMI output for connecting a monitor. The device can work completely standalone or in "pass-through" mode.
2. Analog audio output.
3. Analog audio input. As far as I understand, this input can also be used to connect a microphone.
4. Connector for analog video and audio inputs.
5. HDMI input.


The store's website and the manufacturer's website both show typical connection diagrams for HDMI and analog signals. Oddly, the box is drawn "upside down" in both diagrams. I have no idea why.
HDMI


Component and composite.


The bottom is perforated, presumably for cooling. But since there are no feet and the box can only be placed bottom-down, these holes are pretty much pointless. On the other hand, it barely heats up, so the perforations are probably just decorative.


When power is applied, the blue power indicator LED lights up, along with the green LED indicating the selected input and its status.
Since there are two operating-mode indicator LEDs but three inputs, the manufacturer uses a sort of "code" to indicate the operating mode.
Bottom LED — HDMI input
Top LED — component input
Both at once — composite input

If the input indicator LED is blinking, it means the input is enabled but there's no signal. When a signal is present, the LED stays on.


Pressing the record button also lights the red LED. Pressing it again stops the recording, but the LED doesn't go out immediately; it blinks for a few seconds first to indicate that the recording is being finalized.
During testing, I noticed that once the recording seemed to be running, but the resulting file was zero bytes. So I'd recommend using a flash drive with an activity indicator. It will blink while recording, confirming that the recording is actually in progress. This only happened once, but it's better to be prepared for it.


Before testing, I also took the device apart because I was really curious about how it was built and what components it used. I was sure about which processor was inside, but knew nothing about the rest.


I'll point out right away that it has a built-in battery. Presumably it's used to power the built-in clock, especially since the file timestamps I got were always 9 hours ahead of the actual time. But how to set the time on this box is still a mystery to me.
The battery is not only connected via a connector, but also fitted into slots in the top cover, whereas it is usually just stuck on with double-sided tape. A small detail, but it shows that the manufacturer pays attention to its products.


The board is genuinely well made, with neat, clean soldering, and I have no complaints about the assembly. I was a little surprised, though, by the large number of components. When I first opened the box, I expected to find a couple of chips and a few resistors and capacitors inside, but it looks like there are at least as many components as in a TV box.


Both the HDMI input and output connectors are protected, which is also a nice touch.


A small heatsink is glued to the processor that handles the signal. I wouldn't say it gets hot, though: even after an hour of operation, the device's case is barely warm.


Besides the mode indicator LEDs and the record button, I found another mysterious button inside (between the LEDs), whose purpose isn't mentioned anywhere.


I couldn't find any information on the installed IT9910 processor, but I did find out a few things about the other components. Here goes:
1. IT6613, an HDMI 1.4 transmitter, to which the monitor is connected.
2. ADV7180, an ADC from Analog Devices that digitizes analog signals before they are processed by the main chip.
3. WM8960, an audio codec with a built-in Class D amplifier rated at 2 × 1 W.
4. S/PDIF transmitter WM8804


1. EEPROM, presumably for storing the basic settings.
2. Flash memory for storing the firmware. The device supports firmware updates, by the way, and there's even one on the manufacturer's website. I used it while trying to connect the device to a computer. Updating the firmware is very simple: copy the file to the root directory of a flash drive, insert the drive into the device's port, and power it on. The recording LED flashes for a few seconds, and that's it—the update is complete.
3. A couple of chips I couldn't identify near the USB connector and memory card slot.
4. DC-DC converters for powering the electronics. At the top, I spotted a noise filter that the signal lines of the microUSB connector pass through.


Given the number of components on the top of the board, I expected the underside to be completely bare, but it turns out there are components there too.


The memory card slot is additionally secured with glue.


There are fewer components on the underside, but I'll list them anyway.
1. CAT9883CCQ-150/ITE. I tried to find a datasheet but couldn't; the search mostly turned up sellers on Chinese websites, and no datasheet.
2. I noticed a fairly large copper pour for cooling beneath the main processor, so I decided to point it out.
3. TDA2822, another audio amplifier, this time without a codec.
4. 74HC125, four buffer elements in one package.


That's all for the description; let's move on to connecting and testing it.
Standalone operation was straightforward—well, almost. The one problem was that this device is incompatible with the Beelink Sea TV box.
I connected the two devices, but recording stopped after a minute when the TV box displayed a notification. I tried switching between HDMI 1.4 and 2.0 modes and changing various settings that should theoretically control this, but no luck: one minute, then it stopped. After that, the TV box turned its output back on and the whole thing started over.


Otherwise, I liked how it worked standalone: plug it in, press the record button, and that's it...

But when I connected the device to a computer, that's when I found out what a real problem was.
I tested it with four computers at home—or, more precisely, three computers and a Windows tablet.
Two computers running Windows 7 x64 wouldn't recognize the device at all; it just showed up as an unknown device. Naturally, installing the included driver didn't help at all.


According to the manufacturer's instructions, it should have been very simple. A new device named IT9910 should appear in Device Manager. Then you just click the driver installer, and that's it.
With almost no hope left, I connected the video capture device to another computer running Windows 10 x64, and sure enough, it appeared as shown in the instructions. But the driver still didn't help. That is, The installation process ran, but nothing changed.
After that, I tried again on a tablet running Windows 8.1 x32. It was the same as in Windows 10: the device was detected, but the driver could not be installed.


With absolutely no hope, I went online looking for a solution. After digging through a pile of forums and reading lots of repetitive, pointless advice, I came across a video where a guy explained how to connect this capture device. He talked a lot, often about things that were completely off topic, but I did get some useful information from it. He also said he had problems with Windows 10.
But I was more interested in the driver installation process. The video’s author used a video editing program that installs its own drivers for these devices. He extracted the drivers from the software and posted them separately, for which he has my heartfelt thanks.


Since the video is quite long and it’s not very convenient to hunt for information in it, I’ll show the process in screenshots.



The device description mentioned OBS. No, not “One Old Lady Said,” but a genuinely useful program. It’ll be included in the supplementary materials, and it doesn’t require registration since it’s free.
At the bottom, find the Sources window and click the plus sign in its lower-left corner.
Select the highlighted item.


You can just click OK here, or change the device name.


After that, the name entered above will appear in the Sources window. Click the gear at the bottom of this window to open the settings. Select our input device at the top, and the image appears right away. The computer has two monitor outputs: the main one is connected via VGA, and the capture device via HDMI.


Everything works. The program window shows the second desktop; I dragged the File Explorer window over there.


The program has a ton of settings, and I won’t even try to cover them all. I’ll just note that it struggles quite a bit at 1080p: there’s about a one-second delay and a low frame rate. I’ll probably have to look into the codec.
By the way, I never did find a setting to adjust the time on this box.


Above, I mentioned that the package included software for working with this device. Unfortunately, with these drivers, the software can’t see the capture device—or rather, it thinks the device is busy :(


Almost forgot: if you plan to record video with sound, you’ll need to select the corresponding audio source as the default in your computer’s sound settings.


Now for some tests. I was interested not only in seeing how this device works, but also in comparing it with the Beelink Sea I TV box, which offers similar functionality.
To do that, I recorded two test files, which included:
1. A plain desktop, to assess the quality of a static image.
2. Using File Explorer with scrolling, to assess performance with motion. This test is very demanding for capture systems, since the scrolling window needs the maximum bitrate while the rest of the screen remains static.
3. Playing a video clip. To avoid limiting the quality at the source, I used a 4K demo file with an 80 Mbps stream, but output the video at 1920x1080.
4. Playing a test file with a calibration chart that includes a frame counter, as well as static and moving sections.
5. Checking that the video is recorded “pixel for pixel.”

In the end, I got two files of roughly the same duration. In the screenshots below, Beelink is on the left and the device under review is on the right.
Beelink saves the file in a .ts container (transport stream), while the device under review saves it as .mp4.
Beelink’s file was smaller despite being longer, which is significant. Beelink records at 60 (59.94) frames per second, while the device under review records at 29.97—half as many.


Although the devices use different containers to save video, both use Advanced Video Codec (AVC). But there’s also a fairly significant difference:
Beelink uses the High 4.2 profile for compression, while the device under review uses only Baseline 4.0, an older version.


It’s also worth noting that ezcap records an audio stream at 1534 kbps, while Beelink records at 34 kbps. However, Beelink’s audio file is not 45 times smaller, but only 6 times smaller, because ezcap uses variable bitrate compression, while Beelink uses constant bitrate. So it turns out that ezcap records better-quality audio.



Next came the image tests. The screenshots have been compressed; the actual image looks a little better.
First, I tested a static image of the desktop—more specifically, a folder on the desktop. All screenshots were taken not during recording, but while playing back the recorded video, so they show the final result.
First, ezcap.


Now Beelink.


And, of course, a side-by-side comparison. Beelink is on the left (as in the other comparisons), ezcap on the right. Beelink’s image is brighter, but the text sharpness is almost the same.


Now a test with the same window, but in motion. I scrolled the window while recording and took a screenshot during playback. This isn’t just the first screenshot I happened to get: I tried three times, and the result was the same each time.
ezcap.


Beelink.


Comparison. Here you can see that Beelink performs much better with moving images. I never managed to get a blurred image, just as I couldn't get a sharp one with ezcap.
I also noticed that Beelink rendered text more sharply in motion than in a still image. The processor boosts the bitrate, and does it very quickly.


The video recording test went a little better: both devices produced almost identical results.
ezcap


Beelink.


Comparing the two, I can say that image sharpness is almost the same on both devices, perhaps a little lower on ezcap. Beelink, meanwhile, has lower contrast, making the image look more washed out.


Test chart. I didn't run a side-by-side comparison here; I'll just show two separate screenshots. By and large, both devices produced roughly the same result in this test.
ezcap.


Beelink


And finally, a test of pixel-perfect output—or, more accurately in this case, input.
ezcap failed this test. The horizontal “comb” came out correctly, but the vertical one was wrong. Below is a crop from the full video.


Beelink passed this test without any problems.


Here are a couple of sample videos. They're included in the supplementary materials, but you can judge some things from these alone. To check pixel-perfect output, you'll need to download the file, since YouTube ruined both results.
ezcap


Beelink. By the way, Beelink records audio at lower quality—or rather, it compresses it more.


One more Beelink video, or rather, a demonstration of what only devices like this can do.


I also tested the composite input, though the test was mostly for show, since I used a video camera—an original-brand one, but quite old.
This uses the Baseline 3.0 profile.


The lens isn't quite in focus because I discovered an odd quirk. The capture device can convert the analog input signal to digital and send it to the HDMI output. But the monitor said the signal was outside the supported range, and that was that.
I tried changing the resolution with the remote, but it made no difference.
I could try configuring it in the software, of course, but that's extremely inconvenient.

When testing the analog inputs, I was mainly interested in how color was reproduced. I didn't see any particular problems here; this is how the camera works with the surveillance system, too.
I was also interested in how the capture device handles a poor-quality signal, specifically loss of sync—for example, when copying from a VCR. I know this is obsolete technology, but still. A long time ago, I copied tapes to a computer and even built a special board to restore the sync signal. I came across it again quite recently. But I couldn't find a source to test with, sorry.

Here's what video from the surveillance camera looks like. It's a used Panasonic camera, and the focus isn't set very well. I have video footage too and could upload it, but I don't see the point.


I also made a short introductory video; it might be of interest.


That's all. Time to sum things up, starting, as always, with a quick summary.
Advantages.
1. Can record from an HDMI input.
2. Can operate completely standalone.
3. Composite and YPbPr inputs.
4. Audio output and microphone input.
5. Excellent build quality.
6. Good audio quality

Disadvantages
1. Driver issues, and consequently software problems.
2. Image recordings of text aren't particularly good.
3. Uses the Baseline 4.0 recording profile rather than the better 4.2 profile.

My opinion: I tried to compare two devices with similar capabilities, and honestly, I can't say that I prefer one over the other. Beelink renders text very sharply, but has some color reproduction issues. More specifically, it has no true black or white; instead, they're dark gray and light gray. You can see this in the screenshots.
ezcap produces a nice image, with both black and white, but the fonts blur, quite noticeably. ezcap also failed the pixel-perfect test.
But ezcap has a couple of features that give it an edge over Beelink. First, it's almost completely standalone; second, it has analog inputs. And let's not forget the price: since the Beelink is also a TV box, it costs more.

Ezcap does have one more advantage: you can connect it to a computer, which significantly expands what the device can do. But the driver hassles ruin it completely. Not to mention that I couldn't get it working with Windows 7, although it works quite well with Windows 10.

There's also the opinion that ezcap records in 1080i rather than 1080p. The lack of proper pixel-for-pixel output and the blurred text in the test seem to support this.
As for the Beelink, in my case the playback may have been incorrect because of the .ts format, which is why the screenshots show a reduced brightness range. In any case, I've uploaded the original video files for anyone interested.
I think the Beelink is better suited specifically for HDMI capture, while ezcap is better for analog signals, since the Beelink has no analog inputs and is also a TV box, not a capture device.

That about covers it. If you have any questions, I'll be happy to answer them. I hope this information was useful.
Additional materials:
1. Drivers, software, manual - link
2. Sample ezcap recording - link
3. Sample Beelink recording - link

The store provided coupon LXXV22711, which brings the price down to 53.68. When I was looking for this device on AliExpress, the lowest price I saw was 59.