[YuzukiHD] has provided files for anyone that wishes to build their own HDMI capture card at home. The design is known as the Yuzuki Loop Out HDMI Capture Card PRO, or YuzukiLOHCC PRO for short.
The build is based on the MS2130, a HD video and audio capture chip that’s compatible with USB 3.2 Gen 1. We’re pretty sure that’s now called USB 3.2 Gen 1×1, and that standard is capable of transfers at up to 5 Gbps. Thus, the chip can support HDMI at up to 4K resolution at 60 Hz depending on the exact signals being passed down the line. It’s compatible with YUV422 & MJPEG modes and can be used with software like OBS Studio and FFmpeg. The board itself is relatively simple. It features an HDMI In port, an HDMI Out port, and a USB-C port for hooking up to a computer for capture.
HDMI capture cards can be expensive and fussy things, so you may find it pays to roll your own. Plus, being open sourced under the CERN Open Hardware License V2 means that you can make changes to suit your own use case if you so desire.
We’ve seen some other hilarious video capture tricks over the years, such as a convoluted rig that uses a SNES to turn a Game Boy Camera into a usable webcam. If you’ve got any such madcap hacks brewing up in your lab, be sure to let us know!
The elephant in the room: Does it work with HDCP protected HDMI signals?
from the readme in the linked GitHub repo:
Integrated pre-programmed HDCP keys for Loop Out
Be nice to have a datasheet, taobao wants a login, I CBA to go further either.
Cant find it on AliExpress but semi OT, it’s interesting how cheap the HDMI to CSI boards have got and just how a cheaply a PiKVM project can be built for now
If the pre-programmed HDCP key is not on a block list then yes, if not then no. So all you need to do is contact Intel’s daughter company (Digital Content Protection LLC) sign a billion NDA documents in blood, for a valid HDCP 2.3 key, which they would reject for this purpose.
(But 4K video was added with HDCP 2.2 and there is a reason why HDCP 2.3 exists)
I should probably add that there is a license fee involved in being allocated a valid key.
Gross!
yes, but if a single key is reverse engineered then certainly this opens the flood gates to the protection being broken ? This DRM architecture has always seemed dangerously fragile in that regard.
Team HDCP wants their tithe, but my understanding was that(at least for smaller volumes) you could get HDMI/HDCP-related chips that were keyed up by the vendor on your behalf; so (while you ultimately pay one way or the other) it’s at least possible to get appropriately functional chips without having to do the signing in blood yourself.
Am I mistaken; or is this the case?
The MS2130 (from the manufactures website) has support HDCP 1.4.
1080p possibly yes, since the HDCP 1.x master key was published.
2160p definitely no (because 4K requires HDCP 2.2 it higher)
From the MFR site they seem to have other devices which can convert HDCP ?
Im no expert on this.
Easy bypass: use an HDMI splitter connected to a TV/monitor that supports HDCP. Then on the second output you have you capture device. The TV/monitor will send the right flag back to the source device and the signal will become readable by the capture device. I may know a person that has used this method to rip content they bought on certain streaming services for “backup purposes”.
That’s how I do it. It works well.
Lmfao 🤣 🤦🏽 at HDCP. I knew there has to be an easier way… worst case scenario, tear up the advanced TV and just dump the output LCD trace signals to rip content. I’m sure pro pirates does this.
Know quite a few people who record HDCP protected streams like this. Cheap and simple. Not all splitters will work, but many do. I know one brand available pretty cheaply from a local electronics/hobby store here in Australia that works flawlessly. Connect everything then turn on the splitter and TV first then lastly power up the device on the second output of the splitter.
Is it really open source when the part (MS2130) is so new that it can only be got from the manufacturer MacroSilicon/UltraSemi (I’m guessing with an NDA, and a minimum order size of 1K units).
http://en.macrosilicon.com/
http://www.ultrasemi.com/
Or it is clever viral marketing ?
Open Source with a big black chip in the middle.
Open Hardware should be renamed Open Schematics.
Ayyy this is a pretty nice follow-on from MacroSilicon to the gen 1 Can’tLink cores that only did MJPEG.
Will be trying to build one of these and take it for a spin, and pretty sure you can sideload a key you suck outta a different kit, which I think is available in a gen 1 cant’link memdump, haven’t figured those out yet though.
Would be interesting to run this program against it: https://github.com/BertoldVdb/ms-tools
If the core is similar to ms2109 it allows running code on the chip.
8051 is supported by SDCC. Any exemple around for that chip?
CNX has more infos: https://www.cnx-software.com/2022/10/09/10-hdmi-to-usb-3-2-video-acquisition-board-ms2130/
You need to read the conflicting fine print. It says max loop-out res is DCI 4K, but then it says max video input is UHD 30Hz, but then it says the max USB capture is FHD 60Hz.
That’s a lot of alphabet soup to simply say “It’s a 1080p capture device”. And the spec for USB2.0 is 10Hz, which is pretty useless. Those $15 cheap USB capture devices will do 1080p at higher frame rates.
This is not a 4K capture device, it’s a 1080p capture device that happens to grok and downsample 4K.
I also have a hunch that there is something hinky with the HDMI2.0 and HDR, as in you can’t capture HDR, only pass through.
“Those $15 cheap USB capture devices will do 1080p at higher frame rates.” Yes, with terrible-looking compression. With 5Gbps of bandwidth, this can do *uncompressed* 1080p, which will be a massive improvement.
So is the USB capture doing full uncompressed in MJPEG?
Yes, that is true 1080P overall for usb capture. But the reloop to 4K is good, they have good uses (specially for me) so all is not doomed at 1080P.
I wish something like that would exist for capturing or converting VGA 720×400 @ 70Hz MS-DOS resolutions, available devices usually produce black screen with such parameters. It would make retrofitting moderns screens to old devices easier.
Something like a MS9288C (VGA-to-HDMI conversion chip) might work, but I’ve no idea.
I’ve collected a few old pieces of gear I need to string together to try. I have a Panasonic AG-1980 Super VHS VCR, an industrial looking metal box with an LCD on it that takes S-Video in and upscales to 640×480, 800×600, or 1024×768 VGA out via HDB15 connector, and something else to digitally capture VGA. The stuff is all in a box somewhere and I don’t recall what all the makes and models are of it except the Panasonic VCR.
If the video scaler gets boogered by less than stable VHS tapes (deliberately that way or not) I have a Panasonic DMR-ES15 which when set up properly with the remote works as a time base corrector on S-Video passthrough. I also lucked into a bunch of extra thick, studio grade S-Video cables at a thrift store.
I was going to digitize some old tapes and see if I could upscale and make the video look better, but life in general, other projects, and just not wanting to have to wait on the pesky 1:1 video time to real time ratio for the digitizing have gotten in the way.
i’m just curious, what’s outputting such weird resolution?
720×400 @ 70Hz is standard resolution for MS-DOS text mode or BIOS menus in old computers, so most of the vintage PC hardware uses it (386, 486, Pentium 1 era), it would also be used in older industrial PCs like Vortex86, basically in everything that has only VGA connector. But when you go to graphics mode/Windows etc. you can use standard 640×480 60Hz, which can be easily converted to HDMI.
i was familiar with the 640×480 60hz resolution but i never wondered about text-mode.
thanks for sharing!
Some time back, we used to use the Sony DSC-1024HD Digital Scan Converters for VGA conversions into TV signal for broadcast. Worked very well. Specs can be seen here:
https://budgetvideo.com/catalog/converters/scan-converters/sony-dsc-1024hd-scan-converter
A few used working ones floating around eBay as low as US$75.00.
https://ebay.to/3D3F70J
Pi 4B ARM + RPiOS 64-bit nice … but no PD100W type-c power/hub station connection.
Have 3 at $55, Amazon. Now ~$150.
Our Lenovo flex 3 MediaTek ARM mt8183 chromebook cost $109 Best Buy. Has PD199W usb type-c, one usb type A, nice 11.6 screen, keyboard, mouse, and battery backup. But spyware? crippled gLinux ChromeOS.
Pine64 sbc appears to have PD100W.
Will we soon see a ARM sbc from China with only PD100W and one usb type-A connectors.
Running RPiOS or Ubuntu ARM server? At fraction of the price of a Pi 4B?
gLinus Google Linux … or guano Linux?
Experimenting with porting x86 machine code to ARM platform using gcc c to produce machine code.
c prolog/epllog c calls foo(int x …. rejected Intel MCS BASIC-52 GOSUB or GOTO has no defined prolog or epilog conventions.
.
Second, after ‘hello world’ to see if Ubuntu x86 code ports to ARM Raspberry Pi 4B.
//gcc -o ee 64bit10.c
#include
int main()
{
unsigned char g[8] ={0x00,0x00,0x01,0x00,0x0,0x0,0x0,00};
int i;
unsigned char gg[] ={0x80, 0x40, 0x20, 0x10, 0x08, 0x04, 0x02, 0x01};
long int n;
long int nn=0;
for(int j=7; j>-1; j–)
{ n = 0 ;
for (i=0; i<8; i++)
{ n=n<<1; // n<<1 failed.
if (gg[i] & g[j]){ n++;};
}
nn = nn<<8;// multissply by 512
nn = nn + n;
}
printf("\n End conversion nn = %ld\n\n",nn);
for (int i=0; i<8;i++){
printf("%x ",g[i]); }
printf("\n\n");
return 0;
printf("\n End conversion nn = %ld\n\n",nn);
for (int i=0; i<8;i++){
printf("%x ",g[i]); }
printf("\n\n");
return 0;
}
Code must compile into ARM machine language.
System calls not permitted here.
Other than printfS, of course. :)
Goto bad, use ComeFrom.
Your usb definition is wrong, they just got updated 😉