Closed mxochicale closed 2 years ago
Nhat commented that adding other devices (e.g. GE Vivid IQ) in the config file will be useful for future work. That said, we need to add a config file with such features of different US devices, starting with Venue go and philips devices.
cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')
and Bus 002 Device 003: ID 1bcf:2c99 Sunplus Innovation Technology Inc. MiraBox Video Capture
There are still few things to do in this ticket:
QObject::moveToThread:
$ python framegrabber_capturing_video.py --frame_width 1920 --frame_height 1080 --frames_per_second 60
fps: 60.0
resolution: 1920.0x1080.0
mode: MJPG
Buffer size: 1.0
QObject::moveToThread: Current thread (0x55c501d905e0) is not the object's thread (0x55c5021891c0).
Cannot move to target thread (0x55c501d905e0)
QObject::moveToThread: Current thread (0x55c501d905e0) is not the object's thread (0x55c5021891c0). Cannot move to target thread (0x55c501d905e0)
* [x] check other brands of framegrabbers
* [x] test other resolutions and framerates
640×360 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 640×480 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 848×480 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 960×540 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 1024×768 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 1280×720 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 1280×1024 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 1600×1200 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 1920×1080 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps 1920×1200 | 15, 23.97, 24, 25, 29.97, 30, 50, 59.94, 60 fps
* [ ] test captured aspect ratios https://en.wikipedia.org/wiki/Aspect_ratio_(image)
* [x] test color codecs RGB24, YUY2, Y411, Y800, Bayer8
Capture 4K in perfect fidelity, or use hardware scaling to meet your application needs at any resolution.
- HDMI to USB 4K capture card £504.95
- Color space YUY2 (4:2:2); Color spaces NV12, YV12 and I420 (4:2:0)
- LATENCY: "The hardware latency of the AV.io series is approximately 2 frames. Typically that’s somewhere in the ballpark of 66ms. The software you’re using and even the display you’re connected to will also increase the potential latency." "Basically there is a 2 frame hardware buffer and after that the AV.io sends data as fast as it can. There should however be a reduction in latency from using a 60fps source as opposed to a 30fps source." https://www.epiphan.com/forum/t/latency-for-av-io-4k-and-av-io-hd/1698
- Operating System support: Windows 8.1, Windows 10, Mac OS X 10.10 and up, Linux distribution with kernel 3.5.0 or higher
- Online help: https://www.epiphan.com/userguides/avio-4k/Content/Home-AVio4K.htm
- User guide: https://www.epiphan.com/userguides/pdfs/UserGuide_AVio4K.pdf
- Drawings https://www.epiphan.com/brochures/capture-card-family/AV.io-4K_Mechanical-drawing.pdf
- Linux AV.io configuration tool Operating system: Ubuntu 16.04 LTS 64-bit and similar: https://www.epiphan.com/downloads/avio/ConfigTool_4.0.0_Linux.tar.bz2
- HDMI cables: As a practical test, try using a shorter 3ft or 6ft cable in the workflow to see if it corrects the issue. If the signal is seen with a shorter cable, it could be signal degradation. I’d recommend using an active/directional HDMI cable. https://www.epiphan.com/forum/t/av-io-sdi-compatibility-with-blackmagic-hdmi-to-sdi-converter/3588
- VIDEOS: https://www.youtube.com/watch?v=AVBRCF37Qb4 https://www.youtube.com/watch?v=nn36cT14iMg
- more https://www.epiphan.com/products/avio-4k/
- Resolutions, capture rates and aspect ratios
The simplest way to capture HDMI, VGA or DVI video sources at resolutions up to 1080p.
https://www.magewell.com/products/usb-capture-hdmi-4k-plus
- Is there any place where I can see how the DFG/HDMI is used in Linux (in https://github.com/TheImagingSource)? No, because it is used like an USB camera by v4l2. The DFG/HDMI is v4l2 compatible Therefore, it will work with guvcview, qv4l2 and so on.
- Has the device been used in Ubuntu distros? In Ubuntu 20.04 and Debian.
- Can you share the schematic of HDMI-to-USB converter as I don't see it in https://s1-dl.theimagingsource.com/api/2.5/packages/publications/schematics/sdconvfg/a572e37f-6556-5448-b391-9ff8211888ac/sdconvfg_1.2.en_US.pdf? Sorry this is not available
- It would also help if you provide details on latency performance of the DFG/HDMI . There are no details available for this, sorry for the inconvenience
Hi @huynhatd13
would you please share the "User Manual for the device you used for data collection of the echoes, we currently have this:
us_device_brand: 'GE'
model: 'Venue Go'
software_version: 'R2'.
serial number:?
Transducer: 'GE Convex probe C1-5-D'
Transducer serial number: ?
I would like to cross-check these features for the display:
Size 15.6 inch wide
Resolution 1920 x 1080
pixels (WSXGA)
Number of Bits Min 8 bits per color
Aspect Ratio 16:9
The user manual would help us to see the minimum/maximum settings for depth and field of video of the US image in cardiac mode to create appropriate mask, and post-processing imaging pipelines.
Thanks, Miguel
NOTES:
Nhat and I tested the following resolutions using Mirabox framegrabber with OBS in his machine!
No frame
No frame
^CTraceback (most recent call last):
File "framegrabber_capturing_video.py", line 63, in <module>
main()
File "framegrabber_capturing_video.py", line 60, in main
CaptureVideoTests(args.frame_width, args.frame_height, args.frames_per_second)
File "framegrabber_capturing_video.py", line 47, in CaptureVideoTests
if cv2.waitKey(1) & 0xFF == ord('q'):
KeyboardInterrupt
[ WARN:0@11.648] global /io/opencv/modules/videoio/src/cap_v4l.cpp (801) requestBuffers VIDEOIO(V4L2:/dev/video2): failed VIDIOC_REQBUFS: errno=19 (No such device)
(rt-ai-echo-VE) mx19@sie133-lap:~/repositories/echocardiography/scripts/hardware$ python framegrabber_capturing_video.py --frame_width 1023 --frame_height 768 --frames_per_second 60
[ WARN:0@0.057] global /io/opencv/modules/videoio/src/cap_v4l.cpp (889) open VIDEOIO(V4L2:/dev/video2): can't open camera by index
Traceback (most recent call last):
File "framegrabber_capturing_video.py", line 63, in
SORTED, I think was a problem with the full-drive!
From our meeting, Alberto suggested to made use of adultecho
as a good option for our work link.
Close this one as it is resolved in the docs
🚀 Feature
This ticket will help to track software/hardware integration of models and pretus plug-ins with devices: US device (e.g., software version, image resolution, image anonymization, depth, Field of View, etc), cardiac probe (model, features, etc) and framegrabber (e.g., framerate, performance, etc).
Pitch
Documented integration of hardware and software.
Alternatives
Additional context