Jump to content
  • 0

GigE Camera with Eclypse Z7


Udayan Mallik

Question

1 answer to this question

Recommended Posts

  • 0
Ethernet cameras, at least every one that I know of, rely on video compression, regardless of the video standard or resolution involved. I doubt that and SDK that comes with such a camera doesn't involve a proprietary loadable kernel module or compiled DLL that is very OS and hardware dependent and only provided in pre-compiled form. I could be wrong, but then you'd have all of the sources needed to figure things out.

1 GbE has a raw data rate that's close to, but not quite 125 MiB/s, provided that you are using packets with a very high payload to header ratio. If you are using standard packet types over an internet or intranet infrastructure then the 125 MiB/s maximum rate might drop to an average sustainable rate of 30 MiB/s or thereabouts. Still, even 30 MiB/s would support a reasonable frame rate, video depth, and resolution using a raw data format for some applications. This is true on a strictly hardware basis; software considerations add a whole new dimension to the design problem.

Video over LAN is certainly possible, depending on the requirements. It's the requirements, and details, that drive implementation and whether or not 1 GbE is a realistic media. The characteristics of compressed video is all by itself a limitation for many video applications.

[edit]

Oh... and then there's your platform to consider. Using a 1 GbE PHY connected to logic with direct access to external memory is one thing. Using a ZYNQ PS GEM is another thing entirely. The hardware is clunky... you can only connect it to memory via DMA. The memory,especially if the cores are running out of it, is not ideal for high performance applications. The software is messy at best.. and that's if you are using canned libraries and ethernet-lite. When that doesn't work good luck tuning all of the software dependencies to make it functional. You might think that an UltraScale ZYNQ would resolve all of those issues. I've run the ZC106 ZU7EV with Ubuntu on an SSD and it doesn't compare favorably to a Raspberry Pi 4 with the same amount of memory running off of an SD card if you try and run a few simple applications in multiple windows... much less try and run a USB webcam in a desktop. The Xilinx documentation for the project doesn't tell you about the lousy performance and how small webcam video is going to be. You have to go through the complicated steps of getting everything implemented to see the warts and limitations. This is complicated stuff with no easy one-size-fits-all off the shelf solutions. Please do prove me wrong...

In general a lot of so called demos for what you can do with the tools and FPGA hardware is a trap... unless all you want to just replicate the demo. Once you get into the nitty-gritty details, the finish line starts getting farther and farther away the more work you do. All such project require a bit of up-front experimentation and analysis before committing to an outcome. Edited by zygot
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...