http://www.studiodaily.com/2018/08/magewells-latest-encoder-converts-hdmi-video-to-ndi-streams/

I was hoping Magewell could bring the price of NDI converters down a bit, but at $999 it is still quite expensive to make the leap to NDI...
What do you think?

It's strikes me that the cost differential for different models is slight. For example, a 4k version vs HD. For a given I/O arrangement the bill of materials is very similar.

The BirdDog Studio and Mini makes sense, since there's different I/O.

As long as these things are FPGA-based the cost is not going to fall that dramatically. That's expensive hardware to acquire and program.

As with capture devices, including hardware for scaling and frame-rate conversion adds value.

I don't see the prices falling substantially until there are ASICs to replace the FPGAs.

Some interesting test results performed today using different types of video capture connections - USB, PCIe, Thunderbolt and NDI. The devices included:
1) Magewell Pro Capture Quad HDMI card mounted in an external Zotac AMP Box Mini Thunderbolt 3 PCIe enclosure.
2) Magewell Pro Capture All-In-One card installed internally in a PCIe slot on the Asus Z170 motherboard.
3) Magewell Pro Connect 4K NDI converter.
4) Magewell USB Capture HDMI Plus via a USB3 port.
The signal source was a Blackmagic Mini Monitor 4K card, connected via a 4-way HDMI DA to the capture devices, and outputting a 1080p60 test signal with in-vision frame numbers and fairly complex video content - in order to give the NDI converter some work to do! You can see that the Pro Convert dashboard shows an outgoing data rate of 185Mb/s.

ProConvertDashboard1080p60.PNG (76kb) downloaded 7 time(s).

The vMix test setup was simply those four camera inputs, a text overlay and an image multiview showing a quad split overlay of the four main sources.

QuadTest1-4Oct18.PNG (750kb) downloaded 30 time(s).

What you see is that the NDI converter holds its own against the two PCIe capture cards, with the USB device appearing to be one frame behind. However, in the snapshot frame grabs produced by vMix, all four sources are typically in sync, all showing the same frame.

Sample "image output snapshot" file vMix Capture 04 October 2018 10-50-39.png (3,794kb) downloaded 10 time(s).

Changing the order of the inputs did make a slight difference. By putting the NDI & USB devices as the first two inputs, with the PCIe devices as 3 & 4, all four devices were showing in sync, both in the GUI preview and in the output snapshots. So it may be worth arranging your inputs with what you think may be slower devices as the lower numbered inputs. However, in these tests it was really only the USB device that appeared slightly slower than the PCIe cards, with the NDI converter providing similar overall latency to the PCIe cards.

QuadTest2-4Oct18.PNG (748kb) downloaded 31 time(s).

Sample "image output snapshot" file from the revised input order vMix Capture 04 October 2018 10-35-57.png (3,970kb) downloaded 9 time(s).

I don't have either of those two cards, but I will do a similar test at some point, probably with a Blackmagic Mini Recorder included in the line-up.


I guess that's how vMix works internally - with latency between input and output.

I usually has exactly same frame on in/out. something is wrong, maybe different frame rate between vmix and monitor, maybe is good to look into this and fix for future tests.

About blackmagic quad vs magewell,i got less latency on blackmagic. Depend on type of input. This is because i ask to add on future comparison.

Regards

Jose.

... maybe a screenshot problem?, if you take a photo to the monitor has same latency difference?. Can you test total latency between input and output. You need a hdmi monitor on your source and same model of monitor on the hdmi output of computer, and take a photo of both.