So this guy is using the 1080i out via SDI from the FX6 for live broadcasts for Al Jazeera.
https://twitter.com/stephengreaves/status/1394105437423869953
Has anyone tested to see if the FX6 1080i output is a 'proper' interlaced signal like god intended, or is it more like the pseudo-interlaced you can do in post (from a 50p/60p source)? How can I test this (other than by waving the camera around and looking for artefacts)?
Either way, I'm sure Mr Greaves is fine given the lack of motion in a talking head shot. I'm thinking about doing something a bit more demanding - sending the 1080i signal to an external recorder and using it for a whole news package. Apologies if anyone finds discussion of interlaced video upsetting.
Here is what Alister Chapman says about "basic conversion" vs real interlaced:
"A real interlace image is created by using pairs of lines from the sensor that are – inter – laced.
So with a real interlace signal the first field contains lines that are created by combing these line pairs
2+3 4+5 6+7 8+9…….
The second field contains lines made from the following line pairs.
1+2 3+4 5+6 7+8…..
Overlapping line pairs is necessary to ensure that the two fields blend into each other to ensure there is no temporal or spatial aliasing (it also means the resolution when using interlace is always slightly lower than progressive).
If you take a 50p/60p file and drop it onto a 50i/60i timeline or do a basic conversion from a 50p/60p file to 50i/60i file this overlapping/interlacing does not normally occur so where there is fast motion you will sometimes get temporal aliasing which will appear as combing or tearing of the image. So it will be 50i/60i but in some cases it may have artefacts that would not be seen from a true interlace camera.
I have not yet tested whether this same artefact is present in the FX6’s interlace output as I suspect this will be a simple 50p/60p to 50i/60i conversion so the same artefacts are possible on the cameras output."
https://twitter.com/stephengreaves/status/1394105437423869953
Has anyone tested to see if the FX6 1080i output is a 'proper' interlaced signal like god intended, or is it more like the pseudo-interlaced you can do in post (from a 50p/60p source)? How can I test this (other than by waving the camera around and looking for artefacts)?
Either way, I'm sure Mr Greaves is fine given the lack of motion in a talking head shot. I'm thinking about doing something a bit more demanding - sending the 1080i signal to an external recorder and using it for a whole news package. Apologies if anyone finds discussion of interlaced video upsetting.
Here is what Alister Chapman says about "basic conversion" vs real interlaced:
"A real interlace image is created by using pairs of lines from the sensor that are – inter – laced.
So with a real interlace signal the first field contains lines that are created by combing these line pairs
2+3 4+5 6+7 8+9…….
The second field contains lines made from the following line pairs.
1+2 3+4 5+6 7+8…..
Overlapping line pairs is necessary to ensure that the two fields blend into each other to ensure there is no temporal or spatial aliasing (it also means the resolution when using interlace is always slightly lower than progressive).
If you take a 50p/60p file and drop it onto a 50i/60i timeline or do a basic conversion from a 50p/60p file to 50i/60i file this overlapping/interlacing does not normally occur so where there is fast motion you will sometimes get temporal aliasing which will appear as combing or tearing of the image. So it will be 50i/60i but in some cases it may have artefacts that would not be seen from a true interlace camera.
I have not yet tested whether this same artefact is present in the FX6’s interlace output as I suspect this will be a simple 50p/60p to 50i/60i conversion so the same artefacts are possible on the cameras output."
