XC10 can output 4K>1080 to Ninja Star (95% convinced it's 10 bit)

[Lintelfilm]

It's automatic - the XC10 detects 1080 recorder. You have to use the 60i pulldown feature on the Ninja Star but there are no artefacts.

I'm 95% convinced it's 10bit - colours on the Atomos files are much more varied and separated (even if I adjust exposure to compensate for differences on the recorder).

I'll post some side by side images later if I have time ...

[jgharding]

Please do, I'd be interested to see

[hmcindie]

8-bit is ludicrously panned even though good 8-bits is really good.

Proof? A lot of Canon RAW videos were actually processed into 8-bits. When people use AE to import Canon RAW, it will default to 8-bits. The option to change it has always been there in the importer (it's not enough to set AE into 16 bits) but very well hidden so LOTS of people (I'd say 100%) who used AE to process Canon RAW were actually importing them to 8-bit files and after that they did the grading. Myself included. Just stumbled into the importer options late last year and I was moderately shocked to find the 16-bit switch inside Canon importer that was defaulted to 8-bits.

And even then it looked great. That's because no compression.

[jgharding]

14 minutes ago, hmcindie said:

8-bit is ludicrously panned even though good 8-bits is really good.

And even then it looked great. That's because no compression.

This matches my experience recording C100 with Ninja 2. I was using DNxHD 220x which is 10-bit, and was thinking "is this somehow filling in the extra bits?!" But it isn't, the output is 8-bit, it's just having so much more data really helps.

[Lintelfilm]

Sorry guys don't think I'll have time to share the images yet I'm juggling 2 jobs and a big pitch at the moment. 

For me 10 bit is all about colour rather than banding, artefacts, grading in post, adding FX, etc. Yes high bitrate 8bit can look great but 10 bit colours invariably have more visible depth and hence subjects look more "there"/real. I do seem to be more sensitive to this than most filmmakers I talk to - perhaps because I come from a painting background I think - so I guess you can argue its subjective. I remember Andrew Reid's NX1 JPEG vs RAW spot the difference post - I was able to see which was which immediately, clear as day.

Greater colour depth = better, richer, more subtly varied and separated colour. No question. This only becomes amplified at the grading stage.

However I don't buy the "down sampling to 8 bit in post anyway so no point acquiring in 10 bit" line - look at HD cameras that downsample to HD from 4K (or 4K downsampled in post). The more information you have later in the processing stage the better.

8 hours ago, jgharding said:

This matches my experience recording C100 with Ninja 2. I was using DNxHD 220x which is 10-bit, and was thinking "is this somehow filling in the extra bits?!" But it isn't, the output is 8-bit, it's just having so much more data really helps.

There's a Cinema5D post and video on this from ages ago - it's clear that recording to 10bit 422 ProRes HQ results in a much better image than recording in standard ProRes 422 (8bit), so in one way or another you're right.

https://***URL not allowed***/canon-c100-internal-avchd-vs-prores-compared/

 

[jgharding]

18 hours ago, Lintelfilm said:

There's a Cinema5D post and video on this from ages ago - it's clear that recording to 10bit 422 ProRes HQ results in a much better image than recording in standard ProRes 422 (8bit), so in one way or another you're right.

https://***URL not allowed***/canon-c100-internal-avchd-vs-prores-compared/

 

haha yeah TBH I'm not a scientist, so it's all second hand info, i just know it looks better!

[mercer]

On May 9, 2016 at 6:54 AM, Lintelfilm said:

It's automatic - the XC10 detects 1080 recorder. You have to use the 60i pulldown feature on the Ninja Star but there are no artefacts.

I'm 95% convinced it's 10bit - colours on the Atomos files are much more varied and separated (even if I adjust exposure to compensate for differences on the recorder).

I'll post some side by side images later if I have time ...

Why do you have to use the 60i function?

[Lintelfilm]

5 hours ago, mercer said:

Why do you have to use the 60i function?

Because although the camera is doing 25p it sends it over a 50/60i hdmi signal. I have no idea why but it's quite common. Perhaps to cater for displaying footage on a TV. The Ninja deals with it though I believe.

[mercer]

3 hours ago, Lintelfilm said:

Because although the camera is doing 25p it sends it over a 50/60i hdmi signal. I have no idea why but it's quite common. Perhaps to cater for displaying footage on a TV. The Ninja deals with it though I believe.

Oh, is there any change in motion cadence or blur or anything? Do you have any videos on Vimeo where you used it with the XC10?

[mercer]

Also, is the 60i due to the 4K downscale or does the HDMI output also force 60i with 1080p footage?

I actually really like the 1080p footage in the XC10. I think it's as good as most consumer 4K downscaled in post, but if I could get the 220mbps prores, in 1080p, with the Ninja Star, that would be very interesting. This thing would be a run & gun BM cam with 5-Axis stabilization and Canon color. 

[sanveer]

On 09/05/2016 at 4:24 PM, Lintelfilm said:

It's automatic - the XC10 detects 1080 recorder. You have to use the 60i pulldown feature on the Ninja Star but there are no artefacts.

I'm 95% convinced it's 10bit - colours on the Atomos files are much more varied and separated (even if I adjust exposure to compensate for differences on the recorder).

I'll post some side by side images later if I have time ...

I could be wrong, but good, High Bitrate 8-bit, 4-2-2 does not look very different from 10-bit 4-2-2. Also, it does tolerate grading well, as can be noticed from the various Canon codecs.