Desire: A run-and-gun style unit capable of producing high-quality HD footage.

I've just done a crash course on fully understanding Bayer-type sensors. For my needs, even though these units can capture "4K", I only want HD. In the future, when either a proper 3-chip 4K sensor (or 4K oversampling on a single 8K sensor) emerges, then I'll move to 4K capture and output.

As many know, there are huge differences in feature set on these two units. Here's a list of pros/cons I have found that would be important to me. But would be helpful to hear other's opinions.

Above all, I wish to have the highest possible quality footage. An in most cases, will have to use available light only.

Main X70 advantages over the XC10:
* Three choices of ND that are effectively independent of iris setting. I believe the XC10's single ND filter would only begin to kick in at f/4 or narrower.
* Waveform monitors
* XLR inputs with dedicated audio level dials.
* So many more buttons; less digging through menus

Main XC10 advantages:
* 12 stop dynamic range. My EOS 6D for stills enjoys this range. I would plan on capturing in Canon Log. This should be a bit more gradable-friendly than the X70 (estimated by Alister Chapman as having 10 to 11 stops of DR).
* A decent amount lighter than the X70; and really see the benefit of the rotating grip.
* I can use my existing 58mm lens filters (e.g. CP, Clear) and existing LP-E6 batteries.
* Assuming no XAVC-L fiasco in terms of bringing in footage directly into FCP X.


Other X70 advantages. Spec-wise, these should give higher quality. Practically-speaking, not sure how much so:
* 10-bit 4:2:2 vs. 8-bit 4:2:2
* 50 Mbps (for 23.976p and 29.97p) vs. Canon's 35 Mbps for the same frame rates. Canon only goes to 50 Mbps when doing 59.94p
* 24-bit audio; yes, more latitude, but after much reading, not 100% sold on it.


Now then, saving a couple hundred on having filters/batteries isn't a big deal at all. So this effectively leaves the Canon's higher DR and weight as its main advantages.

Speaking only to image quality, which unit is better? Go for higher DR? Or, will the higher bit-depth, bit-rate and less compression ratio of the X70 win out even if 1 to 2 stops less DR?

Note: I foresee grading all my footage. If that makes any difference.


Side note on audio. I would minimally be using a Røde Stereo VideoMic Pro. Hooked up via 3.5mm on either cam. At maximum, I would use a Stereo VideoMic X. That would be hooked up with XLR on the X70 or 3.5mm on the XC10.

This is an interesting question, I've been thinking of this myself... despite owning an X70.

Main X70 advantages over the XC10:
* Three choices of ND that are effectively independent of iris setting. I believe the XC10's single ND filter would only begin to kick in at f/4 or narrower.

That's pretty huge for me. If the XC10 had the same NDs as the X70, it would be MUCH more appealing.


* Waveform monitors

Sorry to disappoint you -- REALLY sorry -- but the X70 doesn't have waveform; just a basic histogram.


* XLR inputs with dedicated audio level dials.
* So many more buttons; less digging through menus

Those are both very nice. The X70 could actually use more buttons; yes, there are 6 assignable, but so many cool things to assign them to...


* A decent amount lighter than the X70; and really see the benefit of the rotating grip.

Nice for me... portability is a big deal for me. I don't really have a handle on how big it is yet. Once nice thing about the X70 is that the handle comes off and makes it a very pack-able size and shape.


* Assuming no XAVC-L fiasco in terms of bringing in footage directly into FCP X.

This I don't know ... from the video I say it looks like they've come up with YET ANOTHER proprietary wrapper around H.264...? Why on earth not just give us MP4? (I guess licensing fees?) Or why not use XAVC?


* 10-bit 4:2:2 vs. 8-bit 4:2:2

I am REALLY REALLY loving the 10-bit off the X70 -- it seems really organic, and very gradeable. (I'm used to 8-bit AVCHD on the FS700.)


* 50 Mbps (for 23.976p and 29.97p) vs. Canon's 35 Mbps for the same frame rates.

Although the XC10 has a good high bit rate in 4k mode -- so shooting 4k and downsizing should give awesome results. Still 8-bit, though.


Speaking only to image quality, which unit is better? Go for higher DR? Or, will the higher bit-depth, bit-rate and less compression ratio of the X70 win out even if 1 to 2 stops less DR?

In my experience the DR on the X70 is fine. I couldn't put a number on it, but shooting very contrasty outdoors stuff has worked pretty well for me. Eg:

https://vimeo.com/channels/832674

All in all... for me, the XC10 is interesting; and having 4k (REAL 4k, not the cheesy stuff the X70 is getting) would be cool. But I don't actually NEED 4k, so I think I'll be sticking with the X70.

HTH... good luck!

All in all... for me, the XC10 is interesting; and having 4k (REAL 4k, not the cheesy stuff the X70 is getting) would be cool. But I don't actually NEED 4k, so I think I'll be sticking with the X70.

HTH... good luck!

Thank you for the detailed reply, Atticus!

When you say "real 4K", what specifically do you mean? The XC10 doesn't actually have a real 4K sensor. It only has a 4K bayer pattern one. So there would be interpolation going on to fill in missing color values. In HD mode though, each "pixel" is made up of four photo sites; one R, one B and two G. Math is still applied, but now you get all three color components per HD pixel. At least that's how I've read things.

A very interesting experiment would be to record 205 or 305 Mbps 4K footage then downconvert to HD in post. Versus recording HD in camera at 50 or 35 Mbps (bit rates are dependent upon frame rate settings).

Differences in compression artifacts aside, my guess is that the HD footage may have better color info than the post-downscaled 4K.

When you say "real 4K", what specifically do you mean?

I was snarkily referring to the fact that the X70 is apparently getting 60Mb/s 4k, which is what the (cheaper) AX100 had when it came out. Given that this is just 20% more bits for 300% more pixels, and that the AX100 now has 100Mb/s, and that we're expected to pay for it, that seems pretty poor. Still, as an HD camera, the X70 is superb.

Differences in compression artifacts aside, my guess is that the HD footage may have better color info than the post-downscaled 4K.

Hmmm, interesting. I would have thought the opposite, since there's more information to begin with. Plus, how is the HD produced? Really, it's just doing the same kind of downscaling, but in the camera... isn't it? If so, then the difference is the much higher bit rate you have to work with.

Hmmm, interesting. I would have thought the opposite, since there's more information to begin with. Plus, how is the HD produced? Really, it's just doing the same kind of downscaling, but in the camera... isn't it? If so, then the difference is the much higher bit rate you have to work with.

According to what I've read about the Canon XC10 (and its bigger siblings such as C100), the 4K bayer pattern sensor will use 4 photo sites per HD pixel. So each pixel gets one red, one blue, and two green photo sites. Chroma subsampling then applied to arrive at 4:2:2.

But if recording at 4K, each photo site can only capture a single color component and the cam would need to re-construct the missing two components. Then apply subsampling also at 4:2:2

So my thought is that when capturing HD, there is far less re-construction (interpolation) going on.

I believe the X70's sensor is similar in this regard; quantity of 4 photo sites (or more?) per HD pixel.


I found some literature on Sony's F65 (http://pro.sony.com/bbsccms/assets/files/show/highend/includes/F65_Camera_CameraPDF.pdf) which uses a 20 MP sensor. But they arranged things in a different pattern which is better than a 4K bayer pattern sensor. I don't know if the X70's 20 MP sensor shares any similarites in layout to the F65.

...But if recording at 4K, each photo site can only capture a single color component and the cam would need to re-construct the missing two components. Then apply subsampling also at 4:2:2

I see what you're saying. When you downscale in post, you will be averaging 4 pixels to make 1, and so will recover that colour information; but half is already lost due to the subsampling. So yes, that could make a difference, though I'm not sure how much in practice. However the hugely better bit rate in 4k should help.

I believe the X70's sensor is similar in this regard; quantity of 4 photo sites (or more?) per HD pixel.

I think so, though I've seen no details on how that works.

I found some literature on Sony's F65 (http://pro.sony.com/bbsccms/assets/files/show/highend/includes/F65_Camera_CameraPDF.pdf) which uses a 20 MP sensor. But they arranged things in a different pattern which is better than a 4K bayer pattern sensor. I don't know if the X70's 20 MP sensor shares any similarites in layout to the F65.

I doubt it. I thinik it's basically a still-camera sensor re-used for video; like the same sensor as in their 1" compacts. Good news, at least they're reading the entire sensor rather than decimating it (i.e. line skipping).

Adding a bit of closure to this thread. I got my X70 last week and couldn't be happier.

Really liking the footage. Very crisp images; no aliasing. And loving the degree to which I can grade.

Now it's time to practice, practice, practice.

Awesome, glad you're liking it!