Zenmuse X3 smoother than X5 or Hasselblad?

Petah · Jun 3, 2019

I picked up another Inspire 1 V2, this one with a Zenmuse X3. I went out and flew it (of course), and just left the camera on recording video. I wanted to compare / contrast this with my other 2 drones. The result surprised me.

In still images, the Hasselblad wins, least distortion, sharpest images and best color rendition (it shoots in RAW so no real surprise there. 2nd place was predictablly the X5, with less distortion than the X3.
Here's the surprise, for which I have no explanation. All 3 aircraft shoot beautiful video, but the X3 handles lateral panning much more smoothly than either the X5 or the Hasselblad, both of which are jerky when panning side to side. Overall whether yawing the aircraft or just the gimbal, the video from the X3 seems smoother. Could it be because I'm encoding the video as a .mov file on the X3, whereas the X5 and Hasselblad are encoding as MP4? Scratching my head on this one, interested in ideas...

adwb · Jun 4, 2019

The video container makes no difference, I suspect it’s due to the frame rate, shutter speed combination effect.
You need all three to be identical settings to test effectively

Donnie Frank · Jun 4, 2019

Assuming you're talking about mechanically smoother, and not digitally smoother (you should clarify), the wider FOV is probably contributing to a smoother look.

D

Petah · Jun 4, 2019

Donnie Frank said:
Assuming you're talking about mechanically smoother, and not digitally smoother (you should clarify), the wider FOV is probably contributing to a smoother look.

D

Actually, I mean digitally smoother, at least I think I do. I'd read that CMOS cameras struggle when moving laterally to keep up with the changing image, due to the top-to-bottom way the sensor scans the image. Mechanically all 3 gimbals move smoothly in all axes; the resulting video though was choppy when panning left-to-right and vice-versa. Think of a lateral dolly shot, moving the aircraft sideways at 90 degrees to the subject; that's where I'm seeing issues. Much less so on the X3 than the other two. Could be the frame rate or wider FOV; now I'm curious to take the same shot with all 3 cameras and see what I get.

Donnie Frank · Jun 5, 2019

Petah said:
Actually, I mean digitally smoother, at least I think I do. I'd read that CMOS cameras struggle when moving laterally to keep up with the changing image, due to the top-to-bottom way the sensor scans the image. Mechanically all 3 gimbals move smoothly in all axes; the resulting video though was choppy when panning left-to-right and vice-versa. Think of a lateral dolly shot, moving the aircraft sideways at 90 degrees to the subject; that's where I'm seeing issues. Much less so on the X3 than the other two. Could be the frame rate or wider FOV; now I'm curious to take the same shot with all 3 cameras and see what I get.

Thanx for the clarification. I have good news, bad news and good news.

Good news: There's nothing wrong with your camera.
Bad news: As long as you shoot 24 or 30 fps, there's nothing you can do to fix this. I've seen Red and ARRI cameras in well-budgeted productions do the same thing.
Good news: You CAN minimize this problem by simply panning slower or shooting 48 or 60 FPS.

What you're describing is called "judder." Sometimes judder occurs and sometimes it doesn't. And for the love of all that is holy, I can't find what causes it. But obviously I'm not alone...when I see Red and ARRI cameras do the same thing in feature films, I feel a little better about about myself as an amateur video pathologist. And rest assured I ALWAYS follow the 180° shutter rule and ALL rules of cinematography when I shoot. Honestly, I believe its an artifact of the viewing medium and any (codecs that may be involved) and NOT the footage itself, but I've yet to prove that.

If you can stand to listen to this guy talk for more than a minute, he explains how the viewing medium is often the culprit:

Note: You'll only see judder in bright sunlight, which leads one to believe it is an exposure issue. But I have actually UNDER-exposed footage, and STILL gotten judder in sunlight. The same exact shot at night, properly exposed, nets zero judder. Maddening.

Another unavoidable effect you'll encounter some day is the moiré effect, which can be exasperated via aliasing due to a pixel mismatch between the footage and the viewing medium.

I was on a shoot for a television series and the DP complained about moiré on a bridge I had just shot. I was taken aback. I'm like, "Do you not know about this effect, what causes it, and that the only way to avoid it is to either blur the photo or show the footage on a 1:1 pixel ratio???" We were watching 4K footage of a bridge on a 1080p screen, so I fully expected it. Imagine my shock and horror when this DP complained about it? I retrospect, I should've just shot those scenes in 1080p and all would've been fine. But I digress...

Some people claim that slowing your shutter speed will fix this, but they are liars. In this video:

...this guy CLAIMS to fix this problem. But I STILL see judder, even when he's claimed to have fixed it @ 1/50 shutter. Is it the YouTube compression codec? Or the refresh rate of my monitor? Who knows?

If you find a cure, please share it here. But honestly, because we can't control what monitor people use to view video, we may never cure it completely. But the fact that one never sees judder in low-light footage, REGARDLESS of viewing medium, pulls me AWAY from blaming the viewing medium. Like I said, maddening.

D

S White · Jun 9, 2019

Great info. Thanks for all the detail. I fly an I2 w/ x5s, and have a backup I1 w/ x3 and z3. As far as I am concerned the I1 is a much better unit w/ the x3 than z3. Granted the z3 has tele, but I find a video delay often. And the z3 has a bit of blur on the left side of the image. The x3 is a great unit and I record at 30 or 60, at 1080. The 4k stuff is good. I love my I2, but the I1, for being the newest thing at the time is still a great unit. The control features of the I2 w/ the Cendence and Crystalsky is a huge step up, but still the ol' x3 works great when called upon.

Donnie Frank · Jun 9, 2019

S White said:
Great info. Thanks for all the detail. I fly an I2 w/ x5s, and have a backup I1 w/ x3 and z3. As far as I am concerned the I1 is a much better unit w/ the x3 than z3. Granted the z3 has tele, but I find a video delay often. And the z3 has a bit of blur on the left side of the image. The x3 is a great unit and I record at 30 or 60, at 1080. The 4k stuff is good. I love my I2, but the I1, for being the newest thing at the time is still a great unit. The control features of the I2 w/ the Cendence and Crystalsky is a huge step up, but still the ol' x3 works great when called upon.

I don't own an I2 but everything you said about the I1 is spot on. The X3 is a fine camera for "quick-n-dirty" shooting. And up to as late as last year, the X3 was widely accepted in low budget films and commercials that wanted aerial without the big budget of a RAW set up.

And I've always maintained that the Z3 would be a fine inspection camera. But after shooting a few test shots with it, I would never bring it on set. Ever. If I'm shooting in low light, I break out the X5.

If I do decide to invest in a RAW rig, it'll be an M600 Pro with an X5S (which is still a fine camera IMHO).

D

Petah · Jun 10, 2019

Many thanks, Donnie, for the detailed information. Dude, that’s way beyond the call of duty, so I’m grateful. I’m on vacation just now (only one drone along for company) but I plan on doing some testing when I get home to see if I can use the multiple camera setups I have to experiment some.

My plan- take the M2P with the Hasselblad, one I1P with a Z3 and one I1P with a Z5, and shoot video of a fixed object while moving laterally across it at (as close to) the same speed for each aircraft. Tests I plan on trying:
1) All 3 cameras at 60FPS, let the camera pick the aperture and ISO and Resolution
2) All 3 cameras at 30 FPS, with aperture set the same on all 3 cameras (sort of aperture priority in SLR terms)
3) Locked Down test, with all 3 cameras set for 1080P/30FPS/Common Aperture

I’ll compare the results for judder (now that i know what it is) and post here when I’m done.

Unless of course someone’s already tried this...

Donnie Frank · Jun 10, 2019

Petah said:
Many thanks, Donnie, for the detailed information. Dude, that’s way beyond the call of duty, so I’m grateful.

My pleasure.

Petah said:
My plan- take the M2P with the Hasselblad, one I1P with a Z3 and one I1P with a Z5, and shoot video of a fixed object while moving laterally across it at (as close to) the same speed for each aircraft. Tests I plan on trying:
1) All 3 cameras at 60FPS, let the camera pick the aperture and ISO and Resolution
2) All 3 cameras at 30 FPS, with aperture set the same on all 3 cameras (sort of aperture priority in SLR terms)
3) Locked Down test, with all 3 cameras set for 1080P/30FPS/Common Aperture

I’ll compare the results for judder (now that i know what it is) and post here when I’m done.

Unless of course someone’s already tried this...

I believe the most telling test will be to set SHUTTER priority conforming to the 180° shutter rule and do a couple pans at varying speeds at high noon. Then do the same shot again at twilight and then again at night. I believe here is where you will find your most profound differences. You will find that the high noon shots will more than likely show judder at faster pan speeds, with the twilight shots showing less judder and the night shots showing almost none. Those were my findings with the X3. You have me curious now, so I may try this same experiment with my P4P.

As a side note, a CLOUDY day will also allow minimize judder.

One variable I haven't considered is color temperature. I never use auto. Generally speaking, I shoot high noon around 5800K-6000K, cloudy days about 7000K and night around 4200K (season to taste for all three). Is it possible that WB settings have something to do with this? I would think not....but at this point I'm grasping for straws.

D