Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Need Help - Tons of motion blur on very minimal movement Red MX footage

No filtration on most shots, except for a few shots with a pola. F stop would swing between 2.8 and 4 for most shots.

The subject had a fair amount of movement, but even when that was restrained, the blur issue seems to have continued even with his minimal movement.

In terms of lighting, nothing was out of the ordinary from my experience of working with this camera and lighting package, and I've worked with quiet a few Reds and with this particular Red many times. Lighting on set was done primarily with 1 ks, and 650s and 4x4 kinos. There were no big lights on set, so we were not working with loads of light. I would characterize it as a relatively dim set. Plenty enough for proper exposure, but we were not needing to add ND's or close up the aperture much. And we never dropped below a 180 degree (1/48th). So yes I stand by the camera report.

As for the taking out a support ticket, that was going to be my suggestion tomorrow after myself and the DP got a good look at the footage.

I appreciate all your support everyone. I've spend way to many nights staying up and reading these boards over the years so I know how knowledgeable you all are.
 
Shawn,

After reading your posts here, you seem like a competent, and well versed DP. If what you are saying is true, and I have no reason to doubt you at this point, then something might be up with your camera.

If I was to make a guess (and a WAG at that), it sounds like the camera is giving you 360 degree shutter all the time. If that is the case, and you have access to that same camera today, take it outside, set it to 24p, and a 90 or 45 degree shutter and point it at something moving rapidly at 90 degrees to the focal plane, like street traffic. At a 90 or 45 degree shutter, motion blur should be minimal. If it's pronounced, then it sounds like the shutter is not responding the way it should.

I also agree that you should open a ticket and I am sure those folks can help you directly.

Having a camera not do what it should DESPITE doing everything properly, is a very, very frustrating thing at any level.
 
yes Shawn is definitely competent ( 1st AC ) and thats why I don't believe the camera reports are wrong. We were watching for this very problem after noticing it early so I believe when he says he was checking the shutter speed multiple times throughout the day to make sure it was correct.

And I just want to say that I love that my 1st AC is on here trying to help rather than just walking away!!!!! You're the best Shawn!

Shawn, myself, and the DP are meeting in a few hours to look at the footage together. Will report our findings back to you guys soon.
 
Histograms

Histograms

Since the camera is semi-RAW recording, if there was some kind of difference between the shutter angle readout, and the shutter angle the sensor was shooting at, then the histograms should show the error.

If the histograms show the subject peak near where a 18% gray card peak comes out at the same values of T/ stop, ISO, and shutter angle thought to be used, then how could the shutter angle be off, if it was off then the histogram peak would be moved to the white clip by a stop?

That's one good thing about RAW cameras, thre is no gain adjustment change on the recorded data, only in the de-Bayer, so you can see the sensor exposure level in the RAW histograms. You do need to look out for the de-Bayer program compensating for any camera "bug", so maybe you need to override the ISO etc. in the de-Bayer to see the histograms all at the same software "gain" level to compare the histograms.
 
Speaking not so off topic, when I was AC-ing a shoot last summer on the MX, the shutter speed was at 1/48th of a second, but the motion blur looked like it was 1/60th it had a heavy videoish feel, DP confirmed it looking like that.

Though its the opposite problem you're having, I thought I'd mention it.
 
Hey guys, Shawn and I met today and he created a short R3d file of one of the problem clips. I have found now that I converted to apple pro res and can play the clips smoothly that the wide shots are fine. If i arrow key through the frames it looks like bad motion blur all the way through, but to the naked eye when playing the video it looks normal. So this is good news for the wide shots.

As for the mediums and close ups, the problem is still there.

Here is a megaupload download link to download the r3d file example:

http://www.megaupload.com/?d=LV9JZZ26

Its only about 50 MB. hope this helps. We tried some tests with the camera today (i'm sure Shawn will chime in soon with the exact details) but from what I gathered, even after a full factory reset of the settings, we got the same motion blur problem when recording close up of a face going side to side.
 
Is it just me ?, there seems nothing wrong technically with the clip, its definitely a 180 or 200 degree shutter. By checking from one frame to the next you can see the blur distance roughly equals the gap between the start of the same blur in the next frame. The Blur phenomena is simply from the fact that he's moving a fair distance in each 1/48th of a sec. It's obviously worse when you are in close up. If you use a shorter shutter speed the individual frames will be sharper, however then the movement will have a strobe effect. Try a test with say a 1/192nd shutter to see the effect. It's a similar effect to rapid panning.
These effects are obviously more than twice as bad on progressive 24 fps when compared to 29.97 interlaced footage.
 
Marcus Struzina said:
Is it just me ?, there seems nothing wrong technically with the clip, its definitely a 180 or 200 degree shutter. By checking from one frame to the next you can see the blur distance roughly equals the gap between the start of the same blur in the next frame.
Click to expand...

Yep, i can confirm that - looks like perfectly normal 180 degree footage to me. Nothing wrong with it.

The attached file has two consecutive frames overlaid - the blur length roughly equals the gap, as said.
 

Attachments

  • Shutterspeed.jpg
    Shutterspeed.jpg
    54.4 KB · Views: 0
Hey guys.

I spend over an hour staring at the footage today. At first I felt there was nothing wrong, but the more I stared at it the more it just felt wrong. I couldn't put my finger on it, but it just seemed to be off slightly. I know that's not very descriptive, but I just cant figure it out. It just didn't feel smooth when watching, almost like it tired the eyes to watch. (And yes playback was real time, I was previewing in the newest version of RedCine-x at both 1/8 and 1/4 resolution which gives me realtime playback).

Obviously this issue is much more noticeable on the close ups as the movement is quiet a bit more substantial within the frame.

I would guess that we are all just seeing things and that the footage is fine, but after watching the footage for awhile, I couldn't get ride of the nagging feeling something was up.
 
Shawn Seifert said:
I would guess that we are all just seeing things and that the footage is fine, but after watching the footage for awhile, I couldn't get ride of the nagging feeling something was up.
Click to expand...

I know that feeling ;-)

It usually helps to do something else for a while, then come back with fresh eyes...
 
Marcus Struzina said:
Is it just me ?, there seems nothing wrong technically with the clip, its definitely a 180 or 200 degree shutter. By checking from one frame to the next you can see the blur distance roughly equals the gap between the start of the same blur in the next frame. The Blur phenomena is simply from the fact that he's moving a fair distance in each 1/48th of a sec. It's obviously worse when you are in close up. If you use a shorter shutter speed the individual frames will be sharper, however then the movement will have a strobe effect. Try a test with say a 1/192nd shutter to see the effect. It's a similar effect to rapid panning.
These effects are obviously more than twice as bad on progressive 24 fps when compared to 29.97 interlaced footage.
Click to expand...

you're right.. its a 100 mm lens shot,, and the guy is moving so quickly..
 
I pulled this into the editor to have a better look, and laid a BITC on it.

I see a few things going on...1

1. Look at frame 33. When he is apparently quite still, the frame looks good.

2. The DOF is quite shallow. Also looking at frame 33, we can see his left eye (camera right) is in reasonable focus, but his other eye is out as is the zipper of his jacket on that side. So we are not only battling issues of motion blur due to shutter speed, we are also battling shallow DOF pushing things out of focus.

3. If we look at frame #3 (anything in the first 5-6 frames really), we can see that the motion across the frame is well beyond what could be captured at a 1/48 shutter. He moves edge to edge in less than 12 frames. I'd seriously be considering a 1/125 shutter or thereabouts if I was capturing this and wanted it crisp.

Given what I see from this R3D, I don't think the camera is at fault. I simply think the relatively high degree of motion was overwhelming the shutter speed for the camera. As the motion ceases, the subject sharpens.

I concur with Marcus. Nothing wrong technically, simply a factor of photographic choice. In the wides, the apparent motion is less (the motion takes up much less of the screen), so the effect is lessened.
 
Alison,

I think you are going to have to replicate some of the shoot and see if you get the same result. Essentially getting someone to move like your talent did in front of the camera setup like you believe it was. Look at the motion blur and see if it's similar to what you got from this shoot.

You should also consider shooting with another camera, e.g. HVX, HV-20 and set that one also at 24/180. If you frame the shots the same, the motion blur should be the same. This will give you a benchmark. (Obviously the smaller size chip will have greater DoF, so its overall image will look sharper, but the blur should be the same.)

It sounds like you're using RedCine and not RedCine-X for transcoding. I think you should use the latest, which is RedCine-X (actually your AC mentions using RedCine-X) and be sure a full debayer is being done.

You also might want to check the backfocus of the lens. It's also possible that the lens is slightly out of focus and that the movement is just making an already blurry image look even worse.

Good luck.


P.S. Also, realize that using the camera settings you specified and assuming the camera was ten feet away from the talent, that the depth of field was about 7 inches. Depending on how the talent was moving, that might not have been enough. Again, good luck.

P.P.S. If you wind up having to reshoot, sobeit. In "Lord of the Rings" they had to reshoot some of the most emotional scenes b/c of soft focus. As for software solutions to fix it in post, IDK.
 
24p vs. 60i

24p vs. 60i

In clasic filmmaking actors did not move qucikly in close-up shots.

The recommended pan rate tables were made for smaller movie screens and for resolutions much lower than what todays Digital Cinema cameras can produce, so any motion makes the resolution fall more from its higher peak.

Use 60p or 60i for sports and shooting on a pogo-stick, 24p belongs on a tripod and with follow camera. In that shot, the camera does not follow the action, if it did the subject would be less blured and the background more blured, as a moving actor would have been shot in clasic times, a wider shot and follow camera to keep the subject blur under control, and shown at lower resolution on a smaller screen so the angle of the blur to the eye would have also been smaller.

If you have seen silent movies shot at 8fps to 16fps, the motion blur even in medium shots can be quite heavy. We have a 2709 tripod that us geared down so much, that it might take 15 seconds to pan 90 degrees, and that's if you crank fast.

People with a background in Video, need to adjust to the fact that 24p is not a "video" frequency and take camera movement speed and subject movement speed and reduce them by about a factor or 2.5 to 5 since the images are shot with a longer exposure and projected on a larger screen at higher resolution making the motion angle greater to the viewer.
 
Hey guys, thanks again for the insights..

I have just heard back from Red in regards to the ticket we filed:

"Hi,

Thank you for contacting RED. We have taken a look at your R3D file and everything appears to look normal for that shutter speed. The shutter you set was confirmed at 1/48, and if you would like a sharper image for a shot like that you might want to increase your shutter.

Please let me know if we can assist further."

We are reshooting the close ups today with a 5D but maybe we'll try the Red also on a higher shutter..
 
I concur with the consensus here.

The footage is entirely normal.

Given the directorial intent (sharp images) I would have set shutter 90° or narrower to capture movement more sharply, or asked the subject to move less. (Probably not entirely appropriate for a music video.)

I took some time and critically watched some HD music videos today. Shooting with narrow shutter angles seems fairly common. Sometimes if you are pixel peeping you can sometimes see that the motion blur that is present was added in post. (Too uniform, and too linear in cases where rotational movement was evident.)

That indicates to me that the preference for that genre is to shoot sharp images with narrow shutter angles, preferring to err on the side of too sharp in camera and adding blur in post. You can fix too sharp easily, but fixing too blurry is almost impossible.

As far as close ups go, most were either insert shots or cuts to a different setting. They tended towards exaggerated stillness, not "full speed/tempo" dancing. Shots tend to be significantly wider than normal cinema production. Interestingly, perhaps because of the contrast of subject motion together with decreased frequency of use and setting changes, close ups had a more significant effect.

Of course, I rarely watch music videos critically (if at all!)- so my sample might be skewed.
 
You must log in or register to reply here.
Back
Top