Vegas Pro 11 released - Page 19

I'm quite frustrated with the results of the ISO Blu-ray 3D 1080p24 generated by Sony Vegas (MVC 1920x1080-24p 25Mbps video stream). See examples below with MVC 1080p24, MVC 720p60 and 3D SBS 1080p60:

MVC 1080p24:

MVC 720p60:

3D SBS 1080p60:


This video was originally recorded with two camcorders Sony HDR-CX130 attached in a moving car in 1080p60 and edited in Sony Vegas Pro 11.

I wonder of you are having the same problem of fluidity in the generation of ISO 1080p24 Blu-ray 3D with Sony Vegas. Unfortunately I have no other software that render Blu-ray 3D 1080p24. What most intrigues me is because I've seen many movies in 3D 1080p24 and not seen this type of problem. I'm using PowerDVD 12 for 3D videos.

Thanks!
Milton

Yes, I can never get acceptable rendering of significant pans at 24p. Besides the blurring there is a ridiculous amount of judder and sometimes other distortions that look like excessive ghosting. I'm rendering at 720 60p if the video has much motion. I sure wish displays were capable of showing 1080p 3D at 30 fps - or even better, 60 fps.

I think this problem is primarily a shortcoming in the HDMI specs and is not specific to Sony Vegas. It astounds me how shortsided HDMI specs have been in every version since their beginning.

Ok, I came to the conclusion that it is better to generate Blu-ray 3D in 720p 60Hz, since the two camcorders that I use do not record in 1080p 24Hz, only record 1080p 60Hz or 1080i 60Hz. I will create 3D SBS (side by side) content in 1080p 60Hz.

When Sony Vegas can generate 3D Blu-ray in 1080i 60Hz (AVCHD 2.0/MVC), then I think that would be a better option. But from what I saw, the Vegas 12 still is not generating 3D Blu-ray in 1080i 60Hz, only 1080p 24Hz and 720p 60Hz at this moment. We also need a Blu-ray player that reads this format AVCHD 2.0 (MVC).

Thank you!
Milton

Good 3DTVs have no problem displaying 24/48Hz MVC files from 3D blu-ray discs - no judder. The problem is going the other way. It is difficult to convert to 24Hz due to non-integer relationship.If you need 24 fps video, the best way is to capture natively that way. Wolfgang explained it. It is difficult to do a 3:2 pulldown without judder. Little smear actually helps with the motion, even though it doesn't look great frame by frame.

Yes, and like you say, "If you need 24 fps video..."

Why do we need 24 fps video?

For movie theater systems it is the standard frame rate, set up long time ago. All movies and projectors (pretty much) are that way. It was a limitation of mechanical movie cameras and projectors moving celluloid - and it stuck. If you ever make a video for a movie theater projector use, they'll demand 24 FPS (http://3dff.org/?page_id=52). In 3DBDs it is partially because of that. It also helps to keep the bitrate down for HDMI. It is very demanding for equipment and cables to send serial bits for all the pixels (uncompressed) for two 1080p streams at higher frame rates. Everything is more expensive, and hard to make reliable, especially long cables. The high prices make it difficult to sell to consumers for home use. To change, there are legacy content and equipment cost issues. Part of the reason why ATSC ended up with 36 formats.

Also see: http://www.projectorpeople.com/resources/pulldown.asp and http://en.wikipedia.org/wiki/Telecine.

Of course when going from 30 FPS to 24 (reverse-Telecine), one in 5 frames has to be dropped, which causes judder, where the smearing helps a little overall.
Edited by ComPH3D - 10/5/12 at 3:50pm

Traditional cinematography used a 50% exposure time. This blurred fast motion, and pans. Movie directors worked within the limitations of 24fps. What Vegas has produced for you is not unlike traditional 24fps blur!

Road traffic, or scenes shot from a moving car, will look jerky if captured crisply (short exposure time) at 24fps from a distance of a few metres, and displayed on a computer monitor, or TV, without motion smoothing.

If you examine individual frames in your 60p footage you will see the huge distance a car moves over two frames. 60p is a more natural frame rate to use for this type of material.

Will the display do motion interpolation for Full HD 3D?

I have a problem with some 1080p24 footage I took recently using two webcams (for 3D purposes). Each frame is quite clear, but if I author as a Full HD Bluray at 24fps, the pans look very jittery; and the road traffic too.

In 2D mode my TV set is able to do motion interpolation for 1080p24 (Panasonic Intelligent Frame Creation) and this cleans up the pans very nicely, and lessens the jerkiness of the moving traffic. However there is no motion interpolation available on my set for frame packed (MVC) 3D.

I think a good compromise for fast moving 60p 3D source material is to author as frame packed 720p60.
Edited by MLXXX - 10/6/12 at 3:23am

Hey Jeff:

Thanks for your assessment and opinions above. I've been trying to decide how to shoot a fireworks and outdoor moderate speed motion performance tonight. I want to shoot in 1080 60p but was not sure how to finish it for blu ray. I forgot about the 720 60p option. I'll be shooting with 18" IA and the TD10's as I also need to have zoom. The performance will be at low light too so slower shutter speed and wide open iris too.

I'll try to check in with my smart phone later today to see if you have any additional thoughts on this.

Sorry, just saw your post. Don, all I can do is offer some thoughts. I agree with your decision to shoot at 60p. I think that whenever the action has rapid changes, 60p is safer for 3D using two asynchronous cameras because of the worst case timing disparity of only 1/120th second (rather than 1/48th second).

As for authoring, I have mixed feelings. As far as I know, frame interpolation is generally not available with shutter glasses 3D flat panel displays and projectors. So if panning is rapid or the action itself is rapid, rendering 1920x1080p60 captured material at 24p will result in jitter that may be uncomfortable to watch; and may too big a price to pay for the improved visible resolution for static scenes compared with 1280x720p60.

[ I want to experiment with some footage I recently captured in 3D at synchronous 24p that includes a jittery pan and some jittery action. I am going to try rendering it at 720p. If Vegas does a good interpolation job, the result may look better than 1080p24 on my shutter glasses display.]

One of my projects got cancelled so that was a bust but I did do the fireworks from a great distance ( at about 2 miles away), at 28" IA and lockdown at 1080 60p. Next time I plan to do the fireworks using the NEX5n much closer and with the 12mm wide angle lenses. None of these are using live zoom or pans.

I completed the editing of one project which is the Magic Kingdom's Wishes Performance. I shot this from two locations using two wide stereobase TD-10 systems.

One camera set was using 1080 60p and the other was recording at 1080 24p. I mixed the two in editing and rendered three ways-

I could find no artifacts when dissolving a 24p with a 60p. Vegas handled the rendering of these dissolved clips surprisingly well. However, the timeline would not play live without crashing at the dissolve. ( I'm using V11, not v12 )

1. 1080 24p for BluRay compatibility
2. 1080 30p wmv for YT upload processing as I post this.
3. 720 60p as a comparison for Blu Ray playback.

So far the motion is quite smooth on 1 and 2. 3 is still rendering. The wmv is a tad softer due to the rendering at half rez SBS and only 8K Bitrate. It is currently uploading and processing on YT.

I decided not to shoot the Fireworks using the NEX 5N pair because of the one project at that location was cancelled. So, I ended up shooting it at great distance and needing zoom capability so I shot using twin TD10's as opposed to twin NEX5n.

Edited by Don Landis - 10/13/12 at 3:17pm

Don Landis Congratulations your work is impressive. It is very difficult to capture the fireworks in 3D.

I request help in correcting the color of underwater footage with Sony Vegas 11.
I recorded coral reefs to 12 meters deep with unfiltered TD10 and need instructions to correct the color.
Anyone know of any guide.

My legacy is underwater film and video. I'm also a retired PADI Master Instructor and YMCA Gold Instructor with photo and video instructor specialties.

Here's what I can tell you about underwater color. Water is a selective color filter and the more distance between the light source and the camera, reflecting off the subject, the less reds you will have in your exposure. The simple solution to achieving the best color is to use a local light source, usually a mounted strobe or flood light near the camera. To reduce the back scatter of particulate matter in the shot, locate the light source off to one side or overhead aimed down on your subject. In very clear water such as Tropical warm water or underwater caves, you can even place light sources close to the subject and get back farther away from the subject for a wider shot and still maintain color. But the color in a natural light exposure from the sun will contain a fair amount of red content up to about 15 ft maximum. As you go deeper the sunlight from the surface begins to lose reds first then oranges and then yellows so that at about 60 feet of total light distance traveled under water the greens begin to disappear. This leaves only the blue light in the spectrum illuminating your subject in deeper water beyond 100 ft.

With that explanation, please understand that filters can correct the color balance if there is something there to correct. If the deep water has filtered out all the warmer colors, then you will have nothing to rebalance ( assuming the remaining light is all blue light, there are no reds left to build up. Any attempt to rebalance may result in nothing more than a darkening of the blue and give you a darker, browner image as you begin to boost and change blues to browns. This is true whether you use a deeper color correction filter or a try to correct the colors using color correction adjustments in the editing software. You may have some success using colorizing black and white methods where you strip all the color back to a black and white image and then build the colors back into the shot one at a time, like they do when they color old B&W movies. This is a very laborious process.

FWIW- My last U/W video system used two 250 watt flood lights located on each side of the camera, plus one 350 watt flood I would give to my dive assistant to locate in the shot near the subject. Underwater lighting is no easy task but it gives you the best color, far superior to any correction filter except in shallow, less than 10 feet of depth. Here is where the color correction ( rose colored) filter will be better. Not at 12 meters! Smaller lights are commercially sold to underwater shooters but these generally have effective distance only up to about 6-8 feet maximum. My 850 watt system would work up to 30 feet in clear water. But it weighed in at 85 pounds!

Hi Don,
I downloaded the YouTube 1080p version and obtained a 298MB mp4 at 29.97p, which I imported into Vegas Pro. I saw that there was a lot of Vegas plug-in animated horizontal offset adjustment in the first 4 minutes.

There are so many possible choices for horizontal offset! For my own vision I like to use a setting that keeps the Left-Right separation at a minimum for the prominent parts of the content [subject of course to avoiding extreme separations elsewhere in the picture]. So, with the distant view, I would probably have chosen to set the convergence on the distant building, except for the climactic finale when the water is noticeably illuminated. Early in the video (23 seconds) where the boat passes by in the foreground I'd probably have set the convergence on the boat. My partner's vision can tolerate wider offsets than mine.

When watching the video I didn't notice any glitches with frame rate, apart from a possible anomaly at 1min 56.583 sec, where the Left view of the tower is brightly illuminated but the Right view isn't. (This might have had to do with with angles of reflection. Whatever caused it, it caught my eye.)

Thanks for this entertaining video.

Thanks Jeff.

I made no Horizontal offset corrections in either clip that I recall.

In assembling the edit I used my original stereographic auto corrections from last December for the close clip. But, after I uploaded the version you saw, I noticed some problems with the opening scenes where the castle did not converge visibly. It seemed OK on my 32" screen but didn't work on the 105" screen. So, I made some new auto corrections and re-rendered. It looks cleaner now. Note- In Vegas the auto-correction does not affect Horizontal, only vertical, Keystone, zoom and rotation.

Normally, when I slice up a long clip to intercut with the distant shot, I would go back in and verify the auto correct key frames at the beginning and end of the new shorter clip, maybe additional auto corrects mid way as well. These corrections need to be done often especially when I might zoom with the camera or change the IA (rarely I do this).

This whole edit was mostly an experiment to test the ability to mix 60p and 24p video in edit with dissolves. As far as I can see that part was a success. It is possible without obvious problems. But, in future projects I will try to shoot all video in a single frame rate. In January, I do plan to shoot 1080 60p on my wide IA NEX5n rig, wide angle camera rig with a single TD10 ( 1080 60i) in the center to record a close up view of the same scene. Then render to 24p, 60p and 30p as was done for the fireworks.

What still is a huge problem I deal with on every wide IA project is the compromise on horizontal convergence between different screen sizes. What you described.


The recent movie, Prometheus 3D was shot with Red Epic cameras in a beam splitter configuration. Mostly using 4 twin camera locations. In this production I saw some scenes where the stereo did not converge on my 105" screen yet looked fine on my 32" screen. But, this production also was shot in sets that were huge in size, mixed with normally close in shots. Based on this observation, I don't think I'm alone in this screen size conflict. Whether this is an human error or a problem with the science in screen size compromise remains to be answered. The solution so far seems to be to use less IA so that the separation of left and right images is much less than what I like to see in the depth. In other words, go for more of the flat object look and less exaggerated roundness in the scene.

Yes, the manual horizontal offset slider does not move when launching a Vegas stereoscopic auto-adjust. However the actual horizontal offset (and z value for convergence) can end up being affected, sometimes by quite a bit. I think this is because Vegas needs to crop and resize the raw Left and Right frames to implement some of its stereoscopic adjustments..

(Talk about large IAs, I recently tried a 70" angle bracket, supported in the middle with a tripod, and with a webcam attached at either end. Not for use in a crowded location! Most of the footage was unusable - too much stereo content captured at short distances.)

I'm not sure I've explained myself sufficiently.

I am saying that the autocorrect feature for the Sony Vegas Stereoscopic Adjustment plug-in can change the horizontal separation of Left and Right if the source frames require substantial corrections.

In recent times I've used interaxial distances of around 600mm, 1200mm and 1800mm (24", 48" and 70") for two webcams. My workflow has been as follows:

1. Choose a point in the video where the cameras were static for a second or two (i.e. not in the middle of a pan), and capturing distant objects (nothing too close).

2. Activate autocorrect on the stereoscopic plug-in. [This causes at least a slight change in Left-Right separation, and sometimes a considerable change.]

3. Set the horizontal adjustment for convergence as desired. (Often I do this as an animated adjustment along the time line.) It's easy to see the extent of horizontal disparity if using a shutter glasses monitor operating in 3D mode but without wearing shutter glasses.


Don's recent YouTube fireworks video shows a variety of movement in the convergence in the first 4 minutes. See for example the change in the appearance of the castle from the 1min 36 sec mark (Left and Right images of the castle have nil, or near nil, separation) to the 2min 5 sec mark (Left and Right images of the castle have a large separation). I can't tell whether these changes were somehow associated with panning and zooming in the editing but in that particular time interval there seems to be very little panning or zooming.

I can say that with my own editing that I get convergence changes solely through activating autocorrect (no manual change in zoom, etc). And I find Vegas calculates slightly different correction parameters (with slightly different impacts on convergence), depending on the point in the timeline chosen to launch an autocorrection.

The main thing for me is to minimise eye strain. Where feasible, I like to set the convergence on the main object of interest. Of course that isn't always possible if the scene includes both very close and very distant objects and a high interaxial distance). Such scenes may necessitate a convergence in the middleground, even if the main object of interest is elsewhere (foreground/background).
Edited by MLXXX - 10/16/12 at 7:25am

Jeff, repeating, I made some corrections to the closer shot clips that fixed the disparity issues after that video was uploaded. Originally, I was not concerned over the original keyframes in that older clip since my test was mainly to check for issues in the dissolves between clips of different frame rates. I took a look at the specific location you mention and the keyframes did need to be adjusted to correct the changing settings.

The one thing I have observed, is if you set an auto correct for a long clip and then slice it up into smaller clips, you really need to go back and do additional auto correct keyframes as the original settings never seem right after slicing the clip up. I don't think this is a bug, but rather just the nature of the beast.


If the cameras and subjects remain static there should be no need to do any more than one auto correct at the beginning of the clip. If you zoom in or out with the cameras, or have subjects change distance along the depth axis, or pan along a perspective depth line axis, then you need to do an auto correct at the beginning of the change, and at the end of the change. This will create a ramp of correction through the image change. If the rate of image change happens then you may need another keyframe set at the point of change. These corrections are needed for twin independent cameras because it is virtually impossible to achieve 100% perfect alignment at the huge distances wide IA cameras are used for. So, we do final keyframe alignment in post.

The horizontal offset keyframe, if incorrect, can generate lots of ghosting. For my completed projects, I set all the horizontal offset keyframes to reduce ghosting. Generally this balances the primary subject in positive, negative, or zero parallax based on the lens focal length, distances and I.A. I used for the shot.

I have spot checked auto stereo adjustment on self contained 3D clips from my Panny 3D1 and my TD-10's but they never need any adjustment. Only the paired clips from the twin TD10 and the Twin NEX5n's require auto stereo adjust.

My old Sony 3D Bloggie also needed adjustment because it suffered a vertical alignment issue, first spotted by Frank right after I got it. However, I haven't used that camera in over a year.

Sorry to hear about your Z10k being defective. You seem to get quite a few defective cameras. I recall you had several TD10's that didn't align properly too.

One thing I learned in class is if you have to make any adjustments to a paired camera set, then if there is a change in the shot itself where objects move, either by themselves or by moving the camera or zoom lens then you will need to make additional key frame adjustments to compensate for the geometry change. It's just the math. You may recall the classroom exercise used by Sony with the Airplane taxiing on a runway.