Thanks very much indeed Chris.

Chris,

When importing that footage used with your mask, I am assuming that you import it as 16:9 and have a 16:9 project loaded to accept it, correct? Are there any problems you have come across with this? Do DVD players ignore the black bars? Any issues trying to go from that 16:9 mask to a 4:3 project?

I actually really wouldn't want to shoot in 16:9 on the camera...It's really not wide screen at all, you don't fit more information into the picture than 4:3, so you aren't actually gaining anything except for pure superficial cosmetics..True widescreen works because it's shot in a way that's easier on your eyes, it is truly wide in regards to more objects in the periphreal being visible, fake 16:9 just takes the 4:3 picture and squishes it...

For instance, if you have want a wideshot of a mountain with two peaks on either end, a 4:3 picture just wont do that justice, you need true 16:9 to fit in the peaks and the actual mountain top for it to look any good.

I seriously stay stick with 4:3 on the VX...

If I may add another option for shooting 4:3, you can do what I did: pick up some Sharpie markers designed for dark surfaces (I got some silver ones, and they're metallic to boot! Fancy lookin', lemme tell ya). Turn the in-camera 16:9 mode on, make four little dashes on the black plastic of your LCD panel that roughly correspond to the corners of the 16:9 frame, turn 16:9 back off, and use that for framing. Still leaves the problem of the viewfinder, but I find it's fairly easy to eyeball it with some practice.

Those familiar with the Memory Mix mask technique, let me ask you, does that mask get recorded to tape? I've been doing the method described above on the assumption that it does (with the mask on tape, I wouldn't be able to reposition my image in post, something I had to do for my DV Challenge short film), but if it's not true, it would certainly make for more accurate compositions.

Truth be told, I've tried the built in widescreen option, and I've been quite happy with the results. Looks pretty darned good, of course I've never been faced with a film transfer. The slight quality hit is hard to notice for most of us, with video on the web (small windows, high compression), but I imagine it may make a difference in your situation. And I'm no expert, but it's been my understanding that the VX2000 widescreen mode, while letterboxed on the viewfinder and LCD, is actually crop-and-stretch when going to tape (or through the cam's outputs); that is, the image isn't simply recorded to tape with the black bars, it's cropped to a 16:9 resolution and then stretched, vertically, back to 4:3. A 4:3 monitor will, therefore, show a distorted image. I suppose you may know that already, but based on the content of your first post, I thought I should chime in just in case. If you can get your hands on a 16:9 capable monitor, you should be able to use the built in option AND get an accurate picture for framing and composition.

Sorry, missed this post.

No, import as 4:3. It is still 4:3. Just has black bars. As I understand it, 16:9 DV is just cropped 4:3. Still 720 wide, but with tops and bottom trimmed.

Robert - I shoot with my PAL VX2000 in the 16:9 mode. I appreciate the fact that both v'finders are undistorted (unlike a lot of Canons and Panasoniocs in the widescreen mode), and the footage plays back off DVD onto my ancient 1990 Sony Trinitron 34" set looking very good indeed.

Of course only the middle 432 lines are being used for the display, but as a lot of braodcast TV arrives anyway in this format, it looks perfectly 'normal' now. If I play the same DVD through the 16:9 Philips TV it automatically switches to fill the screen.

tom.

I must be honest, Tom, I'm not sure what you're getting at; I agree with you, the 16:9 mode looks great to me, and I'm using the NTSC version, lower resolution and all. I'm not sure what it looks like when transferred to film (in the examples I've seen comparing the standard 4:3 image and the built-in 16:9 image, there is just the tiniest drop in quality in certain areas of the picture; complex lines, patterns, and the like, which I imagine will look much worse when blown up to film dimensions), but other than that I'm all for shooting that way. You can't nudge the shot after the fact, in post, but that just encourages care and planning during production, which we should all pay attention to anyway. The big kids, with their VariCams and CineAltas, have to frame it right the first time, we should hold ourselves to the same standard. The only reason I haven't is, I suppose, peer pressure. "Cropping in post provides superior quality", "You've gotta be able to move the frame around during post production", and so on. In fact, I think I'm gonna shoot my next movie with the "inferior" technique, just for the hell of it.

As for the displays automatically identifying the proper aspect ratio of the signal, well, what you describe is the way I thought it was supposed to work; Nick said he had problems with it working correctly, however, and I thought I knew the answer. Apparently not, if 4:3 TVs, even from the era you're dealing with, display things properly. Any idea what his problem might be?

And my marker lines suggestion, well, you ultimately told Nick his best bet would be to shoot 4:3, and I offered my method so he wouldn't have to deal with a permanently recorded letterbox on his footage (using the "Memory Mix" trick). Perhaps there was a miscommunication somewhere, my apologies if I seemed to be contradicting your advice in some way.

To all:

I've noted what I call a "strained" look when shooting the VX2K in 16:9 mode. I don't understand why, because you would think Sony would be doing the same thing as we do with the Memory Stick mask. But for some reason they don't. In fact, it appears some sort of manipulation is being done over and above what normally happens in the 4:3 capture. Some tech Guy can explain it, and the reason, I'm sure.

It is a pain in the neck to have to use memory stick for a mask all the time, and I've often shot just full screen and then masked in post. But thats what we have with this beast. I shot this cam next to my FX1 in some recent things, and I still love the low light latitude versus the FX1.

Trying to get this all:


shooting 4:3 with memory mix = permanent black bars
shooting 16:9 with built in bars = permanent black bars?
shooting 4:3 with marker lines = no permanent black bars, and more adjustable?

Good summary. Works for me.

The second one is incorrect; the viewfinder and LCD are given a letterbox, but the footage is not recorded that way. Some cameras DO record that way, but the VX2000 (and 2100, and PD-150/170) crop the top and bottom of the image, stretch the result vertically to fill the 4:3 frame, and write that to tape (with a flag of some sort that tells monitors and other devices "this footage is stored anamorphically, please unsqueeze it"). The footage as recorded is distorted, and is only unstretched on playback. You may very well NEVER see the distorted version, if all of your equipment is set up properly, but it's there.

More details can be found on Adam Wilt's website, but the important point can be found at http://www.adamwilt.com/DV-FAQ-etc.html, where he discusses the right and wrong ways to do 16:9 ("right" and "wrong" at least from an engineering standpoint; the quality may be acceptable for your tastes):

"The "wrong way" is for the camera to simply chop off the top and bottom scanlines of the image to get the widescreen picture. When you throw the switch on these cameras, the horizontal angle of view doesn't change, but the image is cropped at the top and bottom compared to the 4:3 image (it may then be digitally stretched to fill the screen, but only 75% of the actual original scanlines are being used)."

Emphasis mine. The VX2000 and its family are among the cameras that do the digital stretch he describes.

Which method does everyone think will yeild the most options. I am thinking this.

Use memory mix and just have those with 16:9 use the zoom function.

OR
Use built in 16:9 and hope that it doesn't look like sh^t on 4:3 displays

???? It seems like such a simple answer but I feel like I am making it harder than it is.

I have far too little real world experience to offer advice regarding potential problems with either approach, so someone (namely Tom, since he's participated in this thread already) wiser will have to chime in.

I'm not one to have a big ego, and I'm not looking to push anybody, but if you're looking for versatility, I highly recommend at least trying to frame for 16:9 without any masks from the Memory Mix function. Whether you want to do the little grey marks with the pens that I did is up to you, but I believe the most options would come from a 4:3 source composed so that all major action falls within 16:9 frame boundaries. You can always crop it during editing, compositing, or compression, and if not you'll still have a pleasing 4:3 composition to use for full screen video.

Well, whether you shoot 16:9 with the built-in stretch or shoot 4:3 and crop and stretch to 16:9 in post, it's going to look about the same. That is it will look like 4:3 resolution footage letterboxed rather than the tighter anamorphic 16:9 pixel pattern. It really comes down to convenience.

If you are working with other 16:9 footage and want to pop some VX2000 16:9 footage directly on the timeline without rerendering it, shooting stretch 16:9 will save you a step and look about the same.

If you are using all VX2000 footage you can really do it either way. Cropping and stretching in post will give you more options (you can move the 16:9 frame up or down) and you can do both 16:9 and 4:3 final renders if you want.

Just remember a couple of things. Some things look terrible stretched and interpolated and some things look pretty good. Things like a talking head against an interior background look almost as good cropped and stretched as if they were shot anamorphically while things like leaves or grass look absolutely horrible cropped and stretched. If the points being interpolated are part of a larger line they usually look quite good, but if they are defining a smaller pattern like leaves, grass or a patterned shirt, they can look terrible.

Also, remember that true 16:9 footage and cropped and stretched footage look the same on a 4:3 monitor since the extra lines of resolution are thrown away anyway. The difference between letterboxed and anamorphic footage has more to do with formatting on both aspect ratio televisions than it does with quality considerations. Cropped and stretched 4:3 to 16:9 footage will look no worse than letterboxed anamorphic footage does on a 4:3 set.

Fortunately, in the case of underwater video the interpolation will usually look pretty good. What I would do is just shoot with the VX2000 and not worry about it. If you are doing an all VX2000 project, shoot 4:3 and crop and stretch in post, but if you are mixing the underwater VX2000 footage with topside anamorphic footage, just shoot 16:9 and save yourself an intermediate step.

Also, if you do crop and stretch and slide the 16:9 frame up or down, remember to move it two lines at a time so as not to screw up the interlacing pattern. I do this by zooming out to 50% in Vegas as this automatically means that my movements up or down are going to be an even number.

Okay, one more crack at this. When you use the 4:3 setting, and either lay down the letter box matte in the camera, or lay down the letter box matte in post, you will not have any "stretching" when you rerender it as a 4:3 project. There is no loss of resolution in the picture area. And it plays great on any 4:3 monitor, whether or not it can detect flags for 16:9. Even at this point, the major percentage of TV are still 4:3 aspect ratio.

Now, if you are going to take the 4:3 letterboxed file into a 16:9 timeline, and render as a 16:9 file, you will have to stretch, and the strech will result in eliminating all of the bars at top and bottom.

I think I have it. I am going to film 4:3 with the 16:9 guide marked on the view finder. Than do the rest in post. Thanks a bunch!

I think we're all misunderstanding one another, and further confusing the issue, so let me try and simplify what I was trying to say: if you shoot some footage with THIS camera (I'm not speaking in general terms, referring to any other manufacturer, or any other Sony product) in 4:3, then shoot the exact same footage with the built in 16:9 mode, and compare the two closely, there will be a slight difference. Different, or rearranged artifacts. The nature of the process this camera uses (internal scaling of the picture) produces sometimes-less-than-optimal results.

Is the difference objectionable? Perhaps not, depending on your point of view. Does the picture look okay from a distance? Yeah, sure. Are the two pictures functionally identical as far as most audiences are concerned? It's very likely. Do both versions of the footage look great on a 4:3 monitor? Hell yes! But are we talking about televisions? No. The original poster said he got hired to shoot some footage, with this camera, intended for film. Knowing that a difference in quality, however small and insignificant on television, may have larger repercussions for film--or at least knowing that that's what I've been told by people with more experience than I--I merely meant to caution the gentleman. That's all.

There's an excellent article at DV.com that explains all of this widescreen stuff quite well: http://www.dv.com/columns/columns_it...cleId=59100233 Pay attention to the "Digital Stretch" section, and the part about the PD150 (identical to the VX2000 as far as the image is concerned); the "edgy-yet-soft" detail of the sharpened digital stretch in these cameras is what I was warning Nick about, as while it's true the effect is hard to see in real world shooting conditions when displayed on a production monitor, I have no idea what they'd look like blown up to a thirty-foot screen.

Robert:

I don't think anyone has a problem with the 2 cents everyone is putting in. My problem with this whole thing is I don't get the technical side of it.

From what I can make out, DV 16:9 in the Sony still comes packaged inside a 720 wide frame. And so does 4:3. So it seems to me you should avoid the 16:9 hocus pocus in the camera, and stretch into your 16:9 frame in post where you have more control over the output. Thats why if I know I am going out to 16:9 I would rather choose a matte in camera, over the 16:9 selection. Others just put a tape or guide over LCD to approximate letter box and do it on time line in post. But no matter what you choose, if you go to 16:9 you loose use of some of the pixels, and that is an effective loss of resolution.

I know what you mean; it's hard enough to try and understand this stuff, let alone explain it well enough for others to also get it. Especially since I'm not an engineer.

I believe the article I posted (just edited my post, check it out if you missed it the first time around) does a much better job of sorting all this out than I do--though if you're anything like me you'll have to go over the article several times to fully understand it all. And I think you and I are using the word "stretch" to refer to different concepts. The one I'm talking about is the one presented in the DV.com article.

As for losing resolution, well, as far as I know the only way around that is a 16:9 anamorphic adapter, like the one from Century Optics. Barring that, the only way to get a 16:9 image is to get rid of some scanlines, so "throwing away" resolution isn't so bad: you couldn't use it otherwise. You lose image dimensions when you apply some sort of mask, sure, but not detail in the remaining image area.

Right. And that works expecially nice when you are only letterboxing for a 4:3 output. Well at least this discussion is making it clearer to me, what is going on...

Nice Find
 

Hey all I found this link that gets into nice detail about all the aspect ratios from film to HDTV. Here's the link


http://members.shaw.ca/quadibloc/other/aspint.htm

This one actually has a flash that shows some differences.

http://www.widescreen.org

What I recommend is to try a little of each way to test and compare. When I did this, the camera 16:9 mode looks the same as the post crop and stretch which looks the same as matting the top and bottom and cropping and stretching. No matter how you do it you end up with the same thing: 360 lines of resolution stretched out to 480 lines. As far as I have been able to see, it looks the same no matter how you do it, so you may as well do it with the camera and save the extra steps.

My main camera now is an A1, but I still find myself using the VX2000 for low light shots. When I do this, I just shoot in the 16:9 mode. That way I can put the shots on the same timeline. In bright light the A1 blows the VX2000 away, but in low light, the VX2000 is better, interpolated lines and all.

My personal preference would be to make a mask of transparent coloured lighting gel material with a 16:9 frame cut in it, attach this over the LCD screen and use this to shoot for 16:9 safe image area in 4:3 mode. I assume you are able to use the cam in its underwater housing with the screen opened. If you are using a separate viewfinding device, forget anything I have said.

Which coloured lighting gel works best for underwater light, I cannot tell you, only an experiment can.

An underwater camera is a difficult enough beast to control, so having that little bit of vertical leeway by shooting in 4:3 to recover correct framing in post is worth keeping.

This post is 100% correct. There is such a loss of Vertical resolution that the picture is about VHS quality. Now, put that on a large widescreen!

You're probaly speaking tongue-in-cheek, Lou, but the VX2k in its 16:9 mode is simply miles better than anything I've seen off S-VHS, let alone VHS.

tom.

Thanks for the useful discussion everyone.

I'm set up for the memory mix method now but I'm still recommending to the customer that I shoot 4:3 with the external monitor masked top and bottom with black insulating tape (in lieu of suitable gel).

For film transfer would you shoot interlaced or progressive? Bear in mind my camera is PAL.

Nick

Don't shoot progressive on the Sony. It defaults to 12.5 full resolution fps. Fine if you want to use the camera as a motor-drive still camera, but far too jerky motion for movies.

In general, is the widescreen mode of the VX2100 unusable for anything professional, or can it still be passed off as something good to people that don't really know the difference between it and a camera with native 16:9?

How does it compare to the Canon XL1 and other cameras in its class in terms of widescreen?

The VX2100 does what they term an 'anamorphic' widescreen, although the viewfinders are shown letterbox so there's no horizontal compression distortions. Much nicer than the XM2, say.

I shoot with my VX2000 in the 16:9 mode professionally when the demand is there. Last week for instance I shot a stage show and the 16:9 is a perfect aspect ratio to use. If I'd have shot in 4:3 the bottom of the screen would have been heads of audience and the top of the screen would've been curtains, so effectively my resolution of the actors was unchanged.

tom.

What about going in reverse? Shooting with the in-camera 16:9 and later shrinking it down for 4:3....Anyone?

Just my opinion, but I think the 16:9 digital sampling on the VX/PD series is terrible. My old Canon Elura had a FAR higher resolution 16:9 image - and it used the same basic crop and scale technique. If I were shooting 16:9 on a VX/PD I'd go 4:3 and crop in post. I've got some screenshots of rez charts from my Elura buried on my web site, and its 16:9 was actually pretty decent.

www.philipwilliams.com

I agree Philip, and if you're talking NTSC I'd agree even more with your thought that the VX/PD is not for 16:9.

My PAL 576 lines are reduced to 432 in 16:9 which is just about accep[table on a decent 16:9 TV, but NTSC's 360 lines in the same mode just isn't good enough.

When I replay my stage show footage on a conventional; 4:3 TV it appears masked, but the resolution of the performers is the same as if I'd filled the screen with audience heads and curtains.

Marco - What are you thinking? Go stand in the corner.

tom.

I've been trying to "get" this for a long time. Why is there a need for digital resampling in the VX/PD in the first place. If the DV wide screen and DV Standard are both 720 wide, why is resampling need to turn it to 16:9. It would seem that in wide you just cut the top and bottom off and have less lines, that the picture showing would have the same resolution as a comparable area of 4:3. I am sure I've missed something obvious in my self taught DV 101 class, but its not making sense to me. Anyone ?

It would seem that way based on what one sees in this camera's LCD and viewfinder, and indeed many cameras DO achieve a widescreen image this way. This is also what you are doing when using the Memory Mix function; the image area remaining inside the letterbox has the same detail level as it would in a 4:3 image. But that's NOT how the built-in widescreen function works on the VX2000.

This camera DOES crop the image to make it 16:9, but something else happens before anything is written to tape. The "stretch" I was talking about earlier is that something. The image is cropped, but then digitally resampled vertically to make it a 4:3 image. Due to the vertical resizing, the picture is naturally distorted, but your editing software corrects for that distortion when you import the footage.

Is it strictly necessary? Well, no, I suppose not, which is why so many recommend either the Memory Mix trick you use, or simply shooting a full 4:3 that is framed for 16:9 and cropping after the fact. The aforementioned digital resampling would be bad enough on a progressive image (I think), but performing the same scaling on interlaced footage is even worse. A few posts ago I was saying that it wasn't so bad, apparently I'd forgotten the truth; I went back and looked at some test footage I'd shot using the in camera widescreen mode, and while the nearly-horizontal lines I recorded aren't too offensive, nor is the overall level of detail (only 360 NTSC lines, but it's not too muddy to me), but the ugly, thick black ring around some moving objects is.

I'm not trying to "dis" anybody in this theoretical discussion, because each of us uses practical solutions for our needs. I'm just trying to get the theory through my thick head. This is DV for dummies. So the way I see the 16:9 this is as follows:

1. DV was originally developed as 4:3 medium. 720 x 480

2. As it became popular, someone decided 16:9 should be available in the format, so a something was devised to work that into the same envelope. What that was, I don't know. Because we still have the same number of pixels wide, right?

3. So Sony doesn't want to change its chip, but wants to give its VX2K and later users 16:9 wide screen selection, so it does some kind of electronic magic to widen the existing pixels so it will cover a wide screen if selected. That produces a resolution issue, with some of us.

4. Other cameras, including some Sony models also have a native 16:9 chip. So what actually happens there ? Are the native "pixels" wider, or are there just more of the same size pixels anyway? If the latter were the case, then you would have more electronic interpolation to bring it back to 720 wide.

5. So now, you have this 720 wide file still, that you need to show on a 16:9 monitor. So from what I understand, there is a "flag" in the file that says: "Mr. Signal reader, you need to send me out as 16:9 so the 16:9 TV can get see the video in proper aspect"- and that also tells the modern 4:3 TV to add letterbox stripe top and bottom. Older TVs don't have that capability, so you end up with a "tall effect" in the video output. So what I gather happens there is that 720 field is stretched wider. So that means the signal that was originally recorded was actually squeezed into the narrower 720 width, with some kind of magic that saves the quality we associated with the native 16:9 capture.

Can anybody comment or dispell any of the presumptions in my 16:9 For Dummies process or add to the pixel width question ?

I think that's mostly correct, but I'm no engineer, so I fail to grasp the really intricate details of how these procedures are performed.

NTSC DV is, in my understanding, always 720x480. What makes it 4:3 or 16:9 is the pixel aspect ratio. With 4:3 NTSC, it's .9:1, or thereabouts (it goes to several decimal places, I don't know the exact value) meaning that the pixels are .9 units wide for every one unit they are tall. When you use 16:9, however, the same number of pixels is recorded, but the pixel aspect is set to 1.2 (or something similar, don't quote me on that number), meaning they're wider than they are tall. A device capable of receiving this information and correcting for it will do so, but a 4:3 TV that cannot distinguish between pixel ratios will show a 4:3 picture at all times, regardless of how the footage was recorded.

As for native 16:9 cameras, I think the sensors simply have more of the same shaped pixels. The footage still gets recorded at 720x480, but the finer resolution of the chips provides for a better image (I think downsampling is easier and better looking than upsampling, in most situations).

And yet other cameras, like the PDX10, have 4:3 chips that have more pixels than most; I think that model boasts a one megapixel still photo capability. When taking photos, the PDX10 (and TRV900, by extension) uses the entire chip, and the entire megapixel's worth of resolution. When recording 16:9 DV, it can simply capture data from a wide area and record that. In other words, the chip may be 4:3, but it has more than 720 pixels across its width, and those can be used to generate better looking widescreen footage than this camera can. The details of which and how many pixels are used on the PDX10 for this, however, are a mystery to me, I have no first hand experience with the camera.

Thanks, Robert. The .9 Pixel reference I see in some editing and rendering situations now makes sense.

Matching 16:9 and 4:3 video
 

Hello,

I mistakenly (don't ask me how:)) shot a video shoot this past weekend with one cam (vx2000) in 4:3 mode and the other (vx2100) in 16:9 mode. Are there any suggestions on how to match this in post using a Sony Vegas? I've tried adding running a script that places the black bars at the top and bottom, but of course since I didn't shoot in 16:9 it cuts off the top of some of the video. Is there anything that can be done?

Thanks,
Troy

If you want all in same aspect, you are going to have to open a project (at least in PPro) in the aspect you want, and then trim the non conforming footage. To try to squeeze the footage one way or the other will result in distortion of the non-conforming footage.