Camera Requirements for Internet Display

[J. Anthony Gonzales]

I am looking at working on an episodic Internet project. Does anybody have any advice as to the best camera for the job? There are two project types I'd be looking at:

 

1. Internet only.

2. Internet and DVD distribution at a later date.

 

The budget allows for pretty much any format, BUT I don't want to be overkill if the resolution doesn't require it. I'm thinking DV or 16mm, my friend thinks we can go even lower on the resolution.

 

Any advice? Thanks for your time and help.

 

John G.

[Phil Rhodes]

Hi,

 

Depends very much how you're distributing it. If you're going to make 1920x1080, two-megabit MPEG-4 streams available, then you need to be shooting HDCAM or 16mm. If it's realplayer, DV's fine, but remember that better cameras are also better in areas other than resolution - highlight handling and image control are still worth having irrespective of the resolution you're viewing at.

 

I wouldn't go any lower than DV, if only because anything else is liable to be analogue and noise or instability is pathological for most video codecs. Video compression is a poop in, poop out process, so the better data you can give the compressor, the better your result will look. Also, strongly consider shooting a progressive-scan format, as computer displays are progressive and anything else will need deinterlacing before it's feasible to compress it for the net. This won't hurt you when you come to make the DVD as it'll do either.

 

And of course, if you're going to release it on DVD, the technical requirements are every bit as stringent as broadcast TV.

 

Phil

[nchopp]

BetacamD or SP could be possibilities as well. Personally, depending on the nature of the project, I'd probably go with the DVX-100. Cheap, looks good if you're good with light, and it'll translate well to both computer distribution (it's progressive) and DVD.

[David Cox]

I would avoid 16mm film as the grain in the picture will be seen as "added detail" by the software used to make the highly compressed web-streaming versions. This result is lower final stream quality for a given stream file size. The reason for this is a lot of compression systems work on the basis of only sending what is different between frames. So even in a locked off shot, the grain between frames represents a significant difference to the encoding software.

 

If you were shooting just for the web, keep that frame difference thing in mind. If you run around with a hand held camera, or fill your frame with highly detailed moving elements such as a forest full of trees and leaves blowing in the wind, your final stream will be a lower quality than if your sequences are more similar frame to frame.

 

Thats why when ever you see demos of low bandwidth streaming services, its always with someone small(ish) in frame, against a static background with them being told not to move too much!!

 

David Cox

Baraka Post Production Ltd

www.baraka.co.uk

[J. Anthony Gonzales]

Thanks guys for all the info.

 

This project involves a static, locked down camera in kind of a minimalist sitcom environment. The set pretty much stays constant unless one of the actors moves something around. We plan to do this is a slightly high, wide shot. Actors, at most, would take up about 1/2 the height of the frame. We are leaning more and more towards DV because of the ease of post for a project of this type, although we have some basic 35mm equip available though I think it be a bit much for the project.

 

 

John G.

[Phil Rhodes]

Hi,

 

I tend to agree. Use standard video tricks with your DV kit - underexpose, light for luminance range, grade carefully, and you're good.

 

But do shoot progressive, even if it's an NTSC XL1 and you're stuck with 30p.

 

Phil

[Kitao Sakurai]

One other thing, you'll probably want to compose for tv safe (in case it ever makes it on to dvd, say) but make sure that you protect for the full frame of the video image, as that's how it'll look viewed on the internet. Reallly use the underscan on the monitor that you'll be using on every shot to keep in mind how it'll show up on the internet. I recently did NOT do this on a rather run-n-gun music video shoot with an XL2 with the P+S mini35 adapter, and the shots that I used a Kamio ring-light (that attached directly to the cooke primes we were using) showed the kamio creeping slightly into the corners of the frame... not visible at all when viewed on a tv monitor (which is what the spot was intended for) but when posted on the intenet it was visible. That aside, I thought that the XL2, shot 16:9 24p with the 2nd generation P+S adapter and a set of cooke S4 primes was really some of the nicest looking video I've seen.

 

Kitao

[Phil Rhodes]

Hi,

 

I really think that the entire frame of video should be protected always.

 

The issue with internet video, or even DVDs, because people watch them on their computers and the replay software makes no attempt to simulate overscan (which I consider to be a technical flaw) - you end up with this very loose flappy framing. This is one reason I shoot so much 4:3 composed for letterboxing, because at least you then know where the headroom is going to be. Probably 75% of everything I ever shoot for internet or CD-ROM distro (which is a fair amount of the stuff I shoot, period) is done this way for exactly these reasons.

 

To expand on the topic, the issues they have with graphic design on 16:9 news composed for 14:9 broadcast are quite intractable - watch BBC News 24 off the website and some point and you'll notice how the lower-third ident and clock end up floating somewhere in the middle of the screen, once you've taken overscan and the aspect ratio conversion into account. Quite a dismal state of affairs.

 

Phil