Struggling with Frame Rates, Telecine and other cryptic terms

[David Gibson]

I'll admit that I naively thought that shooting video was pointing a lens at a subject and pressing the record button. That seems like 10 years ago. Now, I spend my time hearing about using different frame rates, what happens to those in post-production and what the final impact to the video will be. So, I've decided that I need to understand a bit more about the plethora of cryptic terms out there in video-land so I can be confident of what I'm doing when I shoot or edit and not just walking around in a haze of technical stuff.

For example, I use a Canon XH-A1 to shoot. The camera shoots in what Canon calls 24f, which is similar to 24p. My reason was that I wanted to give a bit more a cinematic feel to the video I was shooting. I used the "An*Amateur’s*Guide*to*Handling*p24*material*in*Final*Cut*Express" by Brian Wallace and didn't like the results I got. So, I contacted Brian and as much as he tried to help me, I think the weak link in all this is between my ears.

So, can anyone recommend a book or two that will help me get a handle on what all this stuff really means and will help me make better decisions pre and post? 24p, 30p, telecine, blah blah blah ...

Thanks in advance for the help

dave

[Steve House]

Trying think of a recommendation but there are a number of areas of specialization within the videomaking and it's hard to pin down one or two that cover it all. When looking for the definition of a term, my first stop is Wikipedia.

For the specific terms you asked about in your post...

Conventional video is interlaced. Think of the frame as a window covered with a Venetian blond. The slats on the blind represent the scan lines across the frame. Number them from top to bottom 1,2,3,4... etc. Traditional B&W television scans about 30 frames per second but it does it in two passes. On the first pass it scans top to bottom from side to side along every even-numbered slat. Then it goes back to the top and scans the frame again along every odd numbered slat. When you superimpose the scans, every slat has been scanned and the entire frame painted. Each scan take about 1/60 of a second so one entire frame gets scanned every 1/30 of a second - the frame rate is 30 fps (frames per second) consisting of two 1/60 second fields. Nowadays it may be written either 30 fps or 30i or 60i (for 60 interlaced).

Computer monitors didn't need to use the interlaced scheme. They were capable of scanning everyline in the frame in one pass. Instead of scanning lines 1,3,5 ... then 2,4,6..., they can scan 1,2,3,4,5... The scanning is progressive down the frame and when it finishes scanning one frame it goes on the next. Each frame take, let's say, 1/30 of a second so the 30 fps frame rate is written 30p to indicate it's progressive scan as opposed to 30i interlaced. Progressive can be an advantage because in the older system motion of the subject could take place between the odd scan and the even scan and when the two are combined into one image on a good resolution screen it could appear as visible jaggie artifacts.

Black and white television ran at 30 fps (in North America) because it made timing circuits driven from the AC lines cheap to build. When colour TV came out, the signal needed to included additional components to include the colour information in addition to the point-by-point brightness that came from the scan. For compatibility with B&W sets, it was decided to keep 1/30 of a second as time to transmit one frame of information, but that meant the scan part of the frame has to be shortned very slightly to make room for the colour info. The 1/30 scan was shortened, speeded up by 10% so the scanning part now takes 29.97 seconds. Hence modern TV (in North America - European PAL is differnt) is 29.97 fps, sometimes written 59.94i.

Film-based movie cameras run at 24 fps. The 1/30 second television frame rate was chosen because of the relative difficulty and expense of building accurate timing cicuits back in the 30's when the technology was emerging. But nowsdays building timing curcuits is cheap and easy and there's no reason other video frame rates can't be used if desired. A computer a monitor can be just as happy running at 24 fps as it is at 30 fps. Combining the idea of progrssive scan with a frame rate that matches film cameras leads to 24p.

Canon chose to alter the timing of their cameras very slightly and apply pull-down (see below) so the camera can display its 24p images could be viewed on conventional NTSC screens. To distinguish it from pure 24p they've labeled 24f, "f" for "faux." When you transfer it over Firwire it's true 24p but when you connect the camera to a monitor and view it directly, it's 29.97i

Telecine is the process of taking motion pictures shot on film and re-photographing them onto video by shooting the film with a video camera. Think of pointing the lens of a film projector into the lens of a video camera. This involves some timing issues because the film is 24 fps while the video is 30 fps. If it was just converted frame for frame the result would be a Keystone Cops sort of jerky speed up as the frames would be presented to the viewer too fast. This is resolved by duplicating one of the scan fields every so often. If the film frames are identified by A B C ... converting frame per frame and seeing the video fields (two to a frame for conventional TV, remember) it look like AA BB CC DD. To get it to slow down we could duplicate a field now and then and stick the duplicate in sequence to record it like AABBBCCDDDEEFFFGGHHH (note the 2:3 pattern). Now taking those fields two at a time, we end up with frames on the video that look like AA BB BC CD DD EE FF FG GH HH. Those extra fields mean that we end up with 30 video frames for every 24 film frames. That process of adjusting the speed is called pull-down. There are other variations of the specifics, for example there's also a 2:3:3:2 pull down sometimes used. But the effect is the same.

HTH

[David Gibson]

I get the first part of the reply about interlaced v. progressive. Thanks for the venetian blind (reminds me of Mel Brooks in "History of the World- Part 1" when he says, "how do you make a venetian blind?" and then acts like he's poking someone in the eyes with his two fingers) example is does make that more clear. I'm still reading (on wikipedia) about telecine and pulldowns. That will take a bit more time to digest.

I know this will sound reactionary and it may be but its only because I'm a bit lost in this cloud of frame rates so what I believe depends on the last thing I read ... but I'm wondering about the idea of ever shooting at 24f. Seems like more work for little gain if I'm going to be showing my work on the internet and television. Unless I want to shoot at 60i occasionally for the slomo effects, why would you ever want to shoot anything other than 30fps?

I'll keep reading and thinking about all of this. Thanks for the help. It is appreciated.

dave

[Adam Gold]

Originally Posted by David Gibson

Unless I want to shoot at 60i occasionally for the slomo effects, why would you ever want to shoot anything other than 30fps?

You're getting a couple of things slightly mixed up. 60i *is* 30 fps, sort of. The 60 *fields* per second make up 30 full frames. But here's the catch: each of those 60 fields is (are?) exposed at a different point in time. (A lot of people assume that each of 30 whole frames per second are split into two fields, but it's actually two consecutive dissimilar fields that are combined to make a full frame -- and that's why when you freeze a 60i frame you get interlacing artifacts -- little combing edges that show how the two fields aren't identical and aren't exactly lined up.) So even though they are only half of the resolution of a full frame, they do represent 60 pictures per second. As a result, motion is smoother with 60i than 30p (picture two staircases of the same height and length, but one has 60 steps to the other's 30. By necessity, the 60 steps will each be smaller).

You want to shoot 60i because that's part of the official DVD spec, and 30p isn't. All regular TV is 60i and it's the easiest format to work with in post. I can't imagine ever shooting 24p, but a lot of people like it.

Adding to all this confusion is 60p, which is 60 full frames per second. You'd think this would be ideal, but so far the only way to do this is with 720 rather than 1080 lines. 1080p60 is the holy grail, but it isn't part of the HDV spec, there are no cameras so far that can actually shoot and record this, and no NLEs that could edit it anyway.

The good news is that tape is cheap and you should absolutely experiment with all this and see what you like best, both aesthetically and in ease of workflow. You'll quickly develop your own preferences.

[Bruce Foreman]

David,

My Opinion (Like noses, we all have one):

24 frames per second was the standard with film cameras, but it remains (again my opinion) an archaic holdover from the days when that was necessary.

Most of us now operate in a video environment, but there are other ways to get that filmlike look. There is a "flicker" to 24fps, some can see it, I don't but I remember the "flicker" with 8mm movie projection due to a slower frame rate I could see. I don't feel "flicker" really adds that much if anything to a "filmic" look in a video world. And it has been known to give some people a headache.

But many of our cameras have a "cine" preset that uses a different gamma curve resulting in a more extended dynamic range of image tones with reduced contrast and muted color. This moves a lot closer to the way much film looked but I find it kind of flat. Fortunately the cameras I use now let me "tweak" this a bit and the result is something a lot of my viewers say is quite film like.

Like you, the idea of dealing with the issues like "pulldown" don't appeal to me at all so I just go ahead and work in a video medium, select the mode that begins to give me what I want, and modify it with presets. And what needs more gets it in post processing.

I shoot in 60i most of the time and 30p when there is a need and don't worry about it. Especially since I have no need for a "filmout" version of anything. With more and more venues accepting a digital file for projection, the need for such is bound to lesson significantly.

So for me, 24 frames per second is a "Will 'O The Wisp" I don't need to chase.

[David Gibson]

So, looking at the manual for my Canon Xh-A1, if the Canon is recorded in 60i, then the output is, of course, in 60i. If the Canon is recorded in 30F, then the tape output is 30f but the playback mode (I guess that means directly from the camera through the component out) is 60i. It seems like most of you shoot in 60i. I'll do some testing with that in the coming weeks. But, you also mention shooting in 30f at times. When do you choose 30f? I've been using PRESETS on the Canon for the last several weeks, mostly to get the camera closer to real colors. Looks like my next step is to figure out what gamma and all the other stuff means. Kinda stinks being the new kid on the block. So much to learn.

thanks,

dave

[Bruce Foreman]

Originally Posted by David Gibson

But, you also mention shooting in 30f at times. When do you choose 30f?

60i can look a tad sharper because you have the action covered by 2 fields (potential shorter exposure) per frame. And most of the time I select 60i for this reason as much as this mode is also the default on most of our cameras. And since I use consumer grade cams (Canon HF100's) either of the progressive modes, 24p and 30p are actually in a 60i wrapper.

Yet when smoothness of motion might be the prime consideration I will shoot in 30p, the 2 fields that would be interlaced, even first then odd scan lines, are now combined progressively from top to bottom so there are not likely to be positional differences between the scanned fields. With interlaced you have 2 separate images per frame and the potential for them to be different from each other where in progressive you have one image per frame. But with 30p you can also have a bit more motion blur.

Originally Posted by David Gibson

I've been using PRESETS on the Canon for the last several weeks, mostly to get the camera closer to real colors. Looks like my next step is to figure out what gamma and all the other stuff means. Kinda stinks being the new kid on the block. So much to learn.

thanks,

dave

Dave, just pick up the cam (or better yet mount it on a tripod) and start enjoying messing with it. Shoot, try things, and have fun. Don't get so "hung up" on definitions or learning curves, or trying to get your head around things.

Yes, learn, of course. But also learn by doing (which you are already trying to do). we toss these terms around like we really understand ALL about them, but if you really have to, use Google. Here's what I found on what gamma is by typing in Google search field [gamma and definition]

gamma and definition - Google Search

And an excerpt from the second item down:

"The way brightness is distributed across the intensity spectrum by a monitor, printer or scanner."

So now we both know as much as we did before, right?

OK. If you want something that approaches the film look, set your camera for "cine" mode or preset or whatever it has. Shoot some stuff outdoors and indoors and evaluate it. If your camera has custom settings add a bit of contrast boost, a touch of brightness, and a slight bit of sharpening and see if that gives you what you want. Don't concentrate on what we mean with the "mumbo jumbo".

There will be no pop quiz Monday and no test later in the week.

This is a visual medium

Good luck and have fun.