Trying to understand frame rates, etc.

[Jonathan Levin]

This is probably the most basic question, but I’m still trying to understand this. Please bear with me. Here is my question (long):

Years ago, when I was shooting film with Super 8, 16mm camera, you had manual adjustment choices for frames per second i.e 12, 16, and 18, 24, 32, 48, and 64. I understand and shot most everything with the “normal” 24 fps.

I understand that at 24fps, each frame is exposed for 1/24 of a second. If I changed that setting to 48fps, not only would I get a slow(er) motion effect on a prjector running at 24 fps, but I would need to increase the size of my aperature on the camera by about one f/stop because each frame is now being exposed at 1/48 of a second.

I never quite understood what the variable shutter lever was for

I understand this basic concept as it applies to film.

I’ve now entered the world of digital video and considering my first camcorder purchase soon.

On a basic camcorder, from what I under stand, you can shoot in aperature priority, shutter priority, or a combination of both.

So let’s say I want to stick with shutter priority of 30fps, which I believe is the “normal” speed for video. I guess my first qustion is why would one shoot in 24p(fps) in video, if shooting at 30p(fps) would show better detail, simply due to the fact that more “frames” are running thru the “gate” than 24? Or why not shoot at 60(i or p) for that matter. Is setting the shutter speed on the camera a different thing than when I read:

“Records in 1080/60i, 50i, 30p, 25p and 24p; in 720/60p, 50p, 30p, 25p, and 24p”

I understand that 1080 is the 16:9 aspect ratio of wide screen and HD and that 720 is the size for “standard” tv.

Correct me if I’m wrong on this.

Don’t modern HD tv’s “project” at 30?

I’m also starting to think that progressive (p) is better than the old interlaced (i) for picture quality. I am not sure if all newer flat panel HD tv’s are now P or I. And why wouldthey not all be P?

On a camcorder like the Canon HV30, I think there is just a setting for shooting eitrher HDV or DV. I’m assuming for the best quality and HD shooting in HDV (1080) is the main choice.

Sorry for the confusion, and whetehr or not my question even makes any sense.

If you can help great.

Jonathan

[Colin McDonald]

Hi Jonathan,

Lots of parts to your question. I'll start it off by answering the last bit if that's OK:

Originally Posted by Jonathan Levin

On a camcorder like the Canon HV30, I think there is just a setting for shooting eitrher HDV or DV. I’m assuming for the best quality and HD shooting in HDV (1080) is the main choice.

There are several signal setup choices on the HV30. it varies a bit whether it's NTSC or PAL but both will do DV at 4:3 and 16:9 as well as HDV options.
HDV does give the best quality overall but there are some tradeoffs in order to get the signal on to the same kind of tape as is used for DV.

It is also possible to record in HDV and "downconvert" to DV later. Some people find this gives the best DV quality.

[Chris Medico]

I think the basic confusion here is the difference between shutter speed and film speed. Shooting 24fps would suggest that the slowest shutter speed you could get for a single frame exposure is 1/24th. So in this case you would shoot 24fps @ 1/24.

With video if you are shooting interlaced NTSC you will have a shutter speed of 1/60. Each FIELD will represent a unique slice of time (since it takes 60 fields to produce 30 frames of video per second). If you shot progressive 30 then you could drop to 1/30 on the shutter. You could also have any shutter speed shorter than the FPS speed too. You could shoot 30P @ 1/50. You could shoot 30P @ 1/500.

Your shutter speed should be chosen based on the look you are after. Do you want motion blur or do you want very crisp high speed detail. In video you have the choice to use any shutter speed you like. Obviously if you dropped from 1/30 to 1/60 you would have to adjust the aperture open a stop to have the same luminance.

Does this on the right track of what you were asking?

Chris

[Steve House]

First of all, understand that there are two television systems in use for broadcast in the world and part of what you're seeing in that list of options for the camera is frame rates for each system. In North America, Japan, and a few other places video uses the NTSC system which runs at ~30fps (in actually 29.97fps). A video 'frame' is in turn made up of two 'fields' interlaced like the alternating slats of a venetian blind scanned one after the other - all of the odd numbered slats in one top to bottom pass followed by all of the even numbered slats. So 30 frames per second equals 60 fields per second, hence the designation '60i' for 'sixty interlaced fields.' In Europe and most of the rest of the world broadcast video uses the PAL system which is based on 25fps instead of 30. PAL is also interlaced in a similar way so 25fps equals 50 fields per second or '50i.' Computer video cards and monitors generally don't need to use interlacing (though some cheap ones do) and so they can scan all the lines from top to bottom of the screen in one pass rather than breaking the frame into two interlaced fields. This is called 'progressive scanning' and is the meaning of the "p" suffixes in your list.

Understanding variable shutter speeds is a little more tricky because the frames rates themselves are engraved in granite. If you're shooting standard definition NTSC video, your video will be at 29.97 frames per second and that's it. But the number of length of time the sensor block is exposed to light is not necessarily 1/30 of a second. You could expose it for 1/60 of a second, then just wait another 1/60 of a second before reading it, read it, and then expose it for 1/60 of a second for the next frame, wait, and so forth. The result would be frames exposed for 1/60 second each, flowing out of the camera at 30 frames per second. Now the process is not a mechanical opening and closing of a shutter - it's an electornic one where an electronic signal tells the sensor block when to respond to the light falling on it and when to ignore it but the principle is the same. (As Chris pointed out, it's really fields exposed at 1/60, not frames at 1/30, but you get the idea I hope).

1080 and 720 refer to the number of lines per screen, not the ratio of height and width. Standard definition NTSC video is 525 lines per screen, though not all of them are used for picture. In computer monitor terms it is the number of pixels found in the vertical dimension. There are two competing HD standards, 720 lines per screen and 1080 lines per screen. Those screens can be either interlaced or they can be progrssive scanned. Because progressive scanning requires a higher data rate than interlaced scanning, 1080p was pushing the technology envelope just a little too much when the HD broadcast standards were set so 1080i became the "higher definition" high-definition standard. Progressive is a little easier to accomplish with only 720 lines to deal with so the other "lower definition" high-definition standard allows for both 720p and 720i scan rates.

[Jonathan Levin]

Chris, Steve and Colin.

Thank you for the most informative posts. I need to read, re-read, and re-read you're answers again, so that I can ask more educated questions.

Let me go over this stuff, (I'm actually going to print this out) and revisit this in a little bit.

Thanks so much for trying teach an old dog.

Jonathan

[Andy Tejral]

Originally Posted by Jonathan Levin

I guess my first qustion is why would one shoot in 24p(fps) in video, if shooting at 30p(fps) would show better detail, simply due to the fact that more “frames” are running thru the “gate” than 24? Or why not shoot at 60(i or p) for that matter.

I don't think this question was addressed so unfortunately, it will be answered by a cynic.

Some people believe that the 24 fps of film is somehow magic. To some, the holy grail of video is to make it look like film. As you well know, film travels at 24 fps.

Now the confusing part: the video camera itself will capture 24 fps but the video recorder needs to operate at 60 fields-30 frames per second. There is some math involved to make this happen. Its called pulldown--the same process needed to transfer motion picture film to video.

So yes, you will get more detail and smoother motion the more frames you shoot. Once apon a once, there was a system called Showscan (I think) that recorded film at 100fps. Yes, it worked great but it went through film at 4 times the rate! I think the economics killed it.

I’m also starting to think that progressive (p) is better than the old interlaced (i) for picture quality. I am not sure if all newer flat panel HD tv’s are now P or I. And why wouldthey not all be P?

Personally, I think the hype over progressive rather than interlace is doo-doo. The reason for interlacing in the first place (back in the '30s) is for data reduction. Digital or analog, you need twice the bandwidth to move a progressive frame over a interlaced frame. Think of it this way: progressive means you transmit the entire frame at once while interlace lets you take twice as long to transmit the same information.

As a long time video person, starting out with the Sony 1/2" open reel Porta-Pac in the '70s, I think a huge opportunity was missed by not standardizing a single format for HD. The 60 fields per second in our current system was based on the power system. Once color tv changed this to 59.94 (as a form of data reduction to allow color TV to occupy the same space as B&W), we lost that necessity.

So instead of the current mish-mosh of formats, there is no technical reason why the entire world couldn't be using one format.

Hope this helps rather than increase confusion...

[Jonathan Levin]

Andy.

Good thoughts.

I thought that learning the diffferent aspects of still photography was confusing, the digital video world has certainly put that into perspective, with it's numerous formats, codecs, aspects, i this p that, HD, SD, HDV, HDDV, AVCHD, and on and on. My head is spinning.

Not to mention how many camera choices, companies. With still digital you have basically Nikon and Canon. For higher end there is the d300 or d3. Canon has four or five I believe.

Video- Panasonic, Sony, Canon, JVC, and Red.

Oye vey. I'm slowly pecking away at it though.

What I now definatley understand is standard interlace is sort of two frames. So one frame of video is actually two interlaced frames. This was how things were from the begining. Then progresive came along, and that one frame of progresive is more like one frame of movie film.

If I'm clear on this, then the p is newer and that with the technology being as new as it is, has not been full adopted by tv makers, but soon will be, along with the i that has been around for a little longer.

In have also heard that editing P video is better because you can truely edit after a frame of progresive, whereas interlaced, because it is two seperate pictures for every one "frame", if there is motion, like a bird flying each interlace will show a motion blur for that one "frame".

I have to go to sleep now.

jonathan

[Andy Tejral]

Originally Posted by Jonathan Levin

If I'm clear on this, then the p is newer and that with the technology being as new as it is, has not been full adopted by tv makers, but soon will be, along with the i that has been around for a little longer.

Well, from an electronic standpoint, progressive is easier so I'm guessing it must have come first. As I stated, interlacing is a form of data compression, needed when lots of people want to broadcast in a limited amount of space. In the early days of TV, the VHF band had a limited number of channels (say 13). Now, there may have been other technical problems with progressive--painting the lines on the tube that close together for example.

And I'll say again, there is nothing magic about progressive. To some extent, you do get more information with interlace because the two fields are captured at different times.

Now, this gets confusing so I'll just point you elsewhere as I cradle my cranium trying to make sense of it myself.
http://en.wikipedia.org/wiki/Atsc and http://en.wikipedia.org/wiki/1080p

In have also heard that editing P video is better because you can truely edit after a frame of progresive, whereas interlaced, because it is two seperate pictures for every one "frame", if there is motion, like a bird flying each interlace will show a motion blur for that one "frame".

Ehhh, well, no, even though an interlaced frame has two fields, they are treated as one unit.

[Scott Brickert]

Hey Jonathan,
It sounds like you want to do it right, ie learning the details well.

I highly recommend a book that answered so many of these questions for me. The original version is "The Complete Guide HDV What You Need To Know" now in it's second edition.

The author, Douglas Spotted Eagle, has his next book out now, called "The Full HD" which I've not read yet.

If you want great looking video, HD(V) is the way to go. The footage the HV10/20/30, or their competitors, take will put a grin on your face. Or, if your pockets are deeper, the option list expands even further.

[David Heath]

All film and TV systems are engineered to as solutions to delivering the impression of moving reality to the human eye - an alien (with different sensory systems) may find any of them very unsatisfactory.

First facet to think about is "flicker". Flash a lamp on and off and it, well, flashes on and off, do it more rapidly and it flickers, do it faster still and it appears *to the human eye* to be steady. Now, let's flash a series of images instead of a lamp, and start off by updating the image every second. It gives a sort of impression of motion - but extremely jerky, increase the refresh rate and the motion gets smoother.

To stop flicker, the rate needs to be around 50 times per second, though the eye is more sensitive in bright ambient conditions than in a dark room. Motion rate is more difficult to specify, about 15 times per second is a bare minimum, but the faster than that it gets the smoother it looks.

In the early days of film, simple cost argued for keeping it as low as possible, and hence the nominal silent frame rate of 16fps. But simply showing 16 images every second would have given unacceptable flicker, hence the use of a three blade shutter to flash each image on the screen three times, before moving the film to display the next frame. Hence a 48Hz flicker rate.

Then came the talkies, and optical sound tracks. They soon found that the linear speed (1ft/s) was not acceptable for audio quality, and needed speeding up. A 50% increase was judged enough, and the figures worked out well - 24 frames per second, each frame shown twice, so flicker rate still 48Hz.

Come TV and the same basic principles applied - with the added factor that early power supplies gave hum, and hum bars were much less objectionable if they were stationary than rolling through the picture. Hence frame rates needed to be locked to mains frequency, so here the split between the US and Europe over standards started.

To take early UK standards, that meant locking to mains of 50Hz, and that corresponded to about 200 picture lines (about 190 active) every 1/50s to keep within bandwidth constraints - what we'd now call 190p. That was thought too low resolution, so hence interlace was born and the 405 line system. Have 405 lines (377 active), but refresh lines 1,3,5 etc 1/50s after 2,4,6 etc, then the odd ones another 1/50th after that. 25fps, but flicker rate 50Hz, and locked to mains.

Originally Posted by Andy Tejral

Personally, I think the hype over progressive rather than interlace is doo-doo. The reason for interlacing in the first place (back in the '30s) is for data reduction. Digital or analog, you need twice the bandwidth to move a progressive frame over a interlaced frame.

Well, as we come into digits, there are advantages in progressive systems, and one such is that they are easier to compress than interlace. Hence, you may need twice the bandwidth in uncompressed form, but when compressed, then it becomes a lot less than twice for progressive.

The 60 fields per second in our current system was based on the power system. Once color tv changed this to 59.94 (as a form of data reduction to allow color TV to occupy the same space as B&W), we lost that necessity.

The 60 to 59.94 change wasn't to do with data reduction, but to avoid colour transmissions giving buzz on the sound in some older monochrome sets. Colour subcarrier frequency has to be an exact multiple of line and field frequencies, and for NTSC (based on 60Hz) the optimum number was found to interfere with harmonics of the sound subcarrier. Moving it to solve the problem meant unavoidably moving the line and field frequencies by the same percentage, hence 59.94Hz as the new field frequency. The problems really started with the invention of timecode......

So instead of the current mish-mosh of formats, there is no technical reason why the entire world couldn't be using one format.....

You may like to look at http://www.ebu.ch/en/technical/trev/trev_home.html and article 311, "The development of HDTV in Europe — a tale of three cities". People did try, and a 40i universal system was proposed, but........

[Jonathan Levin]

Scott and David.

Thanks for that. I will look into getting that book. In the past several months can't tell you how much I've learned from these forums and others that will remain nameless.

Thanks so much.

Jonathan

[David Heath]

Originally Posted by Jonathan Levin

If I'm clear on this, then the p is newer and that with the technology being as new as it is, has not been full adopted by tv makers, but soon will be, along with the i that has been around for a little longer.

You can think of three fundamental aspects which define a TV system:

1] Progressive or interlace
2] Resolution - normally stated as number of lines, 720 or 1080 for HDTV
3] Frame rate

And the notation used is of the form R(p or i)/F, where R=resolution and F=frame rate. Hence two of the most common European systems are 1080i/25 and 720p/50, the US versions being 1080i/30 and 720p/60. (The system used to use FRAME rate for progressive, FIELD rate for interlace, but some years ago it got redefined such that the last number should always be framerate.)

It then follows that if we consider those three aspects IN ISOLATION, then generally progressive is better than interlace, 1080 is better than 720, and motion portrayal depends on subject - 25Hz for drama, 50Hz for sport etc. So logically, the best system is 1080p/50. Unfortunately, the technical demands of that are too high for most of todays uses, so one of the factors has to be compromised. If 25Hz motion is wanted ("film-look"), it becomes easy, go for 1080p/25 - best resolution and progressive nature. Not so easy for fast motion, one compromise has to be made, so it's either 720p/50 or 1080i/25, and the interlace/progressive debate is in reality a debate between those two systems.

There's pros and cons to each, but for a broadcaster putting out a mix of sport and drama it makes sense to keep the same resolution and just change the motion rate. Since 1080p/25 is unquestionably better than 720p/25 for drama, films etc, the tendency is increasingly to then pick 1080i/25 over 720p/50 for sport etc.

That's why talking about "progressive" in isolation can be ambiguous. There's a lot of difference between p25 and p50.

[Jonathan Levin]

To Bill P. and all who have helped, thank you.

I had to discuss this with a film/graphic design friend who is also trying to straighten this out for me. I think he's ready to kill me. Here is what I've learned:

First, my background with movie cameras in the past has been Super 8, mostly on auto, and I then moved up to shooting with a Bolex H-16 in college.

What I now understand is the Bolex could indeed shoot at variable frame rates (FR) i.e 8, 12, 24, 30, 60, and so on. I understand that on the Bolex there is a spinning "shutter". I think the "shutter" on the Bolex was non-adjustable, and just fixed at 1/24th second.

What I now understand (I think) is that if you want to shoot slow motion, say at a frame rate of 100fps, you now have the option to adjust the shutter speed of a current video camera up to 1/100th s. As bill pointed out to me earlier, there are few video cameras in the prosumer range cameras that can actually do this.

I think I understand as well that shooting at a shutter speed less than 1/30th would result in motion blur to some degree.

I also understand if shooting at a frame rate of 30fps, you can't shoot at anything less than 1/30th sec shutter speed, however, you could shoot at a HIGHER shutter speed, like say 1/60th, 1/125th, and so on.

Either way, I think I'm slowly grasping this concept that fps is independent of shutter speed, though you need to think about both when shooting.

So if I may, is this example right?: I'm out shooting 30fps, normal video speed. In manual mode, I want to keep as shallow DOF as possible, so I'm shooting a (hypothetically) f/2. In order for the exposure to be correct, I would then adjust my shutter speed to (hypothetically) 1/125th sec.

Playing this back on monitor will reveal normal film/video motion quality, right?

I hope I'm getting it. Is there a time limit on this forum for teaching an old dog?

Thanks you so much for your help.

Jonathan

[Nick Flowers]

Very many thanks to David for at last...at long long last....enlightening me on why NTSC went from 60 to 59.94. It seemed just so perverse, even though I suspected there must have been a very good reason for it. So now I know, and can tick off another item on my Must Find Out Why list!