Variable frame rate.

[Stephen]

I understand that historical standards for television systems were often, but not always, a submultiple of the local mains frequency. At first, it was to allow convenient synchronisation by means of utilising synchronous motors for motor driven scanning elements and later to avoid potential display interference patterns from cathode ray displays to due mains-induced "hum".

However, from an early date television receivers had a means to lock onto the frame rate of the video signal, typically my means of an auxiliary phonic motor. Furthermore, a television receiver with a motor driven scanning element may utilise a synchronous motor driven by a power oscillator phase-locked to the line rate of the video signal.

It seems to me then that it would be advantageous to design NBTV cameras and displays to work over a range of frame rates, wherein the display automatically detects and tracks the frame rate of the camera and the camera may adjust its frame rate according to the available bandwidth of the transmission channel, either manually or automatically.

Ignoring flicker effects that may be overcome by various means, such as electronically by means of memory storage or optically by means of interlacing or Mr. Baird's sequential detail scanning system, television images may display relatively smooth motion with frame rates as low as 10 to 12 frames per second. Frame rates of 5 to 6 fps, in contrast, are somewhat jerky, but not necessarily objectionable. See http://www.privatelessons.net/2d/sample/m01_03.html for an illustration of the difference.

I understand that animators often use 12 fps, and sometimes 6 fps, with decent results for web-based animations. I also understand that frame rates of 15 to 18 fps are generally excellent. Therefore, it would be interesting to be able to adjust a television system to work throughout this range. There seems to be no particular reason to fix NBTV systems at 12.5 fps anymore.

I recollect that the original American NTSC standard of 1941 allowed for variable video frame rate, interlace field count and line count, from at least 15 fps with 625 lines interlaced 4-to-1 (DuMont system) to 30 fps with 525 lines interlaced 2-to-1 (RCA system). Properly designed receivers could automatically track changes in the video signal format. Probably few, if any, commercially produced receivers could actually do this.

It would be interesting to have an NBTV system with a similar degree of flexibility.

[DrZarkov]

Another reason for the common frame rates was (is) the flickering of the studio lights! If you use artifical light by light-bulbs you have a nasty stroboscopic effect if you use another framerate than those of your elctricity net.

[Stephen]

Another reason for the common frame rates was (is) the flickering of the studio lights! If you use artifical light by light-bulbs you have a nasty stroboscopic effect if you use another framerate than those of your elctricity net.

I have heard that before, Volker, and perhaps with ordinary television systems that have scan rates at or close to the mains frequency there might be some such effect. But it seems to me that since the period of an NBTV frame at 12.5 fps is over 4 cycles of the 50 Hz mains frequency and if normal lighting had such an effect we would see a pattern of brightness variation over the displayed image. I have not heard of such a problem.

[Viewmaster]

I have used a smudging technique for moving shots in filmed animation to give a less jerky result and to save animation work (by taking 2 frames per animation movement), which at sound film rate of 24 FPS is nearly the same as NBTV's 12.5 FPS.

Maybe if a very slow rate were to be used for NBTV an electronic 'smudge' technique could be introduced on moving shots for that is where the jerkiness is most obvious.
I did my smudging just with a finger.
Albert.

[Stephen]

It might be possible to have an NBTV system with a "universal" scanning arrangement wherein both the camera and display have separate motor driven optical scanning elements for vertical and lateral dimensions, such as vertical and lateral mirror drums, vertical and lateral lens discs or vertical and lateral radially slotted discs. The ratio of their respective speeds would determine scanning direction, either vertical or lateral, frame rate, number of scanning lines, progressive or interlaced scanning, and if interlaced, the number of fields per frame. The only factor that might be non-adjustable with such an arrangement would be the aspect ratio.

[DrZarkov]

Have you seen the "loudspeaker scanner" on page 10 of the recent newsletter? In theory it should be ideal as an all purpose scanner for everything from 30 lines/12,5 up to 625 lines/25 frames (if any LED is fast enough for that...). Should be great for experiments.

[Stephen]

Have you seen the "loudspeaker scanner" on page 10 of the recent newsletter? In theory it should be ideal as an all purpose scanner for everything from 30 lines/12,5 up to 625 lines/25 frames (if any LED is fast enough for that...). Should be great for experiments.

Yes, I think that is a wonderful device for use as a "universal" scanner if it is responsive enough. It would be important to reduce moving mass as much as possible so that the motion of the mirror can accurately replicate the sawtooth vertical and lateral deflection waveforms.

[Stephen]

Thinking it through a little more, let us say that we had a universal scanning system with something like Edward Nowill's loudspeaker scanners in the camera and in the display. The camera could change frame rate and line count and the display could track the camera signal so that the frame rate could be low with high resolution or the frame rate could be high with low resolution, the limits determined by the transmission channel bandwidth.

The camera could be responsive to a motion detector so that with no or little motion it would operate in a high resolution, low frame rate mode to capture detail. With any significant motion it would switch to a low resolution, high frame rate mode to capture smooth motion. The system could even be proportional so that frame rate would be proportional to rate of motion, and of course line rate would be inversely proportional to rate of motion.

This is just a "blue sky" proposal, but I do not see any inherent obstacles to developing such a system.

[Steve Anderson]

Have you seen the "loudspeaker scanner" on page 10 of the recent newsletter?

I haven't seen the article but it sounds directly related to vibratory scanners, Peter Yancer has a short item about them on his web site...

http://www.televisionexperimenters.com/vibrscan.html

As the item mentions it was diven by sinewaves which is no good. Today with some positional feedback applied (as Philips did with some of their loudspeakers a while back) it should be possible to get a reasonable approximation to a sawtooth action.

If a modulated laser was used as the light source the the mirror could be very small, say 1mm². I think that the laser displays in 'light shows' use something very similar.

But it might be a bit noisy.

Steve A.

[Steve Anderson]

....up to 625 lines/25 frames (if any LED is fast enough for that...).

No problem with the LEDs, they are used in fibre optics which go up to Gb/s. The visible ones might not be so fast but I'm sure plenty fast enough for this.

The hardest thing would be getting a 64µs sawtooth scanning waveform being mechanically generated.

Steve A.

[DrZarkov]

Against the noise it's up to you, Steve: Put the device into a glas valve and evacuate it!

[Steve Anderson]

Against the noise it's up to you, Steve: Put the device into a glas valve and evacuate it!

Actually that's a very good idea! Not only to reduce the noise but at the higher scan rates the mass and viscosity of the air must have an effect.

Steve A.

[DrZarkov]

Yes, I know it's a good idea, it wasn't a joke, but I know you obsession for glasware, so I think it's up to you to make such a device. Actually making your own valves is less difficult than it sounds. The is a german site about valves, http://www.jogis-roehrenbude.de/, he describes how to make your own vacuum pump. And there is of course the second book of H.Peter Friedrichs (ac7zl). I've just read his first book, but I'm going to buy his second book, too.

But back to the thread: I would like to experiment with different framerates. For the beginning it would be a good test, if it will make a big difference to increase the framerate of normal "Nipkow-equipment" up to 16 or 18 frames/sec, that was common with Normal 8 (16) or Super8 (18) and looked quite good, no viewable difference to 24 frames, until the mocements are not extremely fast. And don't forget tests with artifical lights!

[AncientBrit]

Stephen,

A frame rate of 12.5fps locked to mains is certainly beneficial.

My solar cell based Nipkow camera exhibited 4 frame bars on scenes that were illuminated by mains powered lighting.

Incandescent lighting was better than fluorescent lighting but the effect was still apparent.

I was able to null the hum to quite an extent by using a secondary pickup cell outside the disc which gave an integrated pickup of scene lighting (mainly 50Hz) and then using a modulator with this signal fed into the one of the ports in anti-phase.

I wrote this up in the mag about 7 years ago.


Graham

[Viewmaster]

The camera could be responsive to a motion detector so that with no or little motion it would operate in a high resolution, low frame rate mode to capture detail. With any significant motion it would switch to a low resolution, high frame rate mode to capture smooth motion. The system could even be proportional so that frame rate would be proportional to rate of motion, and of course line rate would be inversely proportional to rate of motion.

This is just a "blue sky" proposal, but I do not see any inherent obstacles to developing such a system.

If the motion were of a type which rapidly stopped and started, (like an actor moving his face then not, then moving again), the poor old system would be hunting back and forth would it not, the Nipkow motor taking some milli secs to re adjust speed up and down....or am I misssing something?
And why do we need this additional complication?
As per a previous thread on NBTV signal compression, why do something neat if there is no need, or is it just the challenge?

New thought.....Maybe a NBTV system to carry out real life animation frame by frame using the NBTV camera? Store one frame, arrange the animation, press a button on the NBTV camera and another frame is taken and so on.
Albert.