Narrow-Bandwidth Television Association

Having built part of the electronics and LEDs etc I wish to test it out.

As the Nipkow is not yet made there seems no point in running the clubs CD etc as I would not be able to make any sense of the fast LED light variations by eye.

So, using a video application I have made a simple .avi file of a white screen slowly fading to black. Using Gary's Video2NBTV I have converted that to a .wav file and onto a CD.

When I get a portable CD player I hope to play this file and be able to see the LED cluster slowly fade out from full brilliance to black, without using any Nipkow as an initial test.
Albert.

How about a simple cardboard-disc from the internet for experiments and tests? They work amazingly good for 10 minutes work and no costs.

DrZarkov wrote:How about a simple cardboard-disc from the internet for experiments and tests? They work amazingly good for 10 minutes work and no costs.

Yes I guess so, but I have no sync circuits in place yet so would the viewing make any sense to me? A simple fade out on the LEDs I can see and understand, and it would be my first NBTV CD production.
Albert.

Of course it's not real pleasure of viewing because the picture will not stand still for more than a few seconds, but it makes sence, because you can test the contrast of the picture which is impossible by a slight changes frame by frame from white to black.

Viewmaster wrote:Yes I guess so, but I have no sync circuits in place yet so would the viewing make any sense to me? A simple fade out on the LEDs I can see and understand, and it would be my first NBTV CD production.
Albert.

It is quite obvious, even without synch, when there is a valid signal/image coming through, and when there is not. You may not see a locked frame, but there are various harmonics of the disc speed at which you see multiple frame images (in an nxn matrix), and besides that, there is a distinctive 'look' to a video signal even when the disc rotation is totally wrong.

So I'd build a temporary Nipkow disk first, worry about syncrhonisation last. It's quite easy to slow down your disk by pinching the edge with your fingers, and achieve frame lock for several seconds at a time -- and certainly this is good enough to get a good idea of how your circuitry is working.

Albert,

Would it be useful if I created a .wav file that would allow testing of the LED system on its own? I'll add in the the full sync stream as well so it could be useful when you get to that stage....someting like the following...

The scale at the bottom is time in seconds, so it would be a slow fade-up over about 14 seconds, then repeated. Followed by a single white pulse, followed by two decending ramps, each 14 seconds.

The syncs will be included, I just haven't shown them.

Give me a couple of days to do it though...

Steve A.

Will an .wav file play on a portable CD player? I get the feeling not, I think you'll need to use a PC with soundcard. But I could be wrong (usually am). Experts?

Steve Anderson wrote:
Would it be useful if I created a .wav file that would allow testing of the LED system on its own? I'll add in the the full sync stream as well so it could be useful when you get to that stage....someting like the following...

That's very kind of you Steve, but please don't spend the time on this on my behalf.
I have a fade in/out which is good enough for present and a temp disc is a possibility as suggested here. Using a video application I can record any levels or change in levels that I may need.

Out of interest here is the full black and full white pulses taken from the fade in/outs, which I have recorded onto a CD taken from Gary's application. (taken using Goldwave)
Albert.

Viewmaster wrote:That's very kind of you Steve, but please don't spend the time on this on my behalf.

OK, no problem, if you change your mind just let me know. I might do it anyway, just for the heck of it. It's starting to get me very familiar in writing/reading binary files.

I'm a little troubled though by the two waveforms. The sync pulses look severely rounded, almost half sinewaves, or sine² pulses. Is it a software-based scope? Do you know what its bandwidth is?

Compare those with the following from a sound card.....ignore the little glitches and bumps, they're possibly caused by some oversampling. A simple filter would get rid of them.

Steve A.

I've had a look at the Goldwave site and there is a limit to the resolution when used as a scope, hence the 'rounding'. This true of most (but not all) audio editing programs. The intended users of these programs are quite happy with millisecond accuracy, not microseconds.

Here's a screen shot of how Audacity displays a few lines of NBTV within a .wav file, seems somewhat better than Goldwave. It's free..

http://audacity.sourceforge.net/

The screen shot is a few lines each side of the missing sync pulse where the waveform changes from decending back to ascending. The display is showing about 15mS (0.015 seconds) of a 60 second file.

Steve A.

Again I feel we're going tangental...

Steve Anderson wrote:

Viewmaster wrote:That's very kind of you Steve, but please don't spend the time on this on my behalf.

OK, no problem, if you change your mind just let me know. I might do it anyway, just for the heck of it. It's starting to get me very familiar in writing/reading binary files.

I'm a little troubled though by the two waveforms. The sync pulses look severely rounded, almost half sinewaves, or sine² pulses. Is it a software-based scope? Do you know what its bandwidth is?

Compare those with the following from a sound card.....ignore the little glitches and bumps, they're possibly caused by some oversampling. A simple filter would get rid of them.

Steve A.

I've had a look at the Goldwave site and there is a limit to the resolution when used as a scope, hence the 'rounding'. This true of most (but not all) audio editing programs. The intended users of these programs are quite happy with millisecond accuracy, not microseconds.

Actually the syncs are indeed rounded. This is as a result of the low-pass anti-aliasing filtering that is necessary when downsampling from 441 samples to 110.25 (per line) for 44.1kHz. Note that when moving from the digital domain to the analogue domain this will occur anyway for a band limited (i.e. real) system. The sync edges will indeed exhibit a sine wave-ish shape since an infinite number of sines waves, from the cut-off frequency upwards, have been (mostly) removed.

The technique, due to Klaas, allows fixed width sync pulses to be added to the video signal and still retain a fixed width video sequence of samples (of course other factors could be used as it is just a form of sample rate conversion).

An advantage of using 32kHz or 48kHz as the sampling frequency is that this downsampling is not necessary as there are an integral number of samples, but of course when converting from the digital domain to the analogue domain the low-pass anti-aliasing filtering will round of the 'square' edges of the sync anyway. And of course these samples rates are not CD audio compatible.

If the signal is being used as a 'scope for an external source then that signal will be band limited by the ADC and you would see rounding of the square edges, but if the signal has been generated by software, then usually (and I can't think of any exceptions) the rounding off or overshoot 'glitches' that are displayed in audio editors are due to the curve fitting algorithms used to make the displayed waveform "look nice". They only effect the display and do not change the way the audio is edited (sample accurate) or sounds. I'm not sure with GoldWave but I would be surprised that if you zoomed in far enough you wouldn't see individual samples with a spline curve passing through them. When you edit you are only effecting the individual samples and the spline curve changes to suit.

I am, of course, referring to audio editors of linear PCM. Editors that work directly on compressed audio such as MP3 are 'packet' accurate rather than sample accurate, where a 'packet' represents a set of, say 1024, samples that have been compressed, so the editor displays something proportional the power of this packet (for instance).

OK everyone, thanks for all that. I will check the sync pulse shape sometime with my scope but I do have a pulse 'sharpener' in place on the motor control circuit. (not yet tested)

This morning I borrowed a portable CD player from a neighbour, hooked
it all up to my LED driver circuits and played that fading from black to white .wav file I made, on repeat.
At first the LEDs just lit up and didn't show any fading at all.

Stupid as I am I then realised after 15 minutes poking around, (and getting very depressed), that I had the 'black level' pot fully up....there are 4 pots on the driver/sync sep board.

Turning the black level down I saw the LEDs light output fading
up and down correctly.

The LED resistances will have to be reduced as the current through them is rather low, so they do not show as bright as they should, but this is a minor problem.
Incidentally I am going to put 50 ohm wire wound pots for the gamma resisitors, to make the adjustments live on a real picture eventually.

As my machine is a combo (camera and viewer), am building the circuits shown in the handbook. They all have to be built onto vero as the club PCB is for the other circuits on the web site.
Thanks again.
Albert.