Olympus-OM
[Top] [All Lists]

Re: [OM] Super8 vs VHS, NTSC, & PAL

Subject: Re: [OM] Super8 vs VHS, NTSC, & PAL
From: Joe Gwinn <joegwinn@xxxxxxxxx>
Date: Thu, 26 Dec 2002 16:19:33 -0500
Comments interspersed below.

The table and page references given below come from "Television Engineering 
Handbook", K. Blair Benson, editor in chief, McGraw-Hill 1985.  We've been 
having the house painted, and my copy was inaccessible until now.


At 10:04 AM +0000 12/25/02, olympus-digest wrote:
>Date: Tue, 24 Dec 2002 18:29:09 +0100
>From: "Carlos J. Santisteban Salinas" <cjss@xxxxxxxxxxxxx>
>Subject: Re: [OM] On: Dipping our Toe Into Digital (Super8 vs VHS, NTSC, & PAL)
>
>Hi, everyone.
>
> >It's true.  PAL has 625 lines of resolution, compared to 525 for NTSC,
> >so the resolution is better.  Not that either ever really achieved their
> >full theoretical resolution.   Especially on 1/2-inch consumer-level VHS
> >videotape.
> >Many US TV sets had more like 250 lines (pixels).
>
>Don't forget the 525-625 lines are _vertical_ resolution (horizontal
>lines), whereas VHS' 250 lines represent the _horizontal_ resolution (pairs
>of points/pixels on a line). From the 625 lines of PAL, only 575 are
>visible because the rest is needed for sync pulses -- most TV sets show
>even less. I'm not sure about NTSC values, but it should have around 480
>useful lines.

Yes.  See next item.


>Same for horizontal resolution: the screen only shows about 800f each
>line. So, 250 "lines" = 500 pixels would show about 400 _visible_ pixels.

Table 4-2 gives the following summary:

NTSC:  340 vertical by 330 horizontal, 150,000 resolution elements per raster.

PAL:    400 vertical by 390 horizontal, 210,000 resolution elements per raster.

SECAM:   400 vertical by 470 horizontal, 250,000 resolution elements per raster.

These resolution elements are more or less tricolor, although the color 
bandwidth is far less than the luminance bandwidth.  In NTSC, the luminance 
bandwidth is about 4.3 MHz, while the chrominance bandwidth is 1.5 MHz (for 
"I") and 0.5 MHz (for "Q").  (Figures 4-5) 


>Another issue is colour resolution. The chroma (colour) signal has _much
>less_ horizontal resolution than the luminance (B&W) signal. I've read
>somewhere that for the PAL broadcasting (4.43 MHz colour subcarrier) the
>figures are 4.5 MHz for luminance and only 1.5 MHz for chroma -- that's 288
>and 96 "lines", respectively. I'm afraid that some digital cameras do
>something similar.

 PAL will be similar, but in PAL there is no difference between the I and Q 
bandwidths, so colors are somewhat more accurately painted.  Except that the 
alternating phase scheme greatly reduces vertical resolution of color.


>When recording to tape, the thing gets much worse... VHS downsamples colour
>subcarrier to just 627 kHz, so it's no surprise that colours are rendered
>as fuzzy blobs ;-). IIRC, Beta and Video-8 are somewhat better with 689 and
>732 kHz, resp. SuperVHS keeps the 627 kHz, though. Playback circuitry also
>does "nasty" things with chroma signal, too.

Yes.  This is compared to 1.5 MHz.


> >The big difference is that NTSC codes color as the absolute phase of the
> >color subcarrier, while PAL uses the frame-to-frame difference, so PAL color
> >is far more robust.
>
>AFAIK, the reason is a bit different... PAL uses a, say, "negative" coding
>for even lines, instead of the "positive" coding of the odd lines of each
>field. Phase errors will make a colour shift, like NTSC, but with opposite
>directions between even/odd lines, so the global effect would be negligible.

You are right; I misremembered it.  The problem with PAL's alternating-phase 
scheme is that the resulting averaging reduces the vertical color resolution 
substantially, basically by a factor of three.  (page 4-43)  The advantage is 
rock-solid color, despite the vagaries of free-space transmission in cities.  
The tradeoff is a good one, given that low resolution is far less disturbing 
than color oddities and busy edges.


>On the other hand, SECAM colour system has a completely different approach:
>the colour subcarrier is frequency-modulated. But it halves colour
>_vertical_ resolution and compatibility with B&W sets is not as good as in
>NTSC/PAL.

Yes.  SECAM is not much used outside of France (and the former French 
colonies?).


> >I don't know the image size of Super8, but if it scales from 35mm movie film,
> >at 24x18, then the image size will be something like 1/4-size, or 6mm by
> >4.5mm,
>
>Here's a list of movie film formats:
>
>8mm: 4.9 x 3.55 mm
>Super-8: 5.36 x 4 mm
>16mm: 10.05 x 7.42 mm
>35mm: 16 x 22 mm
>70mm: 52.2 x 23 mm

Thanks.  Where did you get the information?

8mm: 4.9 x 3.55 mm =   17.395 square millimeters.  1:1
Super-8: 5.36 x 4 mm =  21.440 square millimeters.  1.233:1
16mm: 10.05 x 7.42 mm = 74.577 square millimeters.  4.287:1
35mm: 16 x 22 mm =    352.0 square millimeters.  17.395:1
70mm: 52.2 x 23 mm = 1200.6 square millimeters.  69.02:1

The 35mm still picture format is 24 x 36 mm= 864.0 square millimeters.


> >which is the rough equivalent of (6*100)(4.5*100)= 270,000 tricolor pixels
> >per frame, or 0.54 Mpix, as digital cameras are rated.  In terms of lines, 
> >this is
> >equivalent to 520 by 520, far more than VHS ever could do.
>
>Based on the above table, and assuming a conservative 50 lp/mm (the limit
>may be higher with the short focal lenghts of Super-8), my calculations
>give 214400 tri-color pixels, or about 230 lines -- roughly the same of
>VHS. However, colour reproduction will be better on film.

With film, all colors have the same "bandwidth" (that is, resolution).  

Super8 is (5.36*100)(4*100)= 214,400 tricolor pixels, actually about the same 
as PAL (as broadcast). With film, all colors have the same "bandwidth" (that 
is, resolution), so the PAL won't have quite as good color as film.

The real problem is that consumer video recorders are far worse than what the 
NTSC and PAL signals can do, so even Super8 blows most or all consumer video 
recorders away.  Digital recording of High Definition TV (HDTV) should solve 
this problem, if HDTV ever really arrives.


Joe Gwinn 


< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >


<Prev in Thread] Current Thread [Next in Thread>
  • Re: [OM] Super8 vs VHS, NTSC, & PAL, Joe Gwinn <=
Sponsored by Tako
Impressum | Datenschutz