ST video-output bandwidth specs?

[Pacman]

I'm trying to determine the signal bandwidth of the three Atari ST video output modes.
What I need to know in order to find out is:

A) number of rows per frame (including blanking and retrace intervals)
B) number of pixels per row (including blanking and retrace intervals)

Does anyone know or can point me to a source?

[Cyprian]

ST-LOW 50Hz:
8MHz Pixel Clock
313 lines per frame (200 for a screen)
512 pixels per line (320px for a screen)

ST-MID 50Hz:
16MHz Pixel Clock
313 lines per frame (200 for a screen)
1024 pixels per line (640px for a screen)

ST-LOW 60Hz:
8MHz Pixel Clock
262 lines per frame (200 for a screen)
508 pixels per line (320px for a screen)

ST-MID 60Hz:
16MHz Pixel Clock
262 lines per frame (200 for a screen)
1016 pixels per line (640px for a screen)

ST-HIGH:
32MHz Pixel Clock
501 lines per frame (400 for a screen)
896 pixels per line (640px for a screen)

[Pacman]

Awesome! Thanks for sharing :)

Apparently the way to calculate the maximum video frequency is: H x V x f/2
(source: Texas Instruments -Signal bandwidth vs. resolution for analog video)
H: pixels per row (including blanking and retrace intervals)
V: rows per frame (including blanking and retrace intervals)
f: refresh rate (frames per second)

..so if my calculations are correct along with the data you supplied here, the numbers should be as follows:

Signal bandwidth -Atari ST video output.png

PS: may I ask where you found those numbers you posted?
All I've ever found regarding the ST video output is the screen size and H-/V-Sync.

[Cyprian]

Actually, those figures are well know for the Atari coders. They are important for the code synced with the beam, e.g. to open left/right/top/bottom border.


Regarding Vertical and horizontal refresh rate, to calculate them you have to know the main clock. There are a few slightly different main clocks, e.g. for PAL, NTSC, ST and STE: More you can find them there:
https://www.atari-forum.com/viewtopic.p ... 27#p335127
https://forums.atariage.com/topic/35376 ... y-choices/

E.g. PAL ST / STE:
Main Clock: 32.084988 MHz

ST-LOW/ST-MID 50Hz:
Hsync: 15666,49805 kHz ( = 32.084988 / 4 / 512 )
Vsync: 50,05270941 Hz ( = 32.084988 / 4 / 512 / 313 )

ST-HIGH:
Hsync: 35809,13839 kHz ( = 32.084988 / 896 )
Vsync: 71,47532613 kHz ( = 32.084988 / 896 / 501 )


--- EDIT ---
I'm not sure about number of scanlines for 60Hz mode. I see on Atari-Forum 262 vs 263.

--- EDIT 2 ---
It seems 263 is correct for 60Hz

[Badwolf]

Awesome! Thanks for sharing :)

Apparently the way to calculate the maximum video frequency is: H x V x f/2

Now I don't know much about analogue filtering, so shoot me down if you will, but looking at the physics, this seems a complicated approach when you know the pixel clock directly.

ST Low is 8MHz, so that would put the max video frequency at 4MHz, I think.
ST Med is 16MHz, so that would put the max video frequency at 8MHz, I think.
ST High is 32MHz, so that would put that max video frequency at 16MHz, I think.

I'd have thought the low figures would be determined by printing an entire line of one colour. So that's basically the horizontal refresh rate of 15.6kHz for PAL and 15.7 for NTSC (and around 31.5kHz for ST High).

The video bandwidth would therefore be the former minus the latter.

Ie. 3.98MHz for ST Low, 7.98MHz for ST Med and ~15.97MHz for ST High.

...which are actually very close to your figures. So fair enough! :lol:

BW

[Pacman]

Awesome! Thanks for sharing :)

Apparently the way to calculate the maximum video frequency is: H x V x f/2

Now I don't know much about analogue filtering, so shoot me down if you will, but looking at the physics, this seems a complicated approach when you know the pixel clock directly.

That may be. I have to admit complete ignorance on the subject, so I wouldn't know if that's it or if there's a better way.
I just need the numbers in order to find suitable spec'd components for a video output board I'm trying to build.

...which are actually very close to your figures. So fair enough! :lol:

:lol: Not bad for someone (me) who has no idea what he's trying to calculate!

@Cyprian : thanks for the links. I read through the postings, but honestly it just confused me.
It might be that different terms are being used for the same things, or that those discussions were mostly concerning the sync rates (where I understand that the Vsync rate actually does make a difference in the bandwidth I'm trying to determine), and not so much the "horizontal pixels per line" and "vertical lines per frame".
Nevertheless, if the differences are minor, can I conclude with the following numbers being approximately correct for the video signal bandwidth?

Signal bandwidth -Atari ST video output2.png

[ijor]

Apparently the way to calculate the maximum video frequency is: H x V x f/2 ...
Nevertheless, if the differences are minor, can I conclude with the following numbers being approximately correct for the video signal bandwidth?

I'm not sure what is the purpose of this, because it might be misleading. The formula is for the maximum frequency that would be transmitted in a SINGLE cable wire, not really the same as the signal bandwidth.

For starters this doesn't consider color resolution. It assumes the signal is monochrome, or that you transmit each color resolution bit on a separate wire. So the formula is per wire, not total (again, unless it's a monochrome signal).

Also there is a technical difference between frequency and bandwidth. The (maximum) bandwidth is just the pixel clock, not divided by 2, and multiplied by the color resolution. Depending on the application the bandwidth might consider the active video only, without the borders.

The frequency is interesting mostly for analysis regarding the cable itself, or whatever means used to transmit the signal. Not so much for actually processing the signal.

[Pacman]

I'm not sure what is the purpose of this, because it might be misleading. The formula is for the maximum frequency that would be transmitted in a SINGLE cable wire, not really the same as the signal bandwidth.

I'm trying to find out which multiplexer chips are suitable for my adapter, so I wanted to find out which minimum requirements need to be met. It appears that one which is capable of the frequency range from around 50 Hz to 16 MHz will suffice.

[ijor]

I'm trying to find out which multiplexer chips are suitable for my adapter, so I wanted to find out which minimum requirements need to be met. It appears that one which is capable of the frequency range from around 50 Hz to 16 MHz will suffice.

Ah, yes, that makes sense. Note that for mono it's slightly over 16MHz (not 15.98 as in your table). The math here is quite simple, it is as simple as the pixel clock/2. And for mono the pixel clock is just the main oscillator frequency, which is slightly over 32 MHz depending exactly on the model and PAL vs NTSC.

[Badwolf]

Also there is a technical difference between frequency and bandwidth. The (maximum) bandwidth is just the pixel clock, not divided by 2, and multiplied by the color resolution.

I can't see why it wouldn't be divided by two. The pixel clock goes through a full cycle for each pixel, but the frequency that the pixels can go from black to white is half that. Ie. the clock goes hi-lo-hi-lo to make the luma go hi-lo (at the maximum rate).

My answer was based on a mono composite cable as RGB would just be that tripled up (but thinking about it, a lower bottom end frequency (as no combined sync on each wire).

I did also briefly consider colour composite, but wasn't sure how much lower the chroma resolution is. Upon reflection the colour bust would probably be considerably in excess of the luma max bandwidth, though.

But since this turns out to be about multiplexers, it's a bit moot!

BW