[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::atarist

Title:Atari ST, TT, & Falcon
Notice:Please read note 1.0 and its replies before posting!
Moderator:FUNYET::ANDERSON
Created:Mon Apr 04 1988
Last Modified:Tue May 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1433
Total number of notes:10312

86.0. "Request for data on ST video chip" by MYRIAD::DEIGHTON () Wed May 11 1988 10:15

Hi,

	I remember asking in a previous incarnation of this notes file
about the video chip used in the ST. In particular I had heard reference
to the fact that this chip could sync itself to an external signal, thus
allowing one to overlay the ST output with another video source. The
reply at the time I originally asked indicated that there was no such pin
(or pins) on the video chip.

	A recent posting in USENET has re-awakened my interest since the
person claims that the horizontal and vertical sync OUTPUTS could be
re-programmed as INPUTS. Further the author of the note indicated he had
connected two STs together in this fashion (in low or medium res since he
mentioned colour display)!

	Since I would dearly like to be able to overlay text etc from my
ST with video I would appreciate anyone who has register/programming info
regarding the video chip to give me further info. If you know of a book
or manual available (other than developer's kit!) which would help
please give me details (mail TAEC::DEIGHTON)....anyone with the developers
kit should be able to help(?).

	A true GENLOCK capability is the ideal (and expensive) solution,
but with some hardware hacking I may be able to come up with something that
it acceptable for my uses for significantly less investment.

	Thanks in advance,
	Nigel
T.RTitleUserPersonal
Name
DateLines
86.1?LEDS::ACCIARDIWed May 11 1988 12:068
    I think that in order do do what you want, the two video sources must
    have the same horizontal and verticle scan rate.  Only then can they be
    displayed in sync. For NTSC, the horizontal scan rate is 15.75 KHz, and
    the verticle is 60 Hz interlaced.  
    
    I believe the ST color scan rate is 31.5 KHz horizontal.  I suspect
    commercial genlocks for the ST must bump the color frequency down
    to NTSC levels.
86.2I want some info tooHLDG03::VELZENWed May 11 1988 12:0910
    
    This subject sounds quite interesting, so please put any answers
    to this note.
    QUESTION: What is genlocking (is it "only" synchronyzing 2
    video-sources, or is it something like reading video into the ST
    memory)? 
    Someone at Usenet said that genlocking was only possible with a GLUE
    chip from one of the early ST's, any remarks on that?
    
              Greetings, Jan-willem 
86.3Internal/External sync ?RDGENG::KEANEWed May 11 1988 12:1523
    Re 86.0
    
    Hi,
    I also read the usenet mail about the external sync. I was
    also interested in a vague sort of way. I had a look at my ref book
    on the ST. The only thing I could find is that in the register address
    space for the SHIFTER, there is  a register address, one of bit
    of which can be toggled for 50/60HZ and another bit can be toggled
    for quote " Internal/external sync ". There is NO mention of what
    or how this is done.
   
    If anyone can expand on the use of this bit I would be very grateful
   
    By the way if you want the pinning of the video connector I will
    bring the data in and put it in a reply to this note
    
    Cheers
    
    Pat K.
    
    
      
                                        
86.4...LEDS::ACCIARDIWed May 11 1988 16:2914
    I think there is another problem to be solved here... how to create
    an interlaced display.
    
    I know that you can record an RF modulated signal from an ST so
    equipped to videotape, but you can't freeze-frame or edit it.  I
    think this is because videotape stores odd and even frames from
    the interlaced video source as adjacent helical tracks on the
    videotape.
    
    I did hear a rumor that the Mac II can be software selected to created
    an interlaced output signal at NTSC frequencies.  Ignore the fact
    that the Mac II cost 10-15X as much as an ST...
    
    Ed.
86.5Output becoming input makes senseLDP::WEAVERLaboratory Data ProductsWed May 11 1988 23:5710
    Actually, it would make sense that the output would become an input
    if external were selected.  I have no way of testing such a hypothesis
    right now, but would be interested in hearing the results.  The
    problem would be in splitting a composite sync source into h&v to
    feed the ST's separate sync lines.
    
    I was the one who checked and couldn't find any external sync lines
    by looking at the schematics, so the output => input makes sense.
    
    							-Dave
86.6TENERE::DEIGHTONThu May 12 1988 20:5044
Re: -?

	An external GENLOCK device would effectively (I assume being, to
put it mildly, not a hardware expert) synchronise two video sources which
used independant clocks (ie horizontal and vertical sync). I suspect that
this would require frame storage hardware and some fairly interesting
logic to determine when and from what point to dump the frame store by
checking the sync information of the video signal to be synchronised with.
Hence the high price of commercial genlock systems for truly independant
video signals. However by slaving one to the sync signals of the other
it is possible to merely mix the video and use the master sync, assuming one
pays attention to not saturating the input of the display device etc etc.

	Thus if the ST can be slaved to the sync of a video source then it
may be relatively simple to obtain reasonable results for video titling
etc (a commercial genlock for the ST was at the Hannover fair I believe
thus negating the USENET comments about never genlocking the ST....price was
high enough to buy several STs!!! but allows all the nice effects you have
become accustomed to seeing where the weather forecaster is overlaid on a
computer generated map).

	Obtaining the sync signals from composite video is relatively simple
either using discrete components or a single chip plus passive components
(as used by most domestic TVs these days) with some voltage level shifting
to provide TTL levels (I used this for a video A/D I breadboarded but never
completed). The question is what levels would the video chip expect?

	The RGB of the Atari would need to be converted to composite
unless the SCART connector of European TVs is able to take RGB and
composite input simultaneously (which I feel may be the case since I use
the composite input to sync the TV and send the video to the RGB inputs
when using my ST in colour).

	The USENET mail about never genlocking an ST does raise one issue
and that is do current video chips no longer have the external sync ability?
I guess my system is old enough (late 1985) to be more worried about other
problems like DMA and GLUE chips that were never within spec.

	Could I suggest we start a topic related to ST's and video
applications/accessories?

	Regards,
	Nigel
86.7Atari is 15khz colorPRNSYS::LOMICKAJJeff LomickaFri May 13 1988 04:1910
To correct Ed A. - The color video generated by the Atari ST is ordinary
15khz television.  The 31 khz horizontal rate only applies to
monochrome.

I also expect that there will be no problem in feeding synch singals
for an interlaced picture into the Atari - assuming that this works at
all - you will just (correctly even) get the same image twice in
slightly different locations.  It would make the horizontal "zebra
stripes" go away.

86.8LEDS::ACCIARDII Blit, therefore I am...Fri May 13 1988 11:3621
    Thanks for the correction... I knew that the ST had 4X the total
    bandwidth of the Amiga in 640 x 400 mode (31.5 KHz H x 70 Hz V versus
    15.75 KHz x 30 Hz V for the Amiga).
    
    I just remembered that some people have used an Atari color monitor
    on their Amigas, which of course implies that same scan rate.  (Sound
    of hand hitting forehead)
    
    Speaking of interlace, there's an interesting card for the Amiga
    that provides a rock steady 31.5 KHz 640 x 400 display.  It's properly
    called a de-interlacer, since it has it's own RAM and combines two
    interlaced frames, and bumps the scan rate up to 31.5 KHz.  It steals
    no CPU time, since the Amiga video hardware is not affected.  The
    fesults are fantastic, actually better than a color Mac II.
    
    However, for the downside, animations can sometimes appear to be
    a little choppy, since any the smallest time interval of motion
    that can be visually detected is now 1/30 of a second instead of
    1/60 of a second.               
    
    Ed.
86.9More on sync from Germany.RDGENG::KEANEMon May 16 1988 08:5784
    
    Hi,
    
    	Heres a bit more on the internal/external mystery !!
    
Newsgroups: comp.sys.atari.st
Path: decwrl!sun!pitstop!sundc!seismo!uunet!mcvax!unido!tub!tmpmbx!netmbx!hase
Subject: synchronizing the ST to an external video source
Posted: 14 May 88 04:53:28 GMT
Organization: netmbx Public Access Unix, Berlin
    
Posted: Sat May 14 05:53:28 1988
    
    
I think this is interesting not only to Michael Kieras, who asked me
about "gelocking" the ST to an external signal.
Please hit "n" if you do not know the right side of the soldering
iron... 
    
>Hartmut Semken, 
    
Just call me hase, like everyone does..  
    
> I read with extreme interest your letter on using the Hsync and Vsync
>lines on the ST as a genlock input.  But, since I really know nothing
about the subject, I am left with many questions.  
    
I'll try to help You; does nobody else know about this?
    
> First, I am guessing that Hsync and Vsynch apply to the RGB lines.  
    
No.  They are just the synchrinisation lines for the ST video output; a
pulse on HSync tells the Monitor (RGB or monochrome) to start a new scan
line, VSync tells it to start a new screen.  The frequency of 35 kHz on
the HSync line (in monochrome video mode/high resolution) can be
hazardous to color monitors (they exspect 15 kHz on this line) so be
careful (a friend of mine killed his green monitor this way..) ! On the
RGB lines the ST transmits the brightnes of the current pixel in its
red, green and blue components.  After pulsing HSync the ST starts to
transmit a new line of 620 or 340 pixels.  These are the basics; ask
your local TV-freak for details (ok?).  When operating "normal" the ST
is source of the sychronisation signals: the GLUE/Shifter chips decide
when to start a new line/picture and tell the monitor so (via H/V-Sync). 
When programmed for external synchronisation, the Shifter (shifting out
the pixels serially after HSync, hence the name) takes the sync lines as
inputs and synchronizes his shifting (and thereforethe ST picture) to
the external signal. 
    
Now all you need are 2 VCRs with EURO AV (also called SCART) connector
or any other [R+G+B+HSync+VSync+ground] input/output, a soldering iron,
some wires, ST monitor plugs (about $3 around here in Germany) and a
weekend: connect the VCRs and let one play (so it becomes SOURCE of a
video (and synchronisation) signal) and the other one record (so it will
synchronize on the input signals, you can use a monitor with RGB inputs
here).  If you now connect the RGB and Sync lines of the ST to the RGB
and Sync lines of the VCR/Monitor *and* set the ST for external sync
(how? hmm, its somewhere in the Abacus book "ST Internals" of
Brueckmann/Englisch/Gerits, I think) and (voila!) the pictures are
simply added (the brightness values for each pixel are added).  To
protect the ST you can do the connection over
one-Microfarad-capacitators (almost never necassary but cheap and
secure).  That is all. 
    
This "genlock" cannot do bluebox tricks, of course.  It is just possible
to add the two pictures by synchronizing the ST to an external video
source (like a camera to put in titles or the like).  A bluebox genlock
is a lot more difficult to make (and too much for my little knowledge..)
and more expensive... 
    
Something wrong about this?
I don't think so; I tried it out: We connected two STs to the same
monitor, and programmed one for external sync.
They both ran NeoChrome with the same picture but slightly different
palettes. Looks great (and makes Amiga-Freaks run for shelter..:-).
Still something wrong? I'm not perfect... Correct me!
    
hase
--  
Hartmut Semken, Lupsteiner Weg 67, 1000 Berlin 37 hase@netmbx.UUCP
I think, you may be right in what I think you're thinking. (Douglas Adams)
   
    
    
    
86.10HOGGAR::DEIGHTONWed May 18 1988 21:1228
I will now search my recent USENET mails for this, basically this fits the
lines I was thinking along except that camcorders don't support SCART
connectors (otherwise they'd be the epitomy of the Japanese idea of portable
Ghetto Blasters!). I have one problem and that is that my system is so old
(early?) that it was never equipped with composite sync and the SCART
connector does not support horizontal and vertical sync seperately. I had
to build a SCART connector with diodes to mix horizontal and vertical
sync and a variable pot to bleed some of this into the another input
(whose function I cannot recall offhand). Even this failed until I realised
that Atari shipped these system with a disk version of TOS that set the
video to NSTC frequencies instead of PAL.....no wonder the TV lost sync
every other frame!

I will eventually (maybe this weekend) breadboard something to extract
horizontal and vertical sync from a composite sync signal and see if I
can blow up the video shifter on the ST.....I still don't know the
voltage levels the thing expects in external sync mode but I can start
small and work my way up.

If my son lets me play for a while I'll post the results, in the meantime
keep the info coming.

Thanks,
Nigel

The best solution wins a titled video of my son eating carrot puree
(disgusting to all but hardened parents, doctors, veterinarians and close
family!).