1. Join Now

    AVForums.com uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

a sideways look?

Discussion in 'Projectors, Screens & Video Processors' started by Chris Bellamy, Apr 10, 2002.

Thread Status:
Not open for further replies.
  1. Chris Bellamy

    Chris Bellamy
    Active Member

    Joined:
    Dec 27, 2001
    Messages:
    96
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Scottish Borders
    Ratings:
    +11
    Biding my time (not patiently) saving for a CRT & scaler, I've had what might be a completely ridiculous thought, but without any credibility to lose I thought I'd air it and hope one of you more knowledgeable people can sort it out for me:

    My thought is as follows - having horizontal raster lines with wide 16:9 images imposes unnecessary strain on the CRT design, since this combines the greatest scanning deflection (wide image) with the highest scanning frequency (tracing out the raster lines). But if the raster lines were vertical, then the high frequency scan would use only 56% of the former beam deflection, and so demanding considerably less current, and giving more headroom within the retrace speed limitations of the projector etc.

    Of course, CRT raster lines are not presently vertical, but could you get more image performance out of a low/mid level CRT when displaying a progressive digital signal if you could turn the projector on its side (don't snigger!) so the raster lines are vertical, and of course suitably counter-rotated the digital image fed to the projector (as easily done with a pc, I imagine), so the projected image remains normally orientated to the viewer, but with the wide axis now at 90 degrees to the raster?

    Is this nuts? Even if you agree the theory is ok, would there be any potential benefit in practice, such as allowing greater field refresh rates or display resolution for progressive digital images? Or would horizontal motion artefacts become more obvious? Obviously this could only be done with progressive digital images, such as created within a hcpc or deinterlacer/scaler, but then that probably applies to most hc CRT setups.

    thanks for any thoughts, I'm sure I'll learn something interesting!

    Chris B.
     
  2. bxd

    bxd
    Active Member

    Joined:
    Oct 17, 2001
    Messages:
    1,482
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    66
    Location:
    Southampton
    Ratings:
    +136
    Chris,

    That's an interesting thought.
    I wouldn't want to comment too much on this issue because I don't know enough about the electrical requirements imposed on a crt to do this. I presume that the current system of horizontal scanning was designed as the best option to give the best compromise between horizontal (left/right) or vertical (up/down) motion.

    You might want to search for information on the Silicon or Grated Light Valve. This is a technology that Sony has bought into. If I remember things correctly... it uses a vertical scan system.

    Brian
     
  3. Chris Bellamy

    Chris Bellamy
    Active Member

    Joined:
    Dec 27, 2001
    Messages:
    96
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Scottish Borders
    Ratings:
    +11
    Thanks, bxd, you're right: it seems horizontal raster is useful with interlaced systems to part cover up the limited ability to resolve motion, and since most motion is sideways, that's why the raster is horizontal.

    I found some interesting papers on scanned linear GLV projection at www.siliconlight.com , especially http://www.siliconlight.com/htmlpgs/glvtechframes/glvmainframeset.html . I had not heard of GLV technology, and never considered their idea of using just one vertical column of single GLV pixels to sweep out the image horizontally as wide as necessary for the aspect ratio. It seems to promise much in the way of performance for low costs. Nevertheless, the technology is still single pixel rather than analogue raster like crt. I did find one fascinating paper on digital video, film and scaling at http://www.microsoft.com/hwdev/archive/TVBROADCAST/convrg.asp#Scan , which had a throwaway aside suggesting using vertical raster scan with crt would be more efficient with progressive images. I found much food for thought in this article, which provides a nice counter (as I understand it) to Charles Poynton's articles favouring 1080i transmission as the most efficient way forward for HDTV, and also is one of the few articles I've found (and understood!) that discuss clearly the limitations of present video and film standards to resolve objects in motion.

    Chris B.
     
  4. Roland @ B4

    Roland @ B4
    Active Member

    Joined:
    Apr 11, 2001
    Messages:
    2,238
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    51
    Location:
    Reading.
    Ratings:
    +32
    Chris
    An Interesting thought but it is probably abit late to start redesigning now. Some of the simulation CRT projectors use a two fold system where the convertional background and detail is draw in the normal scanned way. Then very fine or bright images are drawn seperately on the phospor surface by strearing the beam rather then just scanning it.

    I have installed a projector where the tubes are turned through 90 degrees due to the rear projection room needing to be used as a fire exit. It actually makes no difference to the image. Your brain just needs to work outside the convertional when linning up.

    Your greatest problem is going to one of sources becasue they are all based on a horizontal scan.
     
  5. RichardA

    RichardA
    Guest

    Products Owned:
    0
    Products Wanted:
    0
    Ratings:
    +0
    Vertical raster scan is of course how television started!

    The earliest TV format used the Nipkow disk which scanned the image vertically (well in a slight ) shape really) and used around 40 stripes.

    The problem really omes with the 'benefit' idea that you would reduce scan rate. The problem is that there are more pixels per line than there are lines per frame, so the projector will have to make numerically more scans per frame if it were run on it's side and therefore would actually run at a higher scan rate! and I believe this is exactly why our current TV systems runs as they do, left to right !

    Hope this helps.
     
  6. Chris Bellamy

    Chris Bellamy
    Active Member

    Joined:
    Dec 27, 2001
    Messages:
    96
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Scottish Borders
    Ratings:
    +11
    Thanks, Roland and Richard - very interesting.

    I did not remember at all that the first tv systems used vertical raster (I'm pretty sure I have read that once) - I'm over the age when new information can move in without displacing resident material; the difficulty is that you don't get to choose what is shunted out!

    Chris
     
Thread Status:
Not open for further replies.

Share This Page

Loading...