CurtPalme.com Home Theater Forum

mistagreg · Joined: 09 Jul 2008 Posts: 47

I see 1080p @ ##kHz. Then I see 120Mhz bandwidth as a projector stat. And I see a range for scanning freq. I don't understand what these stats mean? How does it determine the max @## kHz?

friedwart · Last edited by friedwart on Sat Jan 29, 2011 3:29 pm; edited 1 time in total

Ile · Posted: Fri Jan 28, 2011 1:49 pm Post subject:

Bandwidth (MHz) is not limited, picture just start to get softer when you go higher. Rated max BW is just for reference, under that picture should be decently sharp. Usually you can go bit higher without too soft image.

Horizontal scanning freq (kHz) have range where projector will scan properly. Outside this range projector wont show picture.

Same thing with Vertical scanning freq (Hz).

You can calculate Horizontal scanning freq with this.
http://myhometheater.homestead.com/bandwidthcalculator.html
Use only for estimation, because your sources timings (front/back porch, sync width) are probably bit different that calculator use in "with estimated retrace" window. BW values from calculator are much higher compared to straight pixel clock value (above post)...

garyfritz · Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

Just think about how fast the CRT's amplifier has to respond in order to modulate the strength of the electron beam, which is moving across the CRT face. At higher resolutions or higher scan rates, the amplifier has to be able to respond more quickly, and is able to generate higher frequencies, or greater bandwidth.

The formula, in its basic form, for how much bandwith you need is simple: Horizontal pixels x vertical pixels x frames per second, x a compensating factor for signal overhead, providing sync and timing information that isn't displayed in the picture but still takes time to deliver to the sync and timing circuits.

In the case of 1080p-60, it's 1920 x 1080 x 60, or 124,416,000 displayed pixels per second. Add in some compensation for sync and
timing information and you're at a pixel rate of about 140 million pixels per second. If you don't have at LEAST 140 MHz bandwidth,
you can't fully resolve the image. In fact, to FULLY resolve the image with a high MTF value (Modulation Transfer Function, the difference
between the bright section of scan lines and the dark space between them), then ideally you want DOUBLE that bandwidth in the amplifier circuit.

A full bandwidth of 300 MHz through the entire video chain is the Holy Grail of video bandwidth performance, if your top resolution is 1080P-60.

Double it if you need 2x1080P-60 for 3D. But there isn't a CRT projector on the planet that has anywhere close to an honest 600 MHz bandwidth. It just doesn't exist except maybe in a laboratory somewhere. At that rate, the cathode capacitance itself is likely to be
a limiting performance factor.

CJ

ecrabb · Posted: Fri Jan 28, 2011 11:49 pm Post subject:

Enough of this technical talk... Time for a silly metaphor.

OP, think of a water pipe... The water pipe has a cross-section that defines how much water is in one section at any given slice of time. That's like spacial resolution - or 1920x1080 in your example. Bigger resolution, bigger cross section of pipe.

The velocity at which the water is flowing through the pipe is then like refresh rate... It's running at 60hz. How often is that same cross-section of pipe getting "refreshed" with new water... 60hz is 60 times a second.

The total amount of water per second is like bandwidth, then. Bigger resolution, bigger bandwidth. More refreshes, bigger bandwidth.

Wink

SC

garyfritz · Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO

Good analogy, Steve!

Now your projector is a sink that's being filled by that pipe. It has a maximum "bandwidth" it can handle based on the size of the drain. If water comes pouring in faster than the sink can drain, you have a mess. Smile

That's what happens when you feed your projector more bandwidth than it can handle. Except instead of flooding the floor, it just smears the picture.

**Curt Palme** · Posted: Sat Jan 29, 2011 1:22 am Post subject:

dturco · Posted: Sat Jan 29, 2011 1:35 am Post subject:

OK since were talking about bandwidth here, I have a questions. On the DVE test disc the medical test pattern that has the Alphabet on the sides and top and bottom, there are 2 groupings of test lines, 6 boxes each, [pixel frames? one on one off].

Anyway the 6 boxes on the right are fully resolved but the area that should have 6 boxes on the left only has four boxes that are lighter and softer than the other side and the top 2 just aren't there.

Did I run out of bandwidth here? Is that what's going on? Or is that what's supposed to be there?

1080p60hz Blu-ray direct from Sony BDPs550> HDMI> Moome input card.

mistagreg · Joined: 09 Jul 2008 Posts: 47

Thanks for the responses, it helped a lot. I asked because I just got a Vision 1. The specs say a bandwidth of 120, but there are posts of people doing 1080p@72. Which, according to the math puts it over the bandwidth marker. According to the explanation here, this means the picture may be less sharp using that much bandwidth.
-Greg

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

1080p-60 is really quite adequate. I don't notice flicker at 60 Hz and I'd rather stay in bandwidth (or real close to it) and get a sharper image.

I think that those who are running 72 Hz should try going back down to 60 and compare sharpness in complex scenes.

Better yet, use the resolution test patterns. I'm sure that 60 Hz will come out sharper.

However, at 1080p, you have to have your Marquee aligned and tweaked to perfection in order to fully get the best out of the picture,
and even then, it may not be perfect for the simple reason that the smallest convergence steps in a Marquee are more than a pixel in
width at 1080p. This little issue has made me consider modding the convergence system. Reducing the gain on the convergence amps is probably the best approach but it also reduces the total range of convergence adjustment possible. This makes mechanical
setup even more important. But if done right, it would improve the fineness of convergence adjustments.

It's an experiment I want to try.

CJ

Spanky Ham · Joined: 22 Mar 2006 Posts: 5643 Location: Comedy Central

I am not going to look for the link, but there was an excellent discussion on bandwidth about four years ago over at AVS.

garyfritz · Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO

dturco · Posted: Sat Jan 29, 2011 5:24 pm Post subject:

garyfritz · Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO

Those top boxes are the most-detailed and highest-frequency areas of the test pattern, so those are the ones that suffer if you're short on bandwidth. If you're fully resolving the pattern, it looks like your picture. If you're not quite resolving it, you might see lines in both horiz & vert areas, but the horiz pattern will look darker than the the vert. It's easy to do horiz lines -- just turn the gun on (or off) and keep scanning -- but the vert lines require on/off/on/off/on/off really rapidly. If you're bandwidth-limited, instead of getting full-on / full-off in that vert pattern, you end up with a kind of sine-wave-shaped pattern of brightness: mostly on, declining to mostly off, then back up again to mostly on. Because there's no fully-off *black* area in the pattern, it looks brighter than the horiz pattern.

But: if it's due to bandwidth, you should see the same result in all areas of the screen. If you've got good horiz&vert resolution in one area, it's probably not bandwidth. You can get similar effects because your lines aren't sharp due to other causes: poor optical focus (unlikely but possible) due to rings or scheimpflug; poor electrical zone focus; poor zone astig; maybe even poor convergence, though you'd see color fringing with that.

Bottom line: I don't think it's a fundamental limitation of your projector, i.e. bandwidth. I think it's a shortcoming in your setup, and you can probably improve it with careful tweaking.

dturco · Posted: Sat Jan 29, 2011 5:44 pm Post subject:

GaryFritz Wrote: