[CMRA]

Just how important is RBG bandwidth?


Here's what a major CRT OEM has to say:


"RGB bandwidth determines how fast the electron beam can change state as it sweeps across the screen, setting an upper limit on horizontal resolution. The most torturous image for a CRT is an alternating series of one-pixel-wide vertical B&W lines. The input signal for such a pattern is a square wave, but unfortunately, bandwidth measures only sinusoidal frequency components. The sinusoidal bandwidth required to adequately approximate a square wave is roughly three times the binary frequency of the square wave. The additional bandwidth is required to capture the third harmonic, which is essential if a decent approximation of a square wave is to be obtained (most of the signal energy for a square wave is in the first and third harmonics of its Fourier series)."


What was that?


So just how important is bandwidth? Will a 100Mhz bandwidth be visually better than say 75Mhz (which specs exceed HDTV requirements)?

If so, will 120,135,150, and yes 180(Cine 9) show any visual improvement over each other (all things being equal)?

 

[MikeReilly]

I think the point is that you need to have sufficient bandwidth to capture the third harmonic in order to have the best PQ. Anything beyond that is probably overkill. Being able to display an image is different from being able to show it well (ie: third harmonic). It really depends on what resolution you want to display.

 

[CMRA]

techman, secstate, etc. I know you guys have some answers.

 

[RoBro]

Don't forget that a square wave has a 5th harmpnic, 7th harmonic.....


On graphic material (computers...) this may be true, but most video material is already filtered very much until it reaches your PJ, so it will not be that dramatically...

Roland

 

[CMRA]

Quote:

Originally posted by MikeReilly It really depends on what resolution you want to display.

1080p should suffice. Answers?

 

[nidi]

so , 1080P 60Hz needs 186.6 MHz bandwidth, more than even the

Cine 9 can display, have a 3-3 pulldown (72 Hz) and your at

223.9 MHz. third harmonics ??


cheers


Michael


cow and chocolate land

 

[josh-gustafson]

Hmm... The question this raises in my mind:


I'd imagine most of us use our CRTs primarily for film viewing. For films, the pixels in the image represent point samples from whatever's being filmed; not the actual thing itself. In other words, the square-wave nature of pixels is an artifact of the pixels themselves, and is *not* part of the nature of the original image. If that third harmonic (and higher harmonics) is not part of the original image, wouldn't it actually be *better* *not* to reproduce it on the screen, for a more film-like experience?


(It's a different story, of course, if we're talking about computer text and UIs designed with square pixels of internally uniform color in mind, but I imagine that's of secondary concern to most of us...)


Of course, I'm speaking from a theoretical standpoint here. I haven't gotten my new 8500 fixed up yet, and prior to that I've only had an old Panasonic PT-120...

 

[CMRA]

Here's another source of wisdom.


"Scan Rates and HDTV - The ability of a projector to display HDTV depends on its scanning capability. For example for a projector to line double standard NTSC signals (480i at 15.75khz horizontal) from DVD, VCR or cable/ Satellite to 480p, requires 31.5 Khz of horizontal scan rate and enough bandwidth to handle twice the picture information. HDTV at 1080i requires a horizontal rate of 33.75 Khz and 50 Mhz bandwidth. 720p requires a horizontal rate of 45 Khz with 50 Mhz bandwidth. "


It implies a 1080i will need to be doubled, hence 67.5 khz horizontal scan rate and 100Mhz bandwidth. These specs seem within the range of many EMF 8 and 9 inchers out there. Do the facts challenge the theory?

 

[JVanderwalker]

A 1080i camera has a pixel clock of 74 MHz which means the highest frequency that can be output from the camera is 37 MHz. Bandwidth is specified in terms of what frequency is the response down by say 1 dB or 3 dB. So to pass 37 MHz it would be prudent to have 50 MHz frequency response to make sure the response is flat to 37 MHz.

Jim

 

[RoBro]

Ok, 74MHz pixel clock makes a maximum of 37MHz of signal frequency, but that is square wave with all the 3rd, 5th, 7th.... harmonics...

But skipping the 5th and above should not degrade the pic, so you should be good with a 3dB bandwidth of something above 111MHz, say something like 140MHz.

On 1080p you would need the double bandwidth, leaving you with about 280MHz -3dB.

I am glad to live in PAL country, where we can go with 20% less bandwidth

Roland

P.S.:

The beam spot size makes up for an additional low pass filter, so you may not really see a difference between a RGB bandwidth of 50 and 100MHz at 1080i...

 

[techman707]

Quote:

Originally posted by CMRA techman, secstate, etc. I know you guys have some answers.

Your question isn't clear, but if it's what I think you're asking. You may be able to handle the scan rate for 1080p on your projector and you may even get a subjectively decent picture. However, if the question is whether you are displaying the FULL capability of 1080p without meeting the bandwidth requirement, the answer is NO.

 

[josh-gustafson]

Quote:

Originally posted by RoBro Ok, 74MHz pixel clock makes a maximum of 37MHz of signal frequency, but that is square wave with all the 3rd, 5th, 7th.... harmonics...

But again, isn't the sqaure wave just a result of treating pixels as square regions of uniform color? If your image was sampled (as opposed to something like PC text) then what it was you were sampling is most likely *not* a square wave, and the harmonics are just an artifact anyway. Worrying about properly reproducing sampling artifacts seems silly to me.


In fact, the better DACs used in audio, as I understand it, actively work to ensure that they do NOT "reproduce" those harmonics. Of course audio isn't quite the same as video, but...

 

[deronmoped]

What if we made a list of importance of what maters when it comes to displaying a image from a CRT PJ?


1. Size of the CRT's

2. EM focus.

3. Color filtering.

4. Liquid coupling.

5. No streaking.

6. On off black (some CRTs are better at this).

7. ANSI contrast (some CRTs are better at this).

8. Scan rate.

9. Holding convergence.


Can anyone add to the list? I just thinking of things that I have heard were problems before.


Has anyone ever said that they do not have enough bandwidth?


Deron.

 

[techman707]

Quote:

Originally posted by deronmoped What if we made a list of importance of what maters when it comes to displaying a image from a CRT PJ? 1. Size of the CRT's 2. EM focus. 3. Color filtering. 4. Liquid coupling. 5. No streaking. 6. On off black (some CRTs are better at this). 7. ANSI contrast (some CRTs are better at this). 8. Scan rate. 9. Holding convergence. Can anyone add to the list? I just thinking of things that I have heard were problems before. Has anyone ever said that they do not have enough bandwidth? Deron.

While everything you mention is true, with respect to good overall picture quality, none of those things addresses the issue of people that want to use 1080p for HDTV. As I have said, you may get a decent picture (if all the other things you mention are good), but, you STILL won't reproduce "all" the detail that would be in a "true" 1080p picture.


Fortunately (or unfortunately), they don't transmit 1080p for HDTV, so it really make no difference, but if they did, ALL the 9" projectors that I am aware of would fall short of reproducing the full bandwidth.

 

[RoBro]

Even if they transmit 1080i and jou linedouble it your PJ can not resolve it, so it will be softer than 1080i...

Roland

 

[techman707]

Quote:

Originally posted by RoBro Even if they transmit 1080i and jou linedouble it your PJ can not resolve it, so it will be softer than 1080i... Roland

Right AGAIN!

 

[JVanderwalker]

Quote:

Originally posted by RoBro Ok, 74MHz pixel clock makes a maximum of 37MHz of signal frequency, but that is square wave with all the 3rd, 5th, 7th.... harmonics... But skipping the 5th and above should not degrade the pic, so you should be good with a 3dB bandwidth of something above 111MHz, say something like 140MHz. On 1080p you would need the double bandwidth, leaving you with about 280MHz -3dB. I am glad to live in PAL country, where we can go with 20% less bandwidth Roland P.S.: The beam spot size makes up for an additional low pass filter, so you may not really see a difference between a RGB bandwidth of 50 and 100MHz at 1080i...

With a lens on a camera including diffraction limiting and the real world MTF (modulation transfer function) real world scenes from the 1080i output of a camera looks like a sine wave. There is no way you can get a 37 MHz square wave. I maintain that 50 MHz for 1080i and 100MHz for 1080p is good enough.

Jim

 

[techman707]

Quote:

Originally posted by JVanderwalker With a lens on a camera including diffraction limiting and the real world MTF (modulation transfer function) real world scenes from the 1080i output of a camera looks like a sine wave. There is no way you can get a 37 MHz square wave. I maintain that 50 MHz for 1080i and 100MHz for 1080p is good enough. Jim

If "good enough" is ok then go for it.

 

[DAP]

Every device you pass the video signal through acts as a filter. Filter bandwidths are measured at their 3 dB point (unless otherwise specified, the 3dB point is how filters are generally defined). The 3 dB point is the half power point, or the point at which the signal strength drops to half power relative to the middle of the pass band of the filter. Since we are talking here about 75 ohm impeadence systems where the load is resistive (not reactive) the half power point is also half voltage.


Now what happens when we stick two 50 MHz bandwidth filters one after another in the circuit? At the output of the 1st filter, a 50 MHz signal would be half the power of a 10 MHz signal if both were the same power on the input. After the second filter, the 50 MHz signal would be a quarter of the power of the 10 MHz signal. Two 50 MHz bandwidth systems in series produce a system with significantly less bandwidth.


That being said, One DOES want filtering in the system, assuming we are talking about a digital system. A the image that gets to the image sensor should not have a higher resolution than the image sensor. If it does, you will get aliasing (Look up nyquist sampling, you get the same problems in video as you would get in audio) On the output, you also want filtering to eliminate the hard pixel edges.


But that is all. You don't want any more filtering than that, otherwise you loose picture detail. A high bandwidth video system won't display any more information than you give it, but then it also won't throw any of it away.

 

[techman707]

Quote:

Originally posted by DAP A high bandwidth video system won't display any more information than you give it, but then it also won't throw any of it away.

"but then it also won't throw any of it away." That is exactly what you would be doing "IF" they were actually transmitting a 1080p signal. The reason people running 1080p and believe it looks fine is because they really aren't dealing with a 1080p signal, if they were, they would see the difference, if they had a display that WAS capable of displaying it properly. That's why in many instances, it's a waste to run 1080p, especially with sources like DVD.


Even Extron knew back at the beginning of HD projectors that they needed to make interfaces that passed bandwidths of 250-350MHz, even though the projector's they were going to be used with had MUCH lower bandwidths.