Resolution on a CRT HDTV

BobY · 11-27-2006, 08:49 PM

I'm scratching my head on this...

Nowhere did *I* say that a CRT-based HDTV would only display around 830 lines of resolution. In fact, if you had a shadow mask of 1920 x 1080 holes and 1920 x 1080 phosphor triads, it would reproduce 1080i perfectly (assuming the pixels were perfectly aligned to the phosphors).

While the reality is the shadow masks and phosphor triads are not that fine, and therefore reduce the visible horizontal resolution (say from 1920 down to 1150, with some of the loss due to overscan), I wasn't commenting in any way on how good or bad the resolution of a CRT is, simply that the beams are always scanning out a fixed array of pixels, regardless of whether the tube reproduces it accurately, and therefore can be considered fixed-pixel displays.

Since the phosphors are not aligned to the pixels being scanned onto them, it is certainly possible for a phosphor to be only partially excited if the pixel changes in the middle of a phosphor triad. This effect is probably more pronounced with a Trinitron (or other aperture grille tube) as they let more of the electron beam through to the phosphor stripes and would explain how there could be more measurable resolution than the number of phosphor triads. In other words, a single phosphor triad can display portions of more than one pixel. Most of the time, though, the misalignment of the pixel transitions to the phosphor dots will somewhat distort how the pixels appear.

Regardless, I'm not down on CRT's at all. I personally have a CRT HDTV because I haven't seen anything that looks better (OK, the Pioneer 50" 1080p Plasma, but it needs to come down in price quite a bit). I would have preferred a Sony, but the Toshiba was the only thing available at the time that would fit in my entertainment center as the speakers were on the bottom.

maicaw · 11-27-2006, 10:49 PM

I should know better - by now!

BobY · 11-27-2006, 11:10 PM

Yep, by now you should know that I'm right.

I asked a straightforward question and you answered it indicating that you continue to completely misunderstand what I'm talking about.

Oh, and if you'd like to actually understand how an HD CRT works, peruse this Genesis chip spec:

http://www.datasheets.org.uk/datashe...rticle=1556798

It's an older design, but the sets we are talking about are old. The block diagram of the chip internals and the system block diagram clearly show how it's done--analog inputs converted to digital, scan rate conversion into a fixed-pixel array in memory, memory output to three D/A converters. How's that for a weird theory? I guess Genesis must be a weird company

1080PsF · 11-29-2006, 03:25 AM

Quote:

Originally Posted by BobY

If there are 1920 pixels in a horizontal line, then each pixel is 16nS wide and that equates to a pixel rate of 62 Mhz.

BTW the duration of a pixel 13.48nsec. Because actual number of pixels in a 1080i signal is 2200 pixels per line and 1125 vertical lines(you have to include horiz and vert blanking). So you have 2200x1125x29.97/1 sec. and that would equal 13.48ns.

BobY · 11-29-2006, 09:48 AM

Yeah, I was just ballparking to make the idea clear, as bringing up blanking might make it more confusing and it didn't affect the point I was making. 1080i uses a pixel clock of 74.25Mhz.

maicaw · 11-29-2006, 01:33 PM

http://pdfserv.maxim-ic.com/en/an/AN2159.pdf

Quote:

...Component Video.
There are two component-video formats; color difference (Y, Pb, and Pr) found in broadcast and luma/chroma (Y-C), or S-VHS, found in VCRs. Both have a luma (Y) bandwidth that's greater than the color (Pb/Pr or C) bandwidth making it difficult to keep the signals time-coincident...

The last anti-aliasing application is HDTV, which is similar to the XGA filter above. The requirements are based on sampling at 74.25MHz, or 148.5MHz, per SMPTE274M or SMPTE296M , and there is a template for the anti-aliasing filter[11]. Like the template for ITU-601-5, this is too difficult for an analog filter, and the only choice left is a compound filter, like that used in the graphics format. At present, 1080i with 74.25MHz sampling is being used by one manufacturer, which requires a 30MHz, 0.1db bandwidth, with <1nsec group delay variation. The roofing filter suggested here is a 5-pole Butterworth using the 1 in a Sallen-Key realization and a single delay-equalization stage, using a single Quad MAX4393. The schematic with the gain and group delay is shown in Figures 5a and 5b. Although this requires more GBW than a Sallen-Key, the response will not rise back up again, when the GBW of the Op-Amp runs out. Other characteristics, such as LeGendre and Chebyshev, will improve the selectivity at the cost of Group Delay....
Reconstruction Filters
The reconstruction filtering after a DAC is one of the most poorly understood of video applications. Unlike anti-aliasing, the requirements aren't clearly specified, and most people think of it as removing the sample clock, but nothing could be further from the truth.
When an ADC samples a signal, it creates multiple, recurring images, centered on the sample-clock's harmonics. The job of the reconstruction filter is to remove all but the baseband component. If the anti-aliasing filter did its job, the output of the DAC looks like image A in Figure 6, and the images to the right are what we want to remove. Simple enough, except for one thing, the DAC samples are impulses and exist for only for an instant in time. To improve this, the DAC "holds" the impulse for a clock period, creating the familiar staircase waveform seen at a DAC output. This "hold" is a digital filter with a sin (x)/(x) characteristic, shown in Figure 7, from whence it got its name, Sinc[12] corrector. Notice at 0.5FS, the Nyquist frequency, the response is down 4db. Overlaying this on Figure 6 shows the insertion loss as a crosshatched area, that includes the signal. The second job of the reconstruction filter is to restore this loss. The good news is the "hold" has a pole of attenuation centered on FS, so you don't have to filter the sample clock, but applications still use the attenuation at FS as a reference...

etc,etc,etc....

BobY · 11-29-2006, 03:33 PM

I agree with all that. I also agreed with your comment on bandwidth, I just didn't understand how it applied to the discussion.

Everything described in the Maxim App. Note. applies to all digitally sampled systems, not just video. Any sampled system must have input anti-aliasing filters to keep all input components below Nyquist and output reconstruction filters to remove output alias images.

One of the reasons that HDMI *should* be superior to Component Video is it can avoid all of the tweaking necessary to compensate for losses associated with the conversion process, along with the phase shifts, group delays, ringing, crosstalk, etc., but I have yet to see an HDMI-driven display that didn't have worse color-banding than the same display driven with Component Video. Most DVD players process the 8-bit color data from the DVD and use 10 or 12-bit DACs for their Component outputs. Although HDMI supports 10 or 12-bit color, I suspect the majority of HDMI outputs (or HDMI inputs) on the market are actually using 8-bits. Maybe when HDMI 1.3 is standard and Deep Color is heavily supported, HDMI will look better than component.

Amazingly (in another example of cutting corners to save cost), I've seen schematics of displays in which the DVI/HDMI input is converted to analog Component video before sending it to the scan rate converter as one of the Component video signals. This allowed them to save money on the pin connections and extra digital circuitry that would be needed to take the DVI/HDMI digital data and store it directly in memory. In these displays, the HDMI input can't possibly look better than Component video.

BMWDavid · 12-03-2006, 06:48 PM

What I want to know...is the picture quality as seen by the eye better on the CRT vs a LCD? I'm looking to get a LG 32" LCD but now I'm giving thought to a Sony 34" XBR (970) if the picture is better.

David

maicaw · 12-03-2006, 08:50 PM

Quote:

Originally Posted by BMWDavid

What I want to know...is the picture quality as seen by the eye better on the CRT vs a LCD? I'm looking to get a LG 32" LCD but now I'm giving thought to a Sony 34" XBR (970) if the picture is better.David

I've been eyeballing the sony 1920x1080 46"LCDs 46/2500 and 46/XBR2 models - they cost from around $3200 (46V25L1 at Costco) to the nicer $3500 -46XBR2 at BB - and weigh about 80# - the same 52" XBR2 model is about $6000 - it's on my wish list -but nobody around here has one on display yet - I'd take a look at one of those Sony 46" 1920x1080 native LCD panels - being fed a good OTA or Cable CBS HDTV signal or HD-DVD ( not PBS or Sat) -for a reference standard - to compare to the 32" LG before you commit -- even if you don't plan on spending that much - you'll then know what you're missing - The 34" CRT should look as good or better than the LG LCD -providing nobody dropped it or turned it upside down - (my first two 34" XBR800s were ruined by the big box primates before I ever got them) - the weight and size could be a problem if you need to move it -they're 200+ pounds front heavy and tip over easily - the CRT leaded glass screen weighs most of that -
on another thread here someone found a refurb? factory direct outlet store -Phillips 30" CRT with QAM and ATSC for about $440 incl shipping - weighs about 100# - a little smaller than 32" - but quite a deal for the features -IMO
how is this?
http://www.outlet.philips.com/b2c_re...og&shop=OUTLET
http://www.outlet.philips.com/b2c_re...og&shop=OUTLET

AFrost20 · 12-14-2006, 07:38 PM

That technical debate was great....but i am too dumb apparently to know if my question was even answered!! HAHA. Anyway, bottom line....can a CRT HDTV display a "true" 1080i 1920x1080 picture???

BobY · 12-14-2006, 08:08 PM

If by that you mean "display all 1920 x 1080 pixels in a 1080i signal", the answer is no.

That doesn't mean it looks bad, though.

AFrost20 · 12-20-2006, 12:55 PM

Thanks. If it cant display all the pixels, how is it still considered
1080i "capable"??

**rbinck** · 12-20-2006, 01:40 PM

Quote:

Originally Posted by AFrost20

Thanks. If it cant display all the pixels, how is it still considered
1080i "capable"??

Because it will accept a 1080i signal and display a picture. Any set that will display at least 720 lines of video is considered HD by the CEA. Actual resolution does not matter to them.

randytried · 12-30-2006, 09:54 AM

Quote:

Originally Posted by BMWDavid

What I want to know...is the picture quality as seen by the eye better on the CRT vs a LCD? I'm looking to get a LG 32" LCD but now I'm giving thought to a Sony 34" XBR (970) if the picture is better.

David

Hello David and All,

I happen to find this link which appears to have the most in depth discussion of displays I've seen. The physicist in charge compares CRT, LCD, plasmas and the LCoS. All in a "shoot out." The issues surrounding resolution is discussed. Perhaps there is something helpful in the discussion.

He certainly appears to show that the CRT has an excellent picture quality and remains king of displays. Of course it is nice to find something like this that confirms what I think i see in my own XBR 970. For I have yet to see anything at the Sony store I like better.

Many of you may already know of this series of articles. Read all four parts. But here is the link.

http://www.displaymate.com/shootout.html#series1

Enjoy the reading. Randy

PenGun · 01-06-2007, 07:58 PM

Well I have a Sony 34XS955 and I can drive it all kinds of ways because everything goes through my computer.

It's a pure joy. Everyone who has seen it says it's the best picture they have ever seen.

It's a wash for image quality with DVI and Component. I have ended up driving it at 720p with DVI for games ... yup Doom3 is a lot of fun at that size

and 1080i Component for video.

Only real difference is that Component shakes a tiny bit on computer graphics at 1080i and DVI does not. I have convinced myself the colour space is slightly nicer with Component but I'm not really sure.

I sit between 30" and 40" from this and at 30" I can just see lines at 720p but not at 1080i. At 40" it's superb.

I cannot recommend this too highly. I imported mine from NY NY to BC Canada 'cause the only one I could find was the one the owner of the local Sony store has and $2000 cash did not move him at all. That's about what I ended spending anyway.

11-27-2006, 10:49 PM	#17
maicaw Crabtree's Bludgeon Join Date: Apr 2005 Posts: 2,001	Wtf! I should know better - by now! __________________ --------------signature ------- My HDTV equipment is certified organic - mostly plastic No genetically modified electrons or photons are used in it's operation RocketMan -- Tambourine Man -- Lucy {userName=1},{P/w=1} Last edited by maicaw; 11-27-2006 at 10:52 PM.

Similar Threads to Resolution on a CRT HDTV
Thread	Thread Starter	Forum	Replies	Last Post
Which 30 INCH CRT HDTV HAS THE MOST Horozontal Lines Of Resolution?	rickbbb	Direct-View (tube) TVs	32	12-04-2008 03:07 PM
Need help with blurry text on my CRT HDTV	darien95	Computers/HTPCs	7	11-18-2008 07:30 PM
30" Widescreen CRT HDTV $269.00 shipped free	stan123abc	Direct-View (tube) TVs	43	08-14-2007 08:58 AM
Watching SD on HDTV - Why not just watch on a SDTV?	bboyneko	The High Definition Lounge	8	04-20-2007 10:22 AM
CRT HDTV's incapable of progressive scan?!?	EddieDZ	Direct-View (tube) TVs	11	03-28-2006 12:20 PM
Need major helping picking a CRT and resolution issues..	EddieDZ	Direct-View (tube) TVs	22	02-04-2006 04:54 PM

11-27-2006, 08:49 PM	#16
BobY What's all this, then?... Join Date: Jan 2006 Posts: 6,197	I'm scratching my head on this... Nowhere did I say that a CRT-based HDTV would only display around 830 lines of resolution. In fact, if you had a shadow mask of 1920 x 1080 holes and 1920 x 1080 phosphor triads, it would reproduce 1080i perfectly (assuming the pixels were perfectly aligned to the phosphors). While the reality is the shadow masks and phosphor triads are not that fine, and therefore reduce the visible horizontal resolution (say from 1920 down to 1150, with some of the loss due to overscan), I wasn't commenting in any way on how good or bad the resolution of a CRT is, simply that the beams are always scanning out a fixed array of pixels, regardless of whether the tube reproduces it accurately, and therefore can be considered fixed-pixel displays. Since the phosphors are not aligned to the pixels being scanned onto them, it is certainly possible for a phosphor to be only partially excited if the pixel changes in the middle of a phosphor triad. This effect is probably more pronounced with a Trinitron (or other aperture grille tube) as they let more of the electron beam through to the phosphor stripes and would explain how there could be more measurable resolution than the number of phosphor triads. In other words, a single phosphor triad can display portions of more than one pixel. Most of the time, though, the misalignment of the pixel transitions to the phosphor dots will somewhat distort how the pixels appear. Regardless, I'm not down on CRT's at all. I personally have a CRT HDTV because I haven't seen anything that looks better (OK, the Pioneer 50" 1080p Plasma, but it needs to come down in price quite a bit). I would have preferred a Sony, but the Toshiba was the only thing available at the time that would fit in my entertainment center as the speakers were on the bottom.

11-27-2006, 11:10 PM	#18
BobY What's all this, then?... Join Date: Jan 2006 Posts: 6,197	Yep, by now you should know that I'm right. I asked a straightforward question and you answered it indicating that you continue to completely misunderstand what I'm talking about. Oh, and if you'd like to actually understand how an HD CRT works, peruse this Genesis chip spec: http://www.datasheets.org.uk/datashe...rticle=1556798 It's an older design, but the sets we are talking about are old. The block diagram of the chip internals and the system block diagram clearly show how it's done--analog inputs converted to digital, scan rate conversion into a fixed-pixel array in memory, memory output to three D/A converters. How's that for a weird theory? I guess Genesis must be a weird company

11-29-2006, 09:48 AM	#20
BobY What's all this, then?... Join Date: Jan 2006 Posts: 6,197	Yeah, I was just ballparking to make the idea clear, as bringing up blanking might make it more confusing and it didn't affect the point I was making. 1080i uses a pixel clock of 74.25Mhz.

11-29-2006, 03:33 PM	#22
BobY What's all this, then?... Join Date: Jan 2006 Posts: 6,197	I agree with all that. I also agreed with your comment on bandwidth, I just didn't understand how it applied to the discussion. Everything described in the Maxim App. Note. applies to all digitally sampled systems, not just video. Any sampled system must have input anti-aliasing filters to keep all input components below Nyquist and output reconstruction filters to remove output alias images. One of the reasons that HDMI should be superior to Component Video is it can avoid all of the tweaking necessary to compensate for losses associated with the conversion process, along with the phase shifts, group delays, ringing, crosstalk, etc., but I have yet to see an HDMI-driven display that didn't have worse color-banding than the same display driven with Component Video. Most DVD players process the 8-bit color data from the DVD and use 10 or 12-bit DACs for their Component outputs. Although HDMI supports 10 or 12-bit color, I suspect the majority of HDMI outputs (or HDMI inputs) on the market are actually using 8-bits. Maybe when HDMI 1.3 is standard and Deep Color is heavily supported, HDMI will look better than component. Amazingly (in another example of cutting corners to save cost), I've seen schematics of displays in which the DVI/HDMI input is converted to analog Component video before sending it to the scan rate converter as one of the Component video signals. This allowed them to save money on the pin connections and extra digital circuitry that would be needed to take the DVI/HDMI digital data and store it directly in memory. In these displays, the HDMI input can't possibly look better than Component video.

12-03-2006, 06:48 PM	#23
BMWDavid How can anyone watch standard def? Join Date: Dec 2006 Posts: 16	What I want to know...is the picture quality as seen by the eye better on the CRT vs a LCD? I'm looking to get a LG 32" LCD but now I'm giving thought to a Sony 34" XBR (970) if the picture is better. David

12-14-2006, 07:38 PM	#25
AFrost20 Sony! Join Date: Apr 2006 Posts: 60	That technical debate was great....but i am too dumb apparently to know if my question was even answered!! HAHA. Anyway, bottom line....can a CRT HDTV display a "true" 1080i 1920x1080 picture???

12-14-2006, 08:08 PM	#26
BobY What's all this, then?... Join Date: Jan 2006 Posts: 6,197	If by that you mean "display all 1920 x 1080 pixels in a 1080i signal", the answer is no. That doesn't mean it looks bad, though.

12-20-2006, 12:55 PM	#27
AFrost20 Sony! Join Date: Apr 2006 Posts: 60	Thanks. If it cant display all the pixels, how is it still considered 1080i "capable"??

01-06-2007, 07:58 PM	#30
PenGun Avast Mateys Join Date: Jan 2007 Location: Backwoods BC Canada Age: 63 Posts: 217	Well I have a Sony 34XS955 and I can drive it all kinds of ways because everything goes through my computer. It's a pure joy. Everyone who has seen it says it's the best picture they have ever seen. It's a wash for image quality with DVI and Component. I have ended up driving it at 720p with DVI for games ... yup Doom3 is a lot of fun at that size and 1080i Component for video. Only real difference is that Component shakes a tiny bit on computer graphics at 1080i and DVI does not. I have convinced myself the colour space is slightly nicer with Component but I'm not really sure. I sit between 30" and 40" from this and at 30" I can just see lines at 720p but not at 1080i. At 40" it's superb. I cannot recommend this too highly. I imported mine from NY NY to BC Canada 'cause the only one I could find was the one the owner of the local Sony store has and $2000 cash did not move him at all. That's about what I ended spending anyway.