AVS Forum banner
Status
Not open for further replies.
1 - 20 of 59 Posts

·
Registered
Joined
·
94 Posts
Discussion Starter · #1 ·
Is there any relatively reliable source for a comparison of horizontal resolution between various models of HD RPTVs or even direct view TVs?


Also is there anyplace where someone has done side by side comparisons of HD on properly calibrated DTVs.


It is impossible to try to compare HD picture quality and resolution of various HD sets. Few stores even have HD on more than one(if even one!) display, calibration issues aside.


It would be nice to see some standardized measurement of horizontal resolution applied to all sets, and then published! Is this possible?


At least than one wouldn't have to rely so much on owner testimonials and (possibly biased) reviews.


E
 

·
AVS Forum Special Member
Joined
·
11,139 Posts
I've been looking for such sources for some time, too. A while back here someone said an HDTV station at his location regularly put out a test pattern that included resolution images, just as NTSC stations sometimes broadcast.


The problem, I concluded, was reasonably priced signal generators that could put out an adequate signal. A new HDTV-oriented model was just introduced and reviewed here: http://www.avsforum.com/ubb/Forum11/HTML/007925.html If about $2000 for a generator seems reasonably, there's now a source. PCs, usually with add-on specialized software, can also be set up to generate HDTV-level test patterns. The next low-priced source, unless another signal-generator firm surprises us, might be HDTV-capable optical-disc players. Once they're on the market, an Avia-like test disc is sure to follow.


--John




[This message has been edited by John Mason (edited 12-05-2000).]
 

·
Registered
Joined
·
141 Posts
Resolution these days is almost as mis-used as Music Power was in the early hi-fi days.


The most unbiased measure of resolution for TV images is "TV lines per picture height," abbreviated TVL/PH. It is the total number of lines (black AND white) that can be seen on a display NORMALIZED to the picture height.


For vertical resolution, it is simply the number of black and white lines that can fit from the top of the screen to the bottom. For horizontal resolution, take the number of black and white lines that can fit across the screen, turn it 90 degrees and see how many can fit in the distance from the top of the picture to the bottom. (That's the NORMALIZED, or "per picture height" part.)


This way you get one measure of lines that means the same for H and V resolutions, allowing it to be applied to 4x3, 16x9, 1.85:1, 2.35:1, or whatever.


Some folks talk of lines per degree of arc, but that adds viewing distance and visual acuity into the equation (muddying the measurement, in my view.)


TVL/PH! It's the measure that makes the most sense with varying picture sizes and shapes. Try it and see if you agree! (And now, back to our forum...)


Ken
 

·
Registered
Joined
·
94 Posts
Discussion Starter · #4 ·
Now if someone would just make available TVL/PH comparisons for all of the major HDTV sets on the Market, we would be armed with one more criteria for comparing and establishing which properly calibrated sets have the best PQ.


My question now, how does TVL/PH relate to overall visual clarity, and how are CRT size, lenticular pitch, beam focusing, lense quality, etc accounted for in this measurement?


E
 

·
Registered
Joined
·
50 Posts
I have the information you are asking for...but I have to vent, first.


1080i x 1920 = 2,073,600 pixels. A frequency of 30 megahertz.


This is the "Broadcast" standard for HDTV. I believe CE manufacturers are "challenged" to produce HDTV's that can come as close to these numbers & still price the sets so we will / can purchase them.


Some manufacturers make sets that will DO these numbers and have a price tag to prove it. Some cut costs SO dramatically on their sets that it is amazing that these sets can even produce a picture in the HD mode.


August 31st, 1999 the CEA released a "new" set of HDTV definitions to help further "define" what the parameters of an HDTV picture should be. I found it interesting that the CEA only found it necessary to define "vertical resolution" and not horizontal.


Upon investigation, there really is NO STANDARD for measuring horizontal resolution. I guess it must NOT be important then, right? After all, horizontal resolution is where your detail comes from. More horizontal resolution more detail, right? In theory, 1080i x 1000 shouldn't be nearly as detailed as 1080i x 1600 or more. Yet, there are HDTV's out there that ONLY perform to 1000 lines or less. So technically to classify a Television as an HDTV according to the definitions set by the CEA, the set could have a measurement of 1080i x 1 (one) line of Horizontal Resolution & be an HDTV...ridiculous, but true.


IMHO the sad part is that the "normal" consumer, my dad, for example is completely blown away by a 480p picture. No scan lines!!! "...that picture is incredible" he says. When I ask the retailer to show us an HDTV picture...they can't. We are not wired for that...We don't have an HDTV source...These TV's aren't hooked up for that. Yet these retailers SELL (term used loosely) the product & expect the customer to buy these products. I believe that some CE manufacturers realize this & build their sets to meet this 480p "wow" factor. Also,if you look at the numbers below, notice that some of the Japanese manufacturers still seem to build their sets based on the old "MUSE" standard or "Analog" HDTV.


Just a thought, but I have read where people are absolutely AMAZED at the picture they get when they hook up a Progressive Scan DVD such as a Toshiba SD5109/SD6200 to their HDTV's??? To me, that would immediately send up a flag saying "if the manufacturer cut costs on the Line Doubler, what else did they cut costs on?" IMHO, there SHOULD be a difference from an interlaced 480i picture to a Progressive Scan 480p DVD picture, but it should not be like opening your eyes for the first time.


For those of you that have the HDTV's listed below, you bought them for a reason. If you like what you see in your living rooms, home theater rooms...etc. then "great". This post will not change the picture you see in front of you. It may possibly make you research your "next purchase" a little more carefully, though.


Honestly, when I received this information & reviewed it, I was quite "surprised". What I thought was "top-of-the-line", really wasn't.


So how do you measure horizontal lines of resolution? Well according to my source, the CEA recognizes the following method as an honest, realiable & TRUE measure of Horizontal Resolution. However, it is NOT a standard. Remember, there is NO standard for Horizontal Resolution.


The formula for measuring Lines of Horizontal Resolution is: Lines of Picture Height x Aspect Ratio of displayed picture 1.78 (16 x 9) or 1.33 (4 x 3). This is the area "between the black bars" in the HD mode. It could also be between the "Gray Bars" on other sets.


O.K. You asked for it. The source for this information is reliable (I will NOT reveal it, so don't ask). It was done by a 3rd party non-biased testing company. These sets were tested at retail. ALL the sets were converged before testing. The testing equipment was an HD picture generator called, Shiba Soku. This HD generator produces pretty HD pictures & more importantly, what "the testers" called an HDTV "eye chart". This chart shows: Lines of picture height, Lines of vertical resolution, gray scale, interlacing, diagonal detail...The most important chart for this post is "Lines of Picture Height". Enjoy....


Model Format Lines Pic Hght Hor Res HD wideband amp

Hit 61SDX01 4x3 400 1.78x400=712 10-11 megahertz

Mit WS55807 16x9 650 1.78x650=1157 18-19 megahertz

Mit WS55857 16x9 650 1.78x650=1157 18-19 megahertz

Mit WS55907 16x9 700 1.78x700=1246 20 megahertz

Pan PT51HX40 4x3 600 1.33x600=798 12-13 megahertz

Phi 60PP9601 4x3 900 1.78x900=1602 27-28 megahertz

Phi 55PP9701 16x9 900 1.78x900=1602 27-28 megahertz

Phi 60PP9701 16x9 900 1.78x900=1602 27-28 megahertz

Pio SD53HD2 16x9 700 1.78x700=1246 20 megahertz

RCA MM52100 4x3 550 1.33x550=731 10-11 megahertz

Sam PCK5315 4x3 500 1.78x500=890 10-11 megahertz

Son KP61HS10 4x3 600 1.78x600=1068 16-17 megahertz

Son KP57XBR104x3 575 1.78x575=1023 16 megahertz


NOTE: According to my source, these tests were done more than once on these models. The results are "averaged" but I have been told to factor in a margin of error for approximately 8-10%.


Hope this answers your questions...
 

·
Registered
Joined
·
94 Posts
Discussion Starter · #8 ·
LB



Thanks for this info!



Finally some objective horz res comparisons. It is sad that the CEA is allowing marginal equipment to be marketed as top-of-line by using this horizontal resolution loophole (as well as ignorant retailers). Even informed consumers who do there homework will have a hard time getting there hands on info like this.



E



[This message has been edited by EBeast (edited 12-05-2000).]
 

·
Registered
Joined
·
887 Posts
Ebeast, Your question is a good one and LB covers many items.


The missing one, however, is HOW the resolution is "measured" as opposed to defined. The issues you list all affect the obtainable resolution and it's only the visible resolution we're interested in evaluating (it doesn't help if the electronics can handle 2000 lines if the tubes can't display it) The standard that's missing is how do you define whether lines are resolved or not : a light gray line next to a slightly darker gray line might just be perceptibly different, but another display might yield true black next to true white (for the same input signal) and would seem sharper.


The difference between adjacent lines needs to be specified in a "standardized" test so we're all talking about the same thing - it doesn't really matter whether this difference is 20% 50% or 70%, as long as the same criterion is used whenevr the test result is reported. I'm under the impression that ANSI has covered this for display devices (using 30% depth of modulation?) "ANSI lumens" seem to be cited often in terms of brightness for displays, so manfacturers and reviewers could dig out the ANSI lines test and start applying it. If it's like the ANSI lumens test, it will address performance over the whole image, corners and all, not just the best bits


It also strikes me that the set should be calibrated and "standardized" e.g. a la ISF and the resolution measured under those settings (correct b & w levels, grayscale etc) so the test result is the resolution that would be relevant to actually watching the unit (not what it can get when tweaked specially for the test, but not watchable otherwise).


I assume these tests were consistent for the sets reported on and that the conditions can be described well enough so others can be added to the list (by other testers) and still be comparable


Andy
 

·
Registered
Joined
·
1,456 Posts
This information is generally not available mainly due to the fact that test equipment that can produce the required HDTV signals has until recently been too expensive for the average magazine reviewer.


By putting a VE or Avia disk into a good DVD player, even a home user can have the equivalent of what used to be an expensive SDTV test pattern generator. Unfortunately, the same thing can't be done for HDTV, at least not until HD-DVD recomes a reality.


I noticed Sencore has recently introduced an HDTV test generator. Their equipment is priced a test-technician levels, rather than the sky-high prices hat laboratory-grade measurement equipment goes for. So I think you'll probably start seeing more published HTDV measurements in the future as the magazine guys get their hands on this, and similar gear.


--Barry


[This message has been edited by BarryO (edited 12-05-2000).]
 

·
Registered
Joined
·
681 Posts
Quote:
Originally posted by LB:
I have the information you are asking for...but I have to vent, first.


1080i x 1920 = 2,073,600 pixels. A frequency of 30 megahertz.
You need frequency response out to 37MHz to resolve 1920 pixels on a scan line at 1080i. 30MHz of bandwidth is only enough to resolve 1550 pixels. Do the math.


-Jonathan
 

·
Registered
Joined
·
98 Posts
What seems to get lost in all the advertising and hype is that current HDTV resolution is really 720 x 1280, which is just fine for now. Why 720? Because 1080i has the same vertical resolution as 720p due to the vertical filtering that must be done to prevent flicker and line twitter in an interlaced signal. Filtering reduces vertical resolution by about 1/3 or to about 700 or so lines. This is the heart of the argument advanced (correctly) by Joe Kane and others that 720p is as good or better than 1080i.


The horizontal resolution that corresponds to a vertical resolution of 720 lines is 1.78 x 720 = 1280 lines. The RPTVs that deliver about 1200 lines horizontally are delivering true HDTV, albeit at the lower of the two HDTV resolutions. The sets, mostly 4:3, that deliver substantially less cannot be considered to be true HDTV, despite what the CEA says.


Full 1080 x 1920 resolution will have to wait for 1080p to become practical and affordable, which is not going to be anytime soon. That should not be a big concern for now as HDTV at the 720 x 1280 (1080i or 720p) resolution looks really good, is vastly superior to 480i, and can be enjoyed now. We can look forward in the future to seeing further improvements with 1080p.
 

·
Registered
Joined
·
195 Posts
LB,


I know this has been asked before, but are you sure you don't work for Philips? http://www.avsforum.com/ubb/smile.gif


As far as the Phi 60PP9601, I think you have a typo, because its figures are identical to the 4:3 version. I'm interested in what the tests actually said about it, for TVL/PH and horizontal resolution. Do you have that data?


------------------

- Jeff
 

·
Registered
Joined
·
195 Posts
Oops, I meant "identical to the 16:9 version"


------------------

- Jeff
 

·
Registered
Joined
·
1,084 Posts
Quote:
Originally posted by HD Ken:
Some folks talk of lines per degree of arc, but that adds viewing distance and visual acuity into the equation (muddying the measurement, in my view.)
Your lucid and informative post perhaps oversteps a salient point a bit. There may well be gold in that thar mud, as the 49ers found out at Sutter's Creek. With 36" displays, and a viewing distance of ~10', it's doubtful that the apparent resolution differences between two displays - one that measured 800 horizontal lines of resolution and another at 675 lines would be visible to someone corrected to 20/15 acuity. With a 72" display, the difference would probably be apparent given the same parameters. I don't mean to denigrate the importance of striving for best resolution possible, but, let's also not forget that the "big picture" (metaphorically as well as measurably) also has to be considered, or we're guilty of obsessing for its own sake.

 

·
AVS Forum Special Member
Joined
·
11,139 Posts
Assuming test equipment and HDTV test signals become more economically available, as I and BarryO mentioned above, perhaps some tests approximating what the ATSC folks set up might be useful. Their tests are described at: http://www.atsc.org/papers/tsreport.html

Section 2.1 and table 2.3 refer to both expected and measured resolution values. The table values can be confusing unless you remember to DOUBLE everything shown. That's because their data is given in cycles per active picture height, and I at least am used to thinking in terms of alternating lines on a screen (not cycles). From these figures, and taking both vertical and horizontal filtering by receiver circuits, plus the Kell factor (the difference between static and moving images) into account, it's surprising how far from 1080 X 1920 pixels, sometimes referred to as the highest potential HDTV resolution, is from what the ATSC expected and measured.


--John


[This message has been edited by John Mason (edited 12-06-2000).]
 

·
Registered
Joined
·
98 Posts
Why are we still talking about 1080 x 1920 resolution? Current sets and broadcasts are not aiming at that resolution, and no 1080i system can possibly deliver that resolution, as I noted above. The ATSC paper concedes that an interlaced system can only be expected to deliver vertical resolution of 65% of the number of scan lines, which is 702 lines of vertical resolution for a 1080i system.

Quote:
Therefore, the Scanning Formats / Compression Expert Group, in order to establish target specifications for resolution, assumed that an interlaced scanning system should deliver limiting vertical resolution equal to 65 % of the number of active scanning lines, and that a progressive scan system should deliver 90 %. The Expert Group assumed that horizontal resolution would be limited to about 80 % of the number of active samples because of

filtering...
Current HDTV is 720 x 1280, whether it's 1080i or 720p, and LB's numbers show that many current sets deliver resolution very close to that figure. Rather than lament how far current sets are from the highest resolution standard (which they are not even trying to meet), we should note how close they are to the 720 x 1280 HDTV resolution they are aiming for.
 

·
Registered
Joined
·
887 Posts
Yeringto,

Thanks. The report "from the horse's mouth" and your post quantitatively clarify the vertical resolution issue and the greater (non-horizontal) motion sensitivity of the interlaced format, as intuitively expected.


In your quote,"" they assumed 80% limit for the horiz resolution due to receiver filtering."" Thus 1920 could be assumed to be reduced to 1540. In Table 2.3 they show measured values of 460 and 320 ( for 1080i and 720p respectively) horiz. As you say we need to double these "cycles" to get to "lines" if a black and white line pair counts as 2 "lines" ; then we need to multply by 1.78 to go from "per picture height" to "total" lines.


These come out to 1637 and 1139 for 1080i and 720p.


So I'm confused where you justify saying the ""Current HDTV is 720 x 1280, whether it's 1080i or 720p""


I'm OK with the apparent vertical resolution of 1080i and 720p being close to 720 for both formats but you lost me on the horizontal.... 1600± is quite a bit better than 1280, right?


This , of course, is a separate question from the source resolution , encoding compression loss, receiver "quality" and display capability etc all of which have potential effects on "real" pictures we see with our equipment


Andy
 

·
Registered
Joined
·
4,644 Posts
Current HDTV includes 1080p. It is part of the ATSC standard at frame rates of 24Hz and 30Hz (but not 60Hz). It is progressive, so it does not require vertical filtering like 1080i.


I think it is unfortunate that this format is neglected. Material that could potentially be transmitted as 1080p is usually transmitted as 1080i. Even if 1080p were transmitted, it is typically converted to 1080i by the receiver and displayed as 1080i. I wish there was more support for 1080p.
 

·
Registered
Joined
·
98 Posts
AJSJones,

Yes, you are right that 1080i can have a greater horizontal resolution than 1280. I say that current HDTV is 720 x 1280 for two reasons. Given the limit of 700 or so vertical resolution, the matching horizontal resolution for a system with equal horizontal and vertical resolution per picture height would be 1280. This is the "natural" horizontal resolution that goes with a 720 vertical resolution. Most real world, affordable HDTV displays deliver a maximum of about 1200 lines horizontal resolution, so in practice, too, we are dealing with 720 (real or equivalent) x 1280 HD resolution.


There is nothing wrong with providing 1400 or 1600 lines of horizontal resolution, and it is certainly desirable, but it is not really necessary for a 1080i display that has only 700 lines of real vertical resolution - just extra icing on the cake.


Erik,

Having 1080p would be great, but it requires much more bandwidth and storage space. The equipment required to generate, store, transmit, and display it is just too expensive at the moment. Just think of it as a future upgrade path. There are enough problems just trying to make current 1080i/720p broadcasts viable.




[This message has been edited by Yeringto (edited 12-07-2000).]
 
1 - 20 of 59 Posts
Status
Not open for further replies.
Top