Hello.

I have an old Panasonic RPTV "1080i HDTV". I put that in quotes because those early units did not really have the capability of displaying the full resolution of 1080i.

I was wondering if there is a small sample video or software image I can download somewhere on the web than I can then use to measure the real horizontal resolution of my TV. I could connect my laptop through the component dongle and set it to 1080i and use this video or software to make my measurements.

Thanks.

http://www.avsforum.com/avs-vb/showthread.php?t=570937
--John

Thank you very much.
This looks very useful.
I'll give it a shot when I get home.

The problem is your measurement will only be as good as the output of your source test equipment. Since the Tv probably only has analog component inputs, you would need a high quality/bandwidth analog source device. I've read here that most consumer cable/sat boxes really only output 1300 lines of bandwidth on the component out. Who knows what your laptop will really output, probably even less than 1300. Even I wish there was some hard data on the true bandwidth of various component source devices. A Prof HD Pattern Generator would probably be the best test device to use.

^^^Good points about the quality of laptop outputs for those using dr1394's MPEG resolution test pattern software. What might help would be comments from folks in the calibration forum or elsewhere that have tested 1080p fixed-pixel displays with this technique and read ~1920X1080 from DVI/HDMI and RGB or YPrPb. Cable/DBS STBs seem like a somewhat different category because of the distribution path for signals like HDNet's Saturday am resolution-wedge test patterns, easily limited resolution-wise by cable/DBS rate shaping and perhaps STB memory limitations. Similarly, using the included resolution-wedge 'hidden' pattern on Sony Blu-ray movie discs, more reports of full-resolution HDMI/YPbPr readings would help confirm Blu-ray player/disc outputs. Pattern generators, especially calibrated ones, would indeed be best. -- John

I have used these patterns and other PC generated patterns via both DVI-HDMI and Component cables from my HTPC fitted with an nVidia 7800GT at the time.
Feeding my 70” SXRD via DVI-HDMI and Component both connections passed 1920 just fine. There was a slight loss of MTF via component, but it was not significant.
Feeding my 57” Hitachi CRT RPTV via component I was able to resolve the 1440 pattern, however MTF was low indicating that 1440 was pretty much the practical limit.

For those not familiar with MTF (Modulation Transfer Function), it’s a measure of line to line or pixel to pixel relative contrast at a specific spatial frequency. To be fully resolved an alternation black and white line test pattern should show black lines as full black and white lines as 100% white, this is 100% MTF.
It is not possible for a projection display to have 100% MTF because lenses are not perfect, there is always some bleed between lines or pixels due to lens blur. Flat panels are better as they don’t have a lens, however they do have a glass or plastic layer over the panel which acts as a diffuser so 100% MTF at 1920x1080 is not possible with them either.

What seems to be unknown to the most people is that anything shot with a camera and transferred to a 1920x1080 digital video format can never have high MTF at 1920x1080 due to limitations in cameras, film, digitising equipment and the need to low pass filter to avoid artefacts, in fact MTF at 1920 is likely to be 10% at best, probably much worse with most source.
10% MTF is effectively useless for the display of detail in video, a black line would show as 45% grey and an adjoining white line as 55% grey. 50% MTF is considered blurred and 30% MTF is generally considered the minimum useful for video images.

So when using the test patterns, be aware that unless the pattern is displayed with strong contrast that resolution will be of little use for video. What is useful is strong contrast ( MTF) at middling spatial frequencies as this is what makes an image look sharp to the eye. Having a display that fully resolves 1920x1080 is really only important for PC text and graphics which are created at the PC output resolution and therefore have inherently 100% MTF at the output pixel resolution.

^^^Nice summary, Owen. Folks interested in more details on MTF, horiz. rez., and 'sharper' images versus higher-rez images might want download (6.86 MB, PDF) Hans Kiening's tutorial (http://www.arri.de/camera/tutorials/4k_systems_theory_basics_for_motion_picture_imaging.html) about 4k+ systems for motion picture imaging. Concepts generally apply to HDTV. -- John

Related to the OP and comments, spotted a review of JVC's new LT-42X899 42" 120-Hz, 1080p LCD display in the Jan/Feb Widescreen Review. Reviewer Mike Marks found quite a horizontal resolution difference between the HDMI and YPbPr inputs. With an Accupel test generator, the 1920-resolution test pattern was nearly a "solid gray mush" with the analog YPbPr input but "sharp and clear" with HDMI. The analog input rolloff was severe enough to even smear a 6.75-MHz SD test pattern from the SD Avia disc! -- John