Sunday, December 14, 2008

5020 and 6020 have a 3D Y/C comb filter

A Y/C comb filter separates the luma (Y) and the chroma (C) signals from a composite source such as laser disc or an OTA NTSC broadcast. Digital sources already have their luminance and color components separated but for analog composite sources, the video signal quality is highly dependent on the technology of the Y/C comb filter. Pioneer's documentation mentions that the Elite Kuro models have four settings of (Off / High / Mid / Low) for a video processing 3DYC feature. The documentation for the non-Elite Kuro mention that they have no 3DYC feature. So I decided to determine what type of Y/C comb filter Pioneer decided to put in the PDP-5020FD and PDP-6020FD Kuro plasma displays.

I did some testing with a couple of video test patterns, a Pioneer Elite CLD-99 laser disc player, and a Pioneer Kuro PDP-6020FD plasma display. In the late 90's the CLD-99 laser disc player was the first consumer electronics device to have a 3D Y/C comb filter. It was based on a digital 3 line comb filter design and it added a digital frame buffer to compare with the previous frame. This added time as a dimension, hence the 3D label. It was amazing new technology that was expensive, and it was only in the CLD-99. A couple years later 3D Y/C comb filters became more common place and were introduced in some higher-end RPTV's.

The S-video port of the CLD-99 laser disc player uses the 3D Y/C comb filter while the composite port has the combined Y/C signal straight from the optical disc which means that the comb filter in the display is used. This allowed for the comparison of the CLD-99's 3D Y/C comb filter to the comb filter in the non-Elite Kuro.

Static Test
First a color bar test pattern from a laser disc was viewed. The sharp transitions between the different colors is a good test of the separation quality of a Y/C comb filter. Lessor comb filters have an active noise shimmering-like effect between the colors. A 3D Y/C comb filter stabilizes the shimmering but it does have some video artifacts. On a PDP-6020FD display the CLD-99 S-video output looked about the same as the composite output.

Motion Test
For the second test the Snell & Wilcox zone plate from a laser disc was viewed. The zone plate has a fresnel ball that bounces around the middle square section. This motion causes all sorts of problems for Y/C comb filters and the result is chroma flashing. The zone plate also has a couple boxes with vertical and diagonal lines near the top that tend to have a lot of chroma noise with non 3D Y/C comb filters. The comparison of the S-video and the composite signals showed roughly the same image. The CLD-99 had slightly less flashing chroma noise with the moving ball than the Kuro did. The static vertical and diagonal lines in the boxes had slightly less luma noise with the Kuro but increasing the 3DYC slider control from the default value with the CLD-99 equaled and then improved that noise aspect.

Conclusion
The performance of the comb filter in the PDP-6020FD roughly equaled that of the 3D Y/C comb filter in the Pioneer CLD-99 laser disc player. So Pioneer's specifications are incorrect, the PDP-5020FD and PDP-6020FD both have 3D Y/C comb filters. The non-Elite Kuro just lack 3DYC configuration controls.

The CLD-99 had a slightly better Y/C separation than the Kuro but the results were close. This advantage is probably due the CLD-99 being optimized for the laser disc formats 425 lines of resolution while the Kuro's 3D Y/C comb filter is designed to work well with a resolution range from VHS (220 lines) to D1 (480 lines).

The Kuro's 3D Y/C comb filter default also have a bit more Y and C noise reduction (NR) than the standard default in the CLD-99 player. The CLD-99 standard default setting is 50% for Y NR, C NR, 3DYC, and Sharpness controls. Increasing the CLD-99 players Y NR, C NR, and 3DYC settings by a couple ticks to about 75% equaled the image effect of the Kuro's default. It is unknown if this is equal to the 3DYC Med or the High setting of an Elite Kuro display.

No comments:

Post a Comment