“Transparent” Video Switching?
Dear Gary: As you know, many videophiles like to run their video cabling directly to their display, feeling that adding anything else in the signal path, such as receivers or switchers, are likely to degrade the signal. What is your opinion on the question of cable routing? Do you ever use video switchers, or run your video signal through a receiver? To make this question more personal: I have several component sources (DVD progressive scan, HDTV cable box etc.) that I need to route into my single component input on my plasma display. Until now, I’ve run my video cables straight into the display, with excellent results. In fact, I just borrowed a Nordost Optix S-video cable and was surprised at the jump in signal quality (the Nordost S-video cable signal is significantly sharper, with deeper blacks and colors than my component signal routed through my Pioneer VSX 49TX receiver!). Unfortunately it looks like I’ll need a component video switcher, and I’m concerned about maintaining top signal quality. In your opinion (or experience), would a good switcher, like an Extron component switcher, pass the signals with no visible degradation? And further, if I use top quality cables like Nordost Optix, would putting an Extron into the signal path negate any of the quality I find when running the Nordost straight into my plasma? Have you found you can still realize the benefits of excellent cables when routed through video switchers? My Pioneer VSX 49TX, with 45 MHz bandwidth, is THX® Ultra 2 certified, which means it’s supposed to pass through DVD and HDTV signals with no visible degradation. The signal through the 49TX is terrific, but there is a subtle visible change. Colors are deeper and richer, more 3-D, and the picture is also a tiny bit softer. Can you enlighten me as to what is happening to the signal to render these changes?
Rich Harkness
Video Technical Editor Greg Rogers Comments:
I prefer to run analog source signals to a standalone video upconverter (scaler), and then drive a projector or display panel with analog RGB or DVI digital video signals. Therefore, I want an upconverter that not only has exceptional scaling performance, but also multiple component video inputs. Two excellent examples are the FOCUS Enhancements CenterStage CS-2 that I review in this issue, and the Lumagen Vision Pro that I reviewed in Issue 81, February 2004, of Widescreen Review. The CS-2 has two standard-definition interlaced component video inputs, a standard-definition progressive component video input, a component video pass-through input, a DVI input, and an optional SDI input. It also has two S-video inputs, if you must use that format. The Vision Pro has a total of eight analog inputs and an optional SDI digital input. Up to four of the analog inputs can be configured to accept component video signals, and two of those can alternatively be used as high-definition pass-through inputs. Any inputs not used for component video can be used for S-video. The pass-through inputs have a specified 300 MHz bandwidth (-3 dB frequency response). If you don’t have an upconverter with multiple, high quality component video inputs, then you may need to use an external switcher. I haven’t evaluated the Extron switchers that you mentioned, but you can read my review of the Zektor HDS4 Component Video Switch in Issue 78, November 2003 of Widescreen Review. Most video processors and switches specify bandwidth, which normally means the -3 dB frequency response limit. That specification is not very useful because our vision is sensitive to about a 1 dB loss in frequency response. A switch should have a nearly flat response (-0.1 dB or less is desirable) within the applicable signal passband, because it will directly add to the existing frequency response roll off of the source signals. The absolute frequency limit of 720p and 1080i HDTV signals is just over 37 MHz, and the SMPTE specified filter passband is 30 MHz (±0.05 dB). Current HDTV sources have little content at that frequency, but a switch with flat response to 30 MHz can be certain to not soften existing or future 720p and 1080i sources. It isn’t possible to accurately predict the passband response of a switch from its -3 dB bandwidth, because the shape of the frequency response varies significantly from one switch to another. A 300 MHz bandwidth may yield a -0.1 dB response at 30 MHz, or it could easily have a roll off of -1.0 dB or more at 30 MHz. Instead of bandwidth, manufacturers should specify the frequency response loss at 30 MHz for switches intended for 720p and 1080i use, and 60 MHz for 1080p use. It is very difficult to achieve “transparent” analog video switching. Most people immediately think in terms of frequency response and its effect on image sharpness, but that is only one dimension of switch performance. There are two kinds of video switches—active (unity-gain buffered or amplified) and passive. Ideally, passive switches would have the same characteristics as a short piece of coaxial cable—unity gain and tightly controlled 75-ohm impedance from input to output. That would produce the best frequency response flatness and eliminate signal reflections that cause step response aberrations. The latter create edge outlining and ringing artifacts in images. Input and output impedance is also critical for active switches, but transient response anomalies in active circuits can create similar edge artifacts. Signal path lengths must be matched to avoid differential signal delays that produce color-fringing artifacts, which is particularly difficult to do in passive switches. Crosstalk between signal components will alter color accuracy, and, of course, crosstalk between inputs will wreck havoc in an image. Incorrect gain matching between signal components can also alter color accuracy in active switches, and extraneous noise from digital audio processing can be a problem in receivers and preamps with video switching. If I had to use external analog switching, I would favor a standalone switch that can be designed to specifically optimize all aspects of switching performance.
You can E-mail Widescreen Review @ mailto:editorgary@widescreenreview.com