Electrosonic South Africa's technical wizard John Ungerer examines the confusion surrounding scan conversion and scaling techniques.
There appears to be a great deal of misunderstanding in the industry regarding scan conversion and scaling technologies, and while I shall not attempt to delve into their inner workings, I will offer a simple explanation of the processes and their different applications.
Scan conversion
Scan conversion is the process of changing scan rate (sync and data) of a signal, typically from a computer or HDTV device, into a video signal - which can be anything from composite to component video. A typical application would be to convert a computer image, e.g. 1024 x 768 resolution to S-Video, for inclusion in a videoconference over ISDN lines. By transforming a PC signal to video, one is even able to record these images on an ordinary VCR.
The drawback of scan conversion however is a substantial reduction in image quality. This is the result of a high bandwidth signal being transformed into a low band video image (typically 15.75 KHz). One could liken it to reducing a 100 mm water pipe into a 15mm water pipe with the use of a fitting: Even though there is 100 mm’s worth of water volume coming in, only 15 mm’s worth is able to be used because of the reduction.
This does not mean that scan conversion should not be considered, as it is certainly applicable in many instances such as multi-media, boardrooms and video conferencing and also in any application where high quality conversion of graphical data signals to video signals is required.
Scaling
Scaling produces an opposite effect; i.e. the 15 mm pipe becomes a 100 mm pipe. In reality it would be impossible to get more water out than is coming in, and in scaling terms, while one cannot have something in an image that wasn’t there in the first place, by scaling a video image up to data resolutions, one can manipulate and even improve the image to a certain extent.
Scaling is the process whereby any type of video format (composite video, Y/C, YUV and more) is converted into a uniform RGBHV (analogue) or DVI/HDMI (digital) output format at a set resolution with the best possible picture quality. These can be at computer graphics resolutions (VGA, XGA and higher) or HDTV component format.
Most display devices today are equipped with some sort of internal scaling device, however the majority can, at best, be described as mediocre. Far better results will be achieved using a stand-alone scaler to perform the task of matching the incoming signal to the native resolution of the display, taking care of the conversion to progressive scan and performing the other image enhancing processes.
The process of scaling does not end with improving picture quality. One of the biggest benefits of using scalers in a system is that it simplifies connectivity. Scalers essentially come in two types – the basic single format versions and seamless switcher-scalers. In many instances, the latter can almost provide a ‘one box’ solution for small boardrooms and training rooms. All that needs to be added is control.
To provide an example of using the smaller or single format scalers, imagine the following scenario: Two rooms - multiple displays (2 x projectors, 2 x plasma, preview monitor & PC input to video conference). Also multiple sources of various descriptions – 4 x PC (data), 2 x DVD (YUV component), 1 x satellite TV (composite video), 1 x video conference (Y/C)
These devices all have to be routed to allow for maximum flexibility.
The solution, not the problem:
Problem: Different and incompatible formats, but all the display devices are capable of displaying high quality data format.
Solution: 1 x 8x8 RGBHV and audio matrix and 4 x small scalers for video sources, with only RGBHV data cable connected from the matrix to all the displays. The computers already output data, so scaling is not required. Sorted! Not only is the installation simplified, there is flexibility for routing signals and limited control required for all the display devices, as they will always live on the data input. (Which all serves to confirm my favourite maxim: ‘when in doubt, throw in a matrix!)
Switching is faster, cleaner and more stable as one is not relying on the projector to activate its own cheap scaler chip when a video signal is detected. Instead, the work is done properly up front!