We often describe Synchromesh as a 'Multi-Res' playback system, and there's a good reason for that. What we do is keep each video in a native format as long as possible, avoiding scaling or conversion as much as possible. In the case of resolution, and interlace mode, that means trying to only scale the video once.

Let's look at an example - I'll add diagrams when I get a chance - of a DDR unit, with one DVI output set to 1920 x 1080, one output set to 1280x720, one SDI connector outputting 1080i and one outputting NTSC.

We'll play a video - let's say it's 1080i 4:2:2. It's converted to 64-bit floating point RGBA 4:4:4(:4), with bit-depth dithering, but kept at 1920x1080 - the 2 fields are interleaved to make essentially a progressive frame, just for utility. We'll also assume that it's 16:9 (square pixels) and so are the HD outputs.

Now, DVI Program Output 1 will access those 64-bit frames, and de-interlace them - but not scale them, passing them pixel for pixel. DVI Program Out 2 will de-interlace and scale down. SDI Out 1 will not scale or de-interlace, in fact it will make sure that field 1 is output on the output's field 1 and vice-versa. SDI output 2 will de-interlace, scale down (letter-box also) and then output.

Now the user plays a 720p video. Each output again does it's own scaling, but this time, DVI out 2 is pixel for pixel. Note that the video did not get scaled up to 1080, then back down to 720p - that would cause a loss in quality and a decrease in contrast, even with our scaling.

Even an NTSC video, or a custom screen res piece will get treated correctly, with only one point of scaling. 

How the user takes advantage of this system is up to them, but what we aim to do is give you the absolute best quality on each output, and remove the need for a time consuming and lossy resolution conversion.