I was reading some posts in the boards about PAL <-> NTSC standards conversion and a couple of product names were floated. I had never compared our standards conversion to anyone else, but rather simply worked from customer feedback that it was working OK.

I was pleasantly suprised by the results. It would seem that the output from MegaPEG.X for NTSC to PAL conversion is very similar to the commercial Nattress standards conversion package. Basically, in MegaPEG.X Pro, we use our image processing pipeline to resize each field independently, and then resample in time. The result is that the two fields in the NTSC footage map very nicely into PAL, and the byproduct of the time conversion is a light motion-blur effect. I had anticipated that the other products would have a way around this, but it is not the case. In fact, Nattress has a feature for additional motion blur. Who'da thunk it? To my eye, the JES approach is less appealing. It appears they are resizing both fields, and then reinterlacing, perhaps in the time domain. The result is quite different. Here are links to a frame from a fast-moving vehicle sequence shot near the Whole Foods in downtown Austin,TX. And bear in mind that both the JES and Nattress screenshots are taken from a DV version, and the MegaPEG.X screen is after encoding to MPEG-2 (on our highest quality, aka 'Forever Encoder').

   

Standards conversion in MegaPEG.X Pro is not a specific feature. Rather it demonstrates the flexibility of the design of our video processing features. In fact, we're making only minor changes to support upsampling and downsampling HD from 1080i to 720p, 576p and 480p.

The other benefit of using MegaPEG.X to do the standards conversion is that the routines are integrated aspects of the encoder pipeline, all optimized for the Velocity Engine. Not only do you eliminate the need to create an intermediate representation of your material, but you incur no appreciable penalty in MPEG encoding time.