Allows you to specify the video property of the source/input video. This is necessary to retain the correct aspect ratio if your input video does not have square pixels (eg. if it is an anamorphic DVD source). Normally, your encoding software will already have a setting that deals with this (as well as resizing), so "Square Pixels" is the default selected value.
Cropping allows you to remove a part of the picture that you don't want encoded - for example, the black bars on widescreen movies (they will be added back by your DivX player anyway, so there is no point wasting bitrate on them). Generally, your conversion/encoding tool will have a similar feature, so you should just ignore this section.
When the resolution has been changed, this option becomes available. This option allows you to choose a resize filter.
As the various selectable options describe, these filters will determine how soft (blurry) or how sharp the output image will look. Bicubic (Normal) is usually a good middle ground type selection. The Lanczos4 option will produce a very sharp picture, but the problem with a sharp picture is that the noise and jagged edges are enhanced as well, making the picture look very unnatural. Please note that resizing will slow down encoding times. In DivX 6.2, dual-core/hyperthreading support is now supported by the resize filter, making resizing faster.
You can select between H.263, H.263 Optimized and MPEG-2. The recommended quantization is H.263 Optimized. MPEG-2 quantization may not be supported by all DivX certified devices, and hence only recommended for very specific content, but with possibly increased artifacts for other types of videos. In general, MPEG-2 quantization is not very efficient and shouldn't be used.
You can encode the video as progressive or interlaced. If your original source is interlaced, then encoding it as interlaced may result in a better looking picture, although it may increase file size. Similarly, encoding the video as progressive if the source is progressive will be benefitial. There are many ways to check if your video source is interlaced or progressive, but using DVD2AVI and checking for the "Frame Type" is a good way to achieve this.
The middle option allows you to "De-interlace source", which is just a way to convert the interlaced source into a progressive one. This is not really recommended, as it can lower picture quality.
Just like how MP3 encoding removes bits that the human ear can't hear, to decrease file size, psychovisual enhancement does the same with the video. And just like MP3, some people can see (or hear) the difference, but some cannot. Here's what Gej (the creator of DivX) has to say about this setting:
It's hard to make a recommendation here, it depends a lot on people preferences, I personally use the "mask" psychovisual all the time, as I find it to be efficient to reduce size without generating unwanted artifacts but it's the slower option, "shape" psychovisual reduces size even more, but has a tendency to accentuate ringing artifacts (little dots around hard edges), but it's faster than "mask".