For a while now Canon have provided the ability to correct vignetting during RAW conversion using Canon's DPP software. The required correction is stored in a database and applied on a lens by lens basis when one of the database lenses is used. In the Digital Rebel T1i Canon have extended this facility to apply the correction in the camera to images saved as JPEGs. While more extensive correction and greater control is available via DPP, the in-camera correction can nevertheless do a pretty good job if you prefer to save space and minimize effort by shooting JPEGs rather than RAW images.
The EOS Rebel T1i's Auto Lighting Optimizer is designed for use in tricky lighting conditions. It can lighten dark areas of a scene while ensuring that bright areas maintain tonal detail. It's available in all shooting modes, including P, Tv, Av and Manual. It can also be applied (or removed) post-exposure when RAW files are converted using Canon's DPP software. The effect is quite subtle (even in "strong" mode) on most images.
This function is the same as that found on the EOS 40D and EOS 50D. It helps prevent over exposure of highlights by giving about 1 stop more range on the highlight end of the exposure scale. One way to think of how this works is that the highlight sections of the image are exposed at an ISO speed one stop slower than the shadows. It's available over the ISO 200-3200 range (200-1600 on the EOS 40D). So, for example, at ISO 200, the highlights may be effectively exposed at ISO 100, while the shadows are exposed at ISO 200. The main drawback of HTP is that you may get a little more noise in the shadow areas than without it.
Like the XSi, the Digital Rebel XSi has a contrast detection AF mode, which bases AF on the image as recorded by the sensor. This means that during contrats detection AF, the LCD image doesn't black out (since the mirror stays locked up). The downside is that contrast detection AF can be slow and it may take several seconds to get focus. The T1i can also use phase detection AF (the AF system used in normal operation) which is faster and more accurate, but requires the reflex mirror to drop, which blacks out the "live view" on the LCD for a second or so.
Face Detection Live View mode uses contrast AF to recognize human faces. When multiple faces are detected, the largest face closest to the center of the frame is targeted as the AF point.