TMPGenc maxxed out !

[arcorob]

I use TMPG to take my editted DV AVI's and convert to DVD MPEG. These are for wedding video's. Well, I tried something new.

Besides the defaults and some simple color correct and best Motion Estimate Search (very slow) I also set Noise Reduction (High Quality)

My 1 hour and 40 minute movie took 34 hours and 45 minutes to render !!! That would normally be about 3 hours.

Is it worth it ? What am I gaining ? Help appreciated.

[bugster]

Only you can decide if the improvements gained are worth the extra encode time. I would suggest taking a short 5 minute clip and encoding it twice. Once with your old settings, once with the new. Then compare them as best you can and decide if the improvemnets, if any, are worth the extra time.

[DJRumpy]

The higher (slower) motion detection setting is supposed to allow the encoding engine to find motion vectors from frame to frame. Put simply, it a person is walking across the screen, from left to right, and the motion detection properly detects that the object, or person, moving from left to right, is the same object from one frame to the next, then it can simply give the new frame the object's motion information (example: object moved two pixels to the left, and 5 pixels down), rather than actually encoding his image in every frame. The bitrate saved can then be applied elsewhere to improve the image. You should then begin to understand how a high motion movie would benefit from this setting (CAM's do as well, since they tend to be shaking unless you used a tripod).

If an encoder fails to detect motion properly, then it see's each change in a frame as a new object, and applies bitrate to the frame to produce the image. If there are a lot of new 'objects' in a frame (in a fight scene for example..lots of motion), and there isn't enough bitrate, then each macroblock is softened to reduce the overall bitrate requirement for the frame, until the amount of bitrate needed to produce the image is low enough to meet the available amount of bitrate. This is why you get the macroblock effect when a bitrate shortage occurs.

If I understand the methodology correctly (hopefully someone can correct me if I'm wrong) that each frame is made up of 16x16 squares called macroblocks. Bitrate and motion information is applied to each macroblock. This way the encoder doesn't have to see the 'big picture'. It can work on a small piece of it. This makes it easier to see motion in a 16x16 square, than to try and track an entire moving human figure..arms and legs flailing, etc.

If your a quality nut, you should only use Noise Reduction, if your source is noisy (as in analog noise..tape noise, etc). If you use it on a clean source, it can make your image a little 'soft'. If your having bitrate shortages though (macroblocking), then the noise reduction can be a good thing, as it will soften your image, and make it less 'complex' to the encoder, reducing the bitrate requirement.