Dimensionality reduction in Lucene

### Description

Hi. I am opening a new issue to follow up on a [discussion in this issue](https://github.com/apache/lucene/issues/13403#issuecomment-2132043000) regarding using the segment size to decide if we should do vector quantization or not. There are two things to figure out with this, 1) if it would be helpful, 2) how to build/tune?

I like to start the conversation and get some ideas and feedback from the community. Thanks.

Start with the discussion in issue #13403, especially the linked comment, and investigate whether segment size should determine whether vector quantization is used. Review the possible approaches and tuning considerations; done means the community has a clear direction on whether and how to pursue dimensionality reduction.

Written by the indexing model from the issue text.