Hi Antoine,

My initial assumption for this design was to aid faster sample
creation in a long stream of data. Thinking was

   - Jump to a location in the stream
   - Select some random values close to this location (within the 1024
   proximity limit)
   - Find the next location (equidistant) to jump to

However, perhaps that isn't needed as the page is small enough to go for
this design.
Random set of values might end up working equally well. I can give this a
try.

Best
Prateek

Reply via email to