04-27-2022, 02:55 PM
@suzywong
[sidebar]
Kitten-sitting is good for the soul. Our "kitten" is 8 years old (we think; she's a rescue cat) but the principle is the same. We're old enough that three other cats preceded her in my case; four in my partner's case.
[/sidebar]
Given my own experience and the experiences reported by you and others, I'm of the opinion that the observable effect must depend, perhaps even nonlinearly, on the size of the library involved.
I feel the need for a well-characterized test artifact. Toward that end, I'm fleshing out a script to generate a file system containing fake albums. Each album directory contains a few uniquely tagged tracks and a unique random-pixel cover image. Each track contains the memorable 5-tone sequence from "Close Encounters of the Third Kind" which I created with a sine-wave generator (a guy's gotta have a little fun!).
The organizational structure of the file system will be such that I can mount a music source containing 625*N albums depending on which of the N-th subdirectories I publish as an SMB share. I'm thinking N will have to be at least 8 for good coverage; e.g., 5000 albums.
This way I can try to get some reproducible metrics without expending a lot of effort. Of course, this may be irrelevant to the current thread if you discover a magic set of browser settings but I'm too far down the runway to stop now
Besides, I can always use the script to generate similar artifacts for other test purposes.
But first, I have a day's worth of housekeeping chores to attend to..
Regards,
Kent
PS - why 625? Because 42 didn't seem big enough.
[sidebar]
Kitten-sitting is good for the soul. Our "kitten" is 8 years old (we think; she's a rescue cat) but the principle is the same. We're old enough that three other cats preceded her in my case; four in my partner's case.
[/sidebar]
Given my own experience and the experiences reported by you and others, I'm of the opinion that the observable effect must depend, perhaps even nonlinearly, on the size of the library involved.
I feel the need for a well-characterized test artifact. Toward that end, I'm fleshing out a script to generate a file system containing fake albums. Each album directory contains a few uniquely tagged tracks and a unique random-pixel cover image. Each track contains the memorable 5-tone sequence from "Close Encounters of the Third Kind" which I created with a sine-wave generator (a guy's gotta have a little fun!).
The organizational structure of the file system will be such that I can mount a music source containing 625*N albums depending on which of the N-th subdirectories I publish as an SMB share. I'm thinking N will have to be at least 8 for good coverage; e.g., 5000 albums.
This way I can try to get some reproducible metrics without expending a lot of effort. Of course, this may be irrelevant to the current thread if you discover a magic set of browser settings but I'm too far down the runway to stop now
Besides, I can always use the script to generate similar artifacts for other test purposes.But first, I have a day's worth of housekeeping chores to attend to..
Regards,
Kent
PS - why 625? Because 42 didn't seem big enough.
