Figure Recognition not Making Bass

Posted Wed, 2024-01-17 - 16:26

Hi,

I've encountered a situation I don't understand.

I have a small (demo) library where I am attempting Figure Recognition on some phrases of bass line. I've pressed the Auto button on the FR tab of the Take pane and preset 'Bass (melodic)' comes up - all good so far!

Then I run the Figure Recognition on the Phrase Pool, like this:

The resulting Figures though are Blue, not Pink. IOW, they are not Bass figures.

Am I mis-interpreting something, or is the program doing so?

Before (01) and after(02) library files attached as a .zip.

Attachments

bass-not-bass.zip

meestahp

Wed, 2024-01-17 - 19:14 Permalink

You need to apply the Bass recognition preset to at least one item in the pool then run the recognition, on the other hand I found a bug when dirtying the library and trying a revert.

andre

Wed, 2024-01-17 - 20:56 Permalink

The menu command runs FR with the last used settings that were saved after the last run. Settings are only saved when you use them.

sj1

Wed, 2024-01-17 - 22:30 Permalink

Thanks meestahp, thanks Andre!

It seems that the memory of last used settings is per pool.

IOW, if I go to a pool, I can manually do FR on Chunk 1 to get good bass. Then I can do the pool menu item to get good bass for the rest of the pool.

However, when I go to the next pool, I have to do the entire process all over again. Then next pool is not remembering that that last FR I used was 'Bass (melodic)'.

This is way more time and motion than is desirable for library construction scenarios where there are many pools that need the same FR applied to them.

I suggest that there be a way to make the operational memory of FR preset "sticky" across pools for the duration of a session. Of course a pool would still remember long-term how its last past FR was done.

I also suggest being able to do the FR, with the chosen preset, on all the pools in a library in a single operation. IOW, offer a 'Run Figure Recognition' option here:

If possible, we should have a way to be aware of what FR preset is being run in this case.

But if that is not possible, then "run one, then the rest can all be run the same" can be our guide.

sj1

Wed, 2024-01-17 - 22:42 Permalink

Maybe this dialog:

could become:

You are about to run Figure Recognition

Preset: <preset name here>

on all phrases. Are you sure?

The preset could be 'Automatic' too, under some circumstances, I expect.

andre

Thu, 2024-01-18 - 18:35 Permalink

FR settings are local to a phrase pool because they don't work with a wide variety of input. A manual test is needed before they can be applied to the entire pool.

Drums, bass lines and melodies are simple, but polyphonic material always requires manual review and testing. Otherwise you can run FR with the batch import automatically.

But I agree there is room for improving the convenience and workflow. I've added your suggestion to the wish list.