Overhauled Search in Roon

No, why would it? How does that result lead to that conclusion?

You asked why your Pierre Boulez albums weren’t more visible. Search for “Pierre Boulez” is a search on a person’s name. In this case, the person is both a composer and conductor.

I merely pointed out that if you search for the name of a conductor that’s also a composer, the person’s composer results are given priority - vs a conductor that isn’t a composer, where you see the conductor’s conducting albums higher up in the results (less scrolling), as there are no composer albums to display first under that name.

That’s all I wrote. Why read something else into it?

And by the way, “Boulez conducts Stravinsky” gives totally consistent results with what I wrote: highest result is an album with that name, followed by albums where Stravinsky is the composer, followed by tracks. Unlike the search you asked about, it isn’t a search for the name of a person. I don’t see anything inconsistent or illogical in such results.

If you have issues, ask Roon, not me. I’m not in charge of their search results.

Is it -possible to search for playlist in Roon. Sometimes I know the name. But it does not pop up in a search

If the Playlist is in your Library, then it should be searchable. Roon doesn’t search Playlists that are out in the Qobuz or Tidal services.

Excellent search in my classical collection with no connection to streaming sites like Qobuz. I tested on Pierre Boulez and got everything of the composer and the conductor. It works very fluently, even when there is only one work of Boulez on an album in between other composers.

Perhaps some extra work is needed in the future with the search for different recordings of the same work. The number given with a certain recording does not correspond with the actual number of recordings found in my collection. Sometimes the number is too high.

2 Likes

@danny2

sorry to have disturbed you

My own exploration of the new search engine is here:

Today I discovered that Roon has a real fondness for violin concerti!

I think that being a good software developer does not prevent you from asking for help from a real musician, or music lover, or music critic to reference the works. Knowing who is Händel, Handel or Haendel is a music lover’s job, not a developer’s job. No disdain for developers. Only: to each his own. :slight_smile:

All your efforts and explanations applauded. But, yes I still experience questionable results in my search (Roon, Qobuz user). The speed is improved and display of results to my liking (vs. some others here?).
I did test some of the comments here and get similar results. See: Jesper Sorensen i.e. He’s right and the Haendel/Boulez is a thing.
Perhaps share some guidance on BEST PRACTICES for searches (if that exists.) Believe that I know searching classical is the most problematic and true test of capability since the dawn of time, er, Dewey Decimal System.
And I’ll state here and now: I LOVE ROON. I love my dogs, too, but they are unpredictable and misbehave. All the best to you all.

About Gustav Malher, what else ? (!!!)

We’re working on improving search so that it’s able to handle spelling errors (Mahler vs Malher) but it wasn’t something that was ready for inclusion as part of this release. For the moment, you’re only going to see results that actually include the word “Malher”.

As Brian mentioned above, there are lots more improvements to come, and spelling corrections are absolutely part of the roadmap – you can expect this query to hit as expected once that functionality is implemented.

5 Likes

Mike, the problem here is that you keep touting big improvements to search, then we try one simple thing (for me, the first search I tried for “Handel” returned Mozart), and it is really odd. And disappointing.

Why not do all these fixes and improvements you are talking about, and then roll it into your product and crow about it? Instead of the half-baked version you seem to keep giving us?

1 Like

Zenit already explained it to you here:

This is a result of pre-880 data mappings that are cached on your machine and not the new search. I’m going to give you the benefit of the doubt and guess that you missed that post.

A post was merged into an existing topic: Search for Handel Returns Mozart as Top Result

No, I saw the post, but so what? The fact was still that our experience with the “new search” touted in the release notes was odd and disconcerting. The reassurance that such things maybe won’t happen any more at some undisclosed and unknowable point in the future doesn’t really help much with that first impression.

Or am I wrong again? Will some flag pop up when the old outdated set of aliases are finally cleared out, so that the user can know what to think about it?

While the new search is certainly much better, there are still some gross errors. For example a search for “Beethoven Symphony 3” doesn’t find the composition:


And a search for “Mozart Piano Sonata 3” returns the right sonata but by the wrong composer.


@zenit - what is the best way to communicate these issues to you? Do you want them raised individually, or put in the general build 880 thread, or should there be one thread just for build 882 search issues?

1 Like

@mike

Yes I was wrong with the spelling of the name. So sorry. :frowning:

A post was split to a new topic: Search (build 880): typos not yielding good results

Thanks for asking, and thanks for posting. It’s good to keep all these examples here within this thread.

1 Like

First and foremost, thanks for all the continuing work you have put into the search feature. Is it perfect? No. But it does a pretty damn good job, especially compared to other products out there.
Oddities in non-classical searches in Tidal…
The original album “What’s Going On” does not appear in the discography of Marvin Gaye… but it does show up if you search on the title.
A search on Gladys Knight will result in her solo albums and some with The Pips, but not all. A search on Gladys Knight & The Pips will return all of them. BTW, Tidal treats them both as one entity.