[FEAT] Instrumentation Search Criteria

[igosplatt]

I understand that there is a general instrumentation search option available but I'm wondering if you could make the search for certain instrumentation easier and more standardized. Using a score order number system (fl.ob.cl.bsn - hn.tpt.tmb.tba - perc.hrp.pno - vln.vla.vcl.db) users can search for a specific number of parts for a given piece. For example Mozart's Prague Symphony would have the numbers 2.2.0.2 - 2.2.0.0 - 1.0.0 - 2.1.1.1 for 2 flutes, 2 oboes, 2 bassoons, 2 horns, 2 trumpets, 1 percussionist (timpani), 2 violin parts, 1 viola part, 1 cello part, 1 bass part. If there is an English Horn, piccolo, E-flat clarinet, contrabassoon, or bass trombone you can specify those instruments 3(1pic).3(1eh).3(1eflat).3(1cbsn) - 4.3.3(1btmb).1 - 3.2.0 - 2.1.1.1.

This allows for the instrument search to be much more specific. I can search for just string quartets (0.0.0.0 - 0.0.0.0 - 0.0.0 - 2.1.1.0) or string quartets that use two cellos (0.0.0.0 - 0.0.0.0 - 0.0.0 - 1.1.2.0). I could have a very strange ensemble and I want to somehow find music for it (1.0.0.1 - 1.0.1.0 - 3.0.0 - 3.2.0.1 or 1 flute, 1 bassoon, 1 horn, 1 trombone, 3 percussionists, 3 violins, 2 violas, and 1 bass [no cellos]). I should be able to extend my search options so that it finds all pieces with not only the exact instrumentation (which is unlikely) but also pieces with very similar instrumentation (plus/minus a few instruments).

So is there a way to incorporate this more specific instrumentation search into the system?

[steltz]

Since you joined less than an hour ago, I'm going to assume you haven't yet met our "Category Walker" in the "Browse by work genre".

This will do what you are asking. For string quartets with non-standard instrumentations, do the following:

Click "Work genre" in the left column
Scroll down to "Instrumentation"
In the second group you will see: [walk] For 4 players [subgenres]
Click "subgenres"
In this rather long list, if you want all possibilities, you must look under cello, double bass, viola, and violin Here's the list:

cello, 3 double basses
cello, 2 violins, viola (arr)
2 cellos, 2 double basses
4 cellos
4 cellos (arr) = arrangements for 4 cellos
4 double basses
4 viols (arr)
viola, 3 cellos
2 violas, cello, double bass
4 violas (arr)
4 violas da gamba
violin, viola, cello, double bass
violin, viola, 2 cellos
violin, 2 violas, cello
2 violins, cello, double bass
2 violins, 2 cellos
2 violins, viola, cello
2 violins, viola, cello (arr)
2 violins, 3 violas da gamba
2 violins, 2 violones
2 violins, viola, double bass
3 violins, cello
3 violins, cello (arr)
3 violins, double bass
3 violins, viola
4 violins
4 viols

Happy hunting!

[igosplatt]

Whether I joined just now or three years ago is immaterial. The walker is a nice attempt... If you are satisfied with that then so be it, I'm merely offering a simpler search feature suggestion. Thank you.

[steltz]

It's only simpler in some ways. For example, there would be no way to distinguish whether 1.1.1.1 - 1.0.0.0 - 2.1.1.1 is a piece for woodwind quintet and string quintet i.e. one instrument on a part), or a symphony with multiple strings on a part. That would make it very unclear indeed.

And where is your provision for lutes, violas da gamba, accordion (yes, we do have a couple of those), chalumeaux, basso continuo? In fact, just look at the number of non-modern-orchestral instruments in the list of instruments available.

We have almost 90 possible tags in the current system. Even counting the auxiliary orchestral instruments, your numbering system makes provision for less than 15% of those. You might be interested in only the orchestral instruments, but other people might not be.

And apart from the fact that notation for some pieces would be way too cumbersome -- a cello duet would be 0.0.0.0 - 0.0.0.0 - 0.0.2.0 -- the orchestral system doesn't provide for the most basic of chamber groupings, the piano trio.

You use abbreviations for the non-standard instruments orchestral instruments, so that could be expanded, but abbreviations are what we use for the current system. Only 12 of the 90 tags would use numbers, the rest would have to use pretty much what we have now.

Even if this were workable, what would all this notation result in? A list of possible groupings that users have to scroll through, find what they are looking for, and click on the specific instrumentation to get the list of titles. Actually, this is what we have now.

I'm afraid I find it unworkable to try to shoehorn the whole of music history into the modern orchestral notation, useful as it is for that ensemble.

Oops, I guess my instrumental bias is showing - I've completely forgotten the vocal and choral music which also need language tags. Where is your provision for that? Better increase the 90 tags to well over 100!

Sorry, no go.

[igosplatt]

Do your moderators always respond so rudely to mere suggestions? There is a way to say "sorry, no go" without saying "sorry, no go."

Rudeness aside, you make several good points about the variety of instruments in your catalog. What I'm imagining isn't necessarily a stream of numbers but several categories for the numbers. Each category by default is set at - - - - so you don't have to have 0.0.0.0 - 0.0.0.0 - 0.0.2.0 just for a cello duet.

Woodwinds
- - - -
Brass
- - - -
Percussion (perc,hp,pno)
- - -
Strings
- - - -

It would be easy to add a checkbox to search for only chamber music. Chamber Music [x]. There, now it's searching for nonets instead of symphonies. As for the inclusion of other instruments such as the English horn, the notations I mentioned above are standard score abbreviation and number systems that publishers are including in scores today and even on their websites (see Luck's Music). So I'm not making up anything new here, that is a somewhat standardized method of orchestral instruments. Naturally it would be just as easy to create a vocal category.

Voice (S A T B)
- - - -

For more advanced choral pieces you can include other voices in an "advanced search" option.

Voice (S S2 MS A A2 T T2 BT B B2)
- - - - - - - - - -

Aside from the most typical arrangements of orchestra/band/choir scores there is no reason to adapt the search method of non-orchestral instruments. You are correct, to try to summarize the vast instrumentation you already have for typical non-orchestra instruments into a number system would be ridiculous and useless. Therefore, don't summarize those instruments into a number system. They obviously don't need it. As for "normal" (and I say that knowing full well that doesn't really exist) instrumentation this search feature would be much simpler than having to sort through "viola da gamba" stuff just to find a the walker that lets me search for orchestral scores. I would rather type in all the instruments I have and want to search for once than click through several pages and sort through every single instrumentation available on IMSLP to find what I want.

You've got a good thing going with IMSLP here. I really hope this isn't how you respond to all suggestions made to the site.

[KGill]

As for "normal" (and I say that knowing full well that doesn't really exist) instrumentation this search feature would be much simpler than having to sort through "viola da gamba" stuff just to find a the walker that lets me search for orchestral scores. I would rather type in all the instruments I have and want to search for once than click through several pages and sort through every single instrumentation available on IMSLP to find what I want.

I think the central issue that we need to acknowledge here is that IMSLP does not tag for specific instrumentations for orchestral works. Therefore, if you just want orchestral stuff, it will get you nowhere to search for things with the English horn, etc., because nothing you're looking for is even registered in that part of the system. The relevant categories are 'For orchestra' and/or 'Scores featuring the orchestra'. (The system was set up this way for efficiency reasons on the end of the categorization team. It would have taken considerably longer to get even as far as we have if the members of the team had had to open the score of every single orchestral work on the site and figure out the exact instrumentation.)
That said, I still think it's far from a bad idea to provide this alternate method of searching for specific instrumentations - it would probably make it a lot simpler and faster for someone who knows their way around the abbreviations. Hopefully this will come up for consideration to actually be implemented this summer...

[steltz]

Whether I joined just now or three years ago is immaterial. The walker is a nice attempt... If you are satisfied with that then so be it, I'm merely offering a simpler search feature suggestion. Thank you.

I'm sorry you took my reply as being rude, but I took that sentence as being rude and not a little arrogant when I had very nicely given a full explanation of how to find what you wanted. Perhaps that's not how you meant it, but it came across that way.

I am also not sure that running two different systems concurrently is a good idea, but I will bring up my concerns with the tagging team when the time comes.

[pml]

I would only add that the programming of orchestral shorthand into the category walker would introduce an order of magnitude more complexity, and most of the 40,000 works on the site are not orchestral. It’s also possible the 2,400 works which are purely “for orchestra” might resolve out into 1,000 or 2,000 different orchestral permutations, and searching through that — or finding “similar” orchestrations to such a search — is probably non-trivial.

It’s worth noting that presently we aren’t classifying the solo voice types in opera, or the disposition of the chorus within choral-orchestral music: the field is really hugely complex and varied. To take a work for which the authoratative full score has been uploaded, Les Troyens:

Solo roles: 3sop 2mez 1(2)alt 4(3)ten 1(2)bar 9(8)bass — the brackets indicate an optional voicing: and these are the number of roles, not necessarily the number of vocalists, since no distinction could be made between a major role and a four bar solo voice* without adding to the complexity of the notation. (* Literally: le Dieu Mercure, page 589 of IMSLP #108956. A single word repeated thrice; four bars total.)
Chorus: not even going there, but assume SSSATTBB at the very least (possibly broken into multiple groups)
Orchestra: picc+2(picc),2(ca),2(bclar),4 – 4,2+2cornet,3,oph/tub – timp, perc (5) – harps (6–8), str; on-stage: 3 ob, 3 trbn, 8 saxhorns, timp, perc (3)

Also bear in mind that the orchestral shorthand notation is inherently ambiguous when dealing with extra doubling and ad lib. instruments; the more unusual instruments would have to be specified explicitly differently — take for example, the following woodwind spec., noting that it doesn't indicate doublings:
2+4+alt flt.4+1+bss obo.2Eflat clt+4+1+cbs clt.4+1+cbs sar

Instrument doublings are not equivalent (again comparing just woodwinds):
3(picc).3(eh).3(bcl).3(cbs)
1+2.2+1.2+1.2+1 (note stave order!)

To what level of detail would string divisi be necessary to document: if a composer specifies 16.16.12.10.8 does that really mean 62 distinct string parts (and for some composers the answer would be yes, while for others those numbers would be necessary to balance the remainder of the orchestral weight even without divisi in the five major divisions).

I’d be fully in favour of having the category walker support orchestral shorthand — but I wouldn’t want to have to do the programming for it. Or for that matter, both debugging and correctly categorising the orchestral works. :(

Regards, Philip

[KGill]

I’d be fully in favour of having the category walker support orchestral shorthand — but I wouldn’t want to have to do the programming for it. Or for that matter, both debugging and correctly categorising the orchestral works.

Just to make it clear, I only meant to suggest that it be used for non-orchestral chamber works with solo instruments for which we have exact numbers of everything tagged - not to go through all the orchestral works and introduce a new subsystem of tagging for specific instrumentations. I don't think limiting ourselves to the already-tagged solo instrumentations would be an insurmountable task for the hypothetical new search method, and it would still be useful.

[steltz]

I only meant to suggest that it be used for non-orchestral chamber works with solo instruments for which we have exact numbers of everything tagged - not to go through all the orchestral works and introduce a new subsystem of tagging for specific instrumentations.

But would it not equally introduce a new subsystem for the chamber works? The chamber works are precisely where a lot of the non-orchestral instruments are used, so the orchestral system itself is not that well suited to defining it. And at the end of the day, if you had 30,000 works with "2vn va vc" and like tags, you will still have 30,000 works with 0.0.0.0 - 0.0.0.0 - 2.1.1.0 and like tags. I don't see that the list created by the new system would be any different -- it will still have all the chamber works on it.

The idea of categories: Woodwind Brass etc. can be done without changing to a 0.0.0.0 - etc. system, it could be done now, probably very simply without major system changes and re-progamming and, most importantly, re-tagging nearly 40,000 works. After all, the system was built to search for things like "2fl". However, it also needs to be borne in mind that many of these works will be in a "mixed" category anyway, and also that the further one gets back into the pre-Romantic periods, the more "open" the instrumentation, so the system would have to somehow include the works that could be done on any instrument.

I am well aware that 0.0.0.0 etc is a standard system, and I use it a lot myself (and am familiar with Luck's), but it is historically standard for orchestral music, not chamber music. As a tagging system for chamber music, I think it has to be mixed with too many other things in order to make it workable, most especially with non-orchestral music such as chamber music.

Please don't take this the wrong way, but if the original problem that the proposed new system's:

search feature would be much simpler than having to sort through "viola da gamba" stuff just to find a the walker that lets me search for orchestral scores

then the current system will do this with merely two clicks:

1) Click on "Browse scores: Work genre"
2) Scroll down beyond Work Types; in the third section of Instrumentation, you click on "For orchestra"

There was never any need to sort through viola da gamba repertoire to get to orchestral scores, and no new system will make it simpler than two clicks.

[KGill]

But would it not equally introduce a new subsystem for the chamber works? The chamber works are precisely where a lot of the non-orchestral instruments are used, so the orchestral system itself is not that well suited to defining it. And at the end of the day, if you had 30,000 works with "2vn va vc" and like tags, you will still have 30,000 works with 0.0.0.0 - 0.0.0.0 - 2.1.1.0 and like tags. I don't see that the list created by the new system would be any different -- it will still have all the chamber works on it.

I still don't think I've made my meaning clear. I thought the suggestion was more like a shortcut to the tags we already have - so there would be a row of 12 boxes with dividers, you would enter numbers, and the code would interpret your entries to look up tags registered in MW:G. So, for instance, if I entered 2.0.0.0. - 1.0.0.0 - 2.0.1.0, it would be interpreted to mean the same thing as 2fl hn 2vn vc, and the category 'For 2 flutes, horn, 2 violins, cello' (if there is such a thing) would pop up. In addition, a useful feature would be to automatically modify the numbers a bit and bring up a list of 'similar instrumentations'. This would require a bit of code trickery, but we already have 95% of the stuff in place to make it work (ready-made tagging system and possibly the handy new search function that Feldmahler implemented not long ago). And I agree that it is not a replacement in any capacity for the system we already have: it is just a shortcut so that people who are familiar with the abbreviations can find certain things quicker.

[pml]

But the original poster was asking to be able to map orchestral works as well, not just chamber works such as string quartets (with two violins, or unusually with two cellos) — their first example was Mozart’s Prague symphony, and s/he also cited two large orchestral specifications where the type of search wanted would be, “find an exact piece with these instruments, or as close an example to this orchestration as possible.” None of the orchestral works have been tagged in a way to allow this to occur.

Cheers, Philip

[steltz]

Kenny, as Philip says, what you are suggesting is different from what the poster wanted, but it's perhaps not a bad idea for chamber music that uses orchestral instruments, as long as it is clear that for groups using other instruments, it won't work.

What I think we need to guard against, and again, I don't want anyone to take this the wrong way, is requests for huge structural changes when the user didn't really take enough time to see how the current system works, or find the maximum capability of the current system first. After all, a new user will most probably not be on the team that has to implement the changes, so other people have to do all the work. And given our very recent milestone of 100,000 works, nothing is small on this site any more. (I'm not complaining about that . . . . )

[feldmahler]

Using a shorthand search system for chamber works is definitely possible, and may not even be that hard to implement. However, two notes: (1) tags cannot be searched, because tags are treated as malleable, hence the search will need to be of actual categories and (2) I have a few other higher-priority features I am planning, so this may not get implemented for a while.

[pml]

(And again, the search criteria request was not strictly aimed at chamber works alone.)