Music Composers Unite!
Hi! At last I finished designing a structured timeline of composers and styles:
I am ready for your comments and please tell me if something is wrong! I will correct it.
On this timeline composers are grouped into colored areas, which identify styles. Some areas overlap, which means that some composers worked in several styles. This visualisation is meant to help studying and structuring composers and styles.
The start and the end of composer rectangle is the start and end of his life. Composer color shows his country (please see the legend in the timeline for colors). Bold font shows more well-known composers. Most styles have a thick colored line at their top, which shows the main interval when the style was active.
Click on composer, style or period to know more and listen to composer works.
And by the way, here is a similar clickable timeline of music genres:
I glued the egg to the top of the sky and then photographed it upside down.
water bear done said, yes she did:
I have just seen your time line of composers, and I think this is quite interesting, useful, and very pedagogical. I also very much like the hyperlinks from composers to recordings of their music. Just a few words about me. I am a professor of economics at the University of Ottawa, Canada, but I also do research on music and composers.
I would like to make a few comments on your composers’ time line:
As a researcher I can understand that adding new composer names to the time line is intellectually very satisfying. But I would suggest that you also keep in your archives three or four different “levels” or “versions”. The basic version would be of interest to people who start to get into classical music. They do not need to know (at that stage) that there exists a relatively obscure composer from country Y that can fit into the composers’ time line. A second timeline would be slightly more detailed and relevant to people who have more exposure to classical music; and finally a more complete version which would target classical music experts.
What I like from your time line project is that this is not simply a list of composers from, say, Renaissance period, or Baroque, or Classical, or Romantic, or 20th century. There are zillions of time lines like that. Instead you try to discover the interconnections within groups of composers. I understand that by reading music dictionaries (e.g., Grove Music online) and books on history of classical music (e.g., Taruskin, Grout, …) you can improve your understanding of the connections between composers. However, there is perhaps another way to do it, using statistics. As Charles Smith “classical music navigator” was mentioned in this thread, let me tell you that Charles and I have written two academic papers with the objective to track the similarities between pairs of composers. In fact we have built bilateral indexes of similarities across 500 composers (hence, we have generated 500x500 similarity indices). See for example:
Smith, C.H. and P. Georges, "Composer Similarities through the "Classical Music Navigator": Similarity Inference from Composer Influences", Empirical Studies of the Arts, 32(2), 205-29, 2014.
Smith, C. H. and P. Georges, "Similarity Indices for 500 Classical Music Composers: Inferences from Musical Influences and 'Ecological' Measures", Empirical Studies of the Arts, 33(1), 61-94, 2015.
Please let me know if you are interested in these articles (the links above are links to academic journals and if you (or your university) do not subscribe to these journals, you would have to pay some fees).
Finally, and that is where your work is also quite interesting to me, I am currently trying to build a Darwinian model of music evolution with some kind of phylogenetic tree of music speciation and evolution. I am sending you the abstract of the paper, but we could discuss further if you were interested. Here is the title: “Classical Composers Influence Network: A Statistical Analysis of Imitation, Differentiation and Evolution” and here is the abstract: This paper proposes a statistical analysis that captures similarities and differences between classical music composers with the eventual aim to understand why particular composers sound alike if their ‘lineages’ (influences network) are different, or why they sound different if their ‘lineages’ are similar. In order to do this I use statistical methods and measures of association or similarity (based on presence/absence of traits such as specific ‘ecological’ characteristics and personal musical influences) that have been developed in biosystematics, scientometrics, and bibliographic coupling. This paper also represents a first step towards a more ambitious goal of developing an evolutionary model of Western classical music.”
In conclusion, I just want to mention that I really like what you are doing, including your other projects such as for example the association between composers music and painting (another of my interests).
Hi, Patrick. I am glad you find the timeline useful.
1. I tried to add only the most important composers to the timeline. Also, the most well-known composers are marked with bold font. This allows both basic and detailed versions to be on the same timeline.
2. I tried to use both understanding and statistics when making timeline, but your approach seems much more profound from the mathematical point of view. I was able to read one of your articles. What do you think about trying to create a 2D model of dots (one dot per composer), where most similar composers will have greater gravitation (attractions) towards each other? After creating different 2D renders and checking the result with specially developed rules we can create a surface where groups of composers will be visible. Also, one axis can be fixed like in my timeline (x position is birth year).
3. This is an interesting idea. Do you use the same database of characteristics as in "Similarity Indices for 500 Classical Music Composers: Inferences from Musical Influences and 'Ecological' Measures" article?
I feel that the timeline has several problems, that (may be) can be solved using results of your research:
1. Folk. Too big and unstructured.
2. Neoclassical. Too big and unstructured.
3. Neoclassical. Too big and unstructured.
4. Film. Too big and unstructured.
5. Romantic. Its structure is based on my intuition. Also, I do not like that several composers were not structured here (from Paganini to Glazunov).
I will try to look more detailed at your results later.
Thank you for your reply.
I agree with you. But I would be surprised if you do not receive pressures down the line on you. Why no Portuguese composers? Why not Czerny, an important link between Beethoven and Liszt? Why not Henri Dutilleux, a major 20th composer who wanted to follow his own path without imposing his views on others. My coauthor, Charles H. Smith (from The Classical Music Navigator website) tells me that he keeps on receiving pressures from people to add new composers to his list of 500).
This is want I wanted to do. But it is not easy. I do not like the Tables in my paper you have read because I would like something more visual. However, think of it: let us suppose we try to rank composers by birth’s dates along the X axis, and by similarity along the Y-axis. Let us start with one subject composer, say, Vivaldi (born 1678) and let us just pick two other composers that may be similar to him in my similarity ranking (say, Rameau, born 1683 and Handel, born 1685). Now it is easy to visualise that Rameau will be to the right of Vivaldi on the x-axis and Handel, himself, to the right of Rameau. Now, let us think of the vertical axis, going say, from 0 (totally different) to 1 (exactly similar). Vivaldi is identical to himself: He would be positioned at 1 on the y-axis, so we can visualise Vivaldi as a dot (vector) in this space (x-axis: year 1678, y-axis:+1). Rameau and Handel receive, say, a similarity index of 0.7 and 0.8 wrt Vivaldi. We could then visualise Rameau and Handel as one dot for each that will be located to the South-East of the dot identifying Vivaldi (Rameau at (1683, +0.7); and Handel at (1685, +0.8). In fact, as I have computed 500 similarity score wrt to Vivaldi, we could neatly visualise 500 dots/composers, some being very tied together around Vivaldi, and some very distant. Most likely, the cluster around Vivaldi is made of composers from around the same period.
A problem, however, arises as soon as we would redo the same procedure for a second subject composer, say Rameau, on the same graph. If the subject composer Rameau is a dot at (1683 +1), we now have two dots for Rameau: this one and the one identified before: (1683, +0.7). And we will have two dots identifying Vivaldi as well (1678, 1) and (1678, 0.7).
The problem is that the similarity index identifies (of course) a relation between 2 composers. Furthermore, I also worry about the transitivity issue (if Vivaldi is “close” to Rameau, and Rameau is “close” to Handel does that make Vivaldi “close” to Handel? Probably yes with my index. But how close on the graph?)
I think a visualisation of the similarity index with the time index on the x-axis will run into many problems. An algorithm used for social network might work, but I do not know. Perhaps you have a suggestion?
Yes, I have used the same database as in the paper you read. Here, I have visualisations: for any subject composer I am able to visualise which composers have the same “lineage” (personal influences) and which ones were composing in a somewhat similar ecological niche as the subject composer. In fact for any subject composers, I have 499 dots (all other composers) that can fall in one of four quadrants, permitting to infer which composers tried to differentiate themselves from the subject composer, or which composers were simply imitating the style of the subject composer. Also I can use statistical tests of significance due to the similarity formula I have used.
Anyway, thanks for reading. Perhaps this is a bit too technical for the Forum and I apologise if this is the case. Alexey, if you are interested to discuss these issues further, maybe we can pursue by exchanging emails? (I guess it is possible to do it without giving our email address to the whole list.)
Great work - very intelligently done and most informative, thank you.
A small point - in the 'Folk' Section I would be tempted to put Elgar in bold type in preference to Britten.
Patrick, sorry for late response.
I sent you my friend request to exchange emails.
Thanks Alexey. I have sent you my email address.