181,788 users have contributed to 42,186 threads and 254,589 posts.
In the past 24 hours, we have 6 new thread(s), 25 new post(s) and 45 new user(s).
I am interested in concrete Ideas what might be improved in my current project
You can listen the current render here:
Felix Draseke Symphonic Poem "Frithiof"
Constructive and on the real recording related criticism are warmly welcome.
It would be great to give the exact time of the passage (it is displayed in the player) for which you might suggest an improvement.
best
Steffen
Hi Steffen,
first of all I want to state that the Draeseke Symphonic Poem is an
absolutely beautiful work of music. I wanted to listen to the music all
week ever since you posted it however I have been busy and not in the
right frame of mind to appreciate the music fully until today.
Tom
Hi Tom,
Wow, thank you so much for that much really helpful hints. In some aspects you exactly hit points I already was not that satisfied with but havend found a better solution yet.
1) This is especially true for the starting 30seconds. I already acted here (only abit and presumably still not enough) against the advice of the score, which demands Forte in all active instruments from the beginning. When I heard what that would sounded (it was the "wurlizer"-Effect even much worser) I decided to let most of the instruments make a little crescendo, since this is what I feel is the Idea of the vivid increasing passion of the thematic motives.If all accompagning Instruments would have "fired" with full blown Forte from the beginning, no one would ever heard most of the thematic motives. But it seem to me that I still was to shy in leveling down the begin dynamicly. I myself for instance still am kind of disturbed of the Basson which I could not bring with any Velocity X-Fade to beginn more smoothly. But I also feel the stringtremoli a bit less developing as I would have liked (if would have been the composer) but did not dare to, since Draeseke does not notated any hint of dynamic development for this beginning. Your Critic encourages me to perhaps give this beginning (meanwhile the score does not demand) much more dynamic development.
2) The passage at 6'36 of the second Movement gives me another Problem. Draeseke actually demands abrupt dynamic changes each one or two bars. While those with woods and brass are Forte, those without are mezzopiano.The Contrats are composed quite harsch the melody especially of the mezzopianoparts obviously is meant to make an weird impression. I was not really happy with my result, but as I said I dont have any Idea yet how make those fast contrasting bars more convincing. 
3) III.11'30 Yes I hope I understand right what you mean, when it to seems as if I gave the Englishhorn here to much dominance so I will try to bring it back to the ensemble.
4) III12'12 Again a passage I already did not felt so comfortable with the way Draeseke shortens the dominating Cellomotives in this development. But you are right it deserves to try perhaps with other articulations to make it sound more convincing.
5) III 13'17 (or 13'27) At 13'17 I hear a enhousiastic Melody, so what you said seemed to me more likly for 13'27 here Portato Orchesterchords are composed I am also not that much satisfied with my render here, but your critic encourages me to try perhaps other articulations here again.
6) III 14'17 Yes of course the Timpani is much to loud here. I confess here I looked to much at the score and listend not enough how plausible the ensemble sounds after.
All in all that is really a great helpful and inspiring posting, thank you so much for spending your time and diligent attention. I will just wait and collect as much hints like that and will work on the recordings next week when the Woman from the Draeseke-Society will join to discuss and realize all necessary improvements.
best Steffen
Hi William,
First excuse me that I had no time yet for a serious answer on your very much more helpful answer you posted on this thread in the last days, since another project absorbes me curently. I hoped I would be able to answer with a new version of the Frithiof, but the other Project kept me away from it. As Tom also indicated, I should definitly work on the Violins for the first 30 seconds and sure I will. You are absolutly right, that in the way they are written, they tend to loose volume of the sound as more they climb up which makes the Crescendo not very convincing. I think I will try to ad perhaps other Violin samples like Chamber or an additional Solostring in order to give the Violins more substance in the development of the beginning.
I still am not sure if I understood your remark about the horns right, they should play a phrase with repetive triples and sustain notes which is done with repetition-performance and sustain from the basic Articulation set, when I remind it right. I have increased the Velocity oft the three triplets towards the next quarternote. What sounds perhaps a bit like a rhythmic accompagnement has also motivic meaning which comes again in many situations of the symphonic poem. Perhaps could you try once more to explain what you feel uncomfortable with their sound. I hope I remembered the most significant points you indicated, if not please feel free for more advice.
I definitly consent, like you I would and did never programm a vsl-rendition in a notation software, since I am to much used to work on the mididata in the keyeditor and its possibilities to programm the different midichannelcontrolerdata of a track. I use (Smartscore and) Finale only to prepare a clean and neutral midifile for Cubase.
But no, what you currently hear on my site is still the first Version, meanwhile this was already completely edited in CubaseI.
Sorry but I havent found the time yet to work on the Ideas of improvement I collected already for a second version to come, what you hear on my site is still the first version I uploaded on my site. I ill keep posting here, if I managed to work on the different proposals for improvements. in a second version.
You wonder why I invest that much labor for a piece of another composer? The short answer would be: "I like it." But sure I understand, that from your point of view vsl is very helpful to realize your own compositions. Since I all my life hesitated to seriously compose music (meanwhile i had really some fun whenever I tried it) vsl is for me likewise interesting to make hidden treasures of our musical history audible, and there are still some great works to de discovered.
best
Steffen
I am going to listen to more of this and be more specific as I think it is a very good project. ALso, I was unaware of this composer and so it is good to learn of him. I completely understand your desire to do a masterful piece like this in MIDI as it is a good way to study a score very intensively.
Aside from having the time to persue such an astounding project; I just thought i would give credit where credit is due. I also want to thank you for exposing this great composer that over time, have fallen through the crack. He lives. He lives after all. What enormously, tremendous and trajic music. Cannot be denied.. you took upon all of this while teaching. This is an enormous task that is overbearing to any other. You also have so much more material posted, that you have accomplished with your virtous piano skills. Congratulations, your students are lucky to have a truly experienced virtous pianist to learn from. That is a rare entitelment as well.
I thought of something William said a while back about the MIR; That all you have to do is set it and it will mix everything for you. This might be of help to a great player as yourself, where time is of a restrained occupency. You've got William chiming, This is also helpful as well... Will be looking forwards to the final mix. As well as visit your sight. Wish there were more of this serious material out there. I would like to thank you again, for bringing those that have fallen through the cracks over time, out in the open and to the public.
I agree with that and am wondering specifically to help me in what I'm going to suggest -
is this MIDI recorded in such a way that each instrument is on one track? Not splitting into two or more players, but every individual player's part on one track? If so the revision that I can suggest after I listen to all of it will be greatly facillitated.
I discussed today with the woman from the IDG some by internetusers proposed improvements, which we tried to aply. So some improvements are done, but still I am sure there might be enough to think about.
We mainly worked on the beginning giving the accompagning instruments a wider crescendo but still hesitated to let them really start in piano, since that is not meant by the "forte" in all instruments in the manuscript, We also corected the Piccolo for better support the viollins in the higher regions and changed finally some articulations of the violins in the end of the start-development But we also fixed several things in the second and third movement someone suggested.
@ couch potato (first let me thank you for your kind remarks about my website) What I still of course definitly have to do but did not found the time yet, is the acustic especially of the winds. One Mail proposed me, to change the predelay of the winds in order to let them apeare not so near to the listener. I konw "MIR" would do the job much easier for me and I do intend to upgrade my hard and software to work with it in the near future, But the Idea to change the predelay in the convolution reverb of the Vienna suite sounds interesting enough to give it a try (as soon I' find some time for).
@william: sure each track is one "voice" (player/section), but still there are for instance passages, where first Violins play two notes together. But winds are always splitt in two, three or four tracks depending on how much player a section should have (f.i. piccolo, flute 1, flute 2 makes three tracks).
Still I am very interested for any further hint that might improve my rendition.
best
Steffen
I wanted to say this is an important project in general, to bring to light this fine composer. I am amazed by how this great piece is not already in the orchestral repertoire. Also, this is one of the things that is so good about the VSL Forum - one actually encounters things here that you simply don't find elsewhere.
I cannot do too much time on this but am listening when I can. ALso, since it is all in separate MIDI tracks that is great, because any changes can be made very practically. There are some things that can be done in general, divorced from any specific timings, and I can give those right now.
One thing I noticed is the overall mix sound is very dry, which increases the artificial effect. As was pointed out, using MIR if you can do so would be a great thing since it allows you to do fantastic sounding mixes without having to do a huge amount of tweaking. MIR tends to make almost any sound very realistic, with no effort, and so all the effort spent on the huge amont of music in this could benefit from that.
Also, I feel strongly that there are too few articulations used in general. This is of course the most basic aspect of doing a MIDI performance - selecting articulations. It is how one creates the basic sense of realism as well as expressiveness. And so, it is up to the performer to decide exactly what he likes. This involves a HUGE amount of work, because of the vast number of possibilites that VSL has created. You could take one line of music and do it literally a hundred different ways with all the articulations that are available.
So that is why I was originally say, it is very important to take this piece and listen to EACH TRACK SOLO. And then ask oneself, is that one starting note just a sustain, or is it a sforzando - and even more, the particular sforzando that has been recorded for this instrument. All of them vary, based upon the performance practices, and how they were recorded, the player who recorded it, etc. So this is where AUDITIONING the samples available is so crucial. What I am saying here is that there is no magic instant solution for this aspect - it is simply hard work figuring out exactly what is the best selection for the articulations of each line. But you can do that, because you very correctly have each instrument separate which is huge for making this possible.
There is one general fix for instruments sounding too "computerized" for lack of a better term. On a piece like this, you can take each section - for example, the flute section, which might consist of three players. On first, you can keep it at quantized value. On second and third you should apply different humanize presets of anywhere from 10 to 25 % randomness. Then, listen to each track separately, and see if that humanize resulted in any problems, such as overlapping notes, cut off of note ends, or objectionable rhythms, etc. If each one sounds o.k., then you can route each of those tracks to separate instrument sets. Such as 1st flute, 2nd flute, piccolo, or whatever you are using. If you have to do some transpose/pitch-shifting, which you might on clarinet or others depending on what collection you are using, then each one of those would be an individual instrument. The point is, to make each track totally separate from the others, and make is sound good ON ITS OWN. If it does, then when they are blended, with the humanize effect combined with one quantized track, it will sound very natural.
One thing on this use of one quantized track and other humanized tracks, it is important to determine which ones you apply these to. For example, I have often used a very strictly accurate quantized (string) bass part, since it underlies the entire orchestra and forms a rhythmic basis for everything else. And also - importantly - it does not exactly double any other instrument. That is important because, if you apply quantize (or leave the notation file intact without humanize changes) on instruments across orchestral groups such as flute 1, violins 1 oboe 1, even trumpet 1 if it happens to be playing - if these instruments which often play in unison in a piece in a tutti are all quantized, then again you will start having that artificial sounding "too perfect" rhythm that instantly makes the ensemble sound fake. So it is good to use the humanize on each instrument that would likely be playing the same rhythm, even across sections thoughout the orchestra, and only use quantize on those that will be playing very different parts. Anyway, that is a general principle you could apply.
Those are some few general things I thought of that could actually accomplish a lot in making this more realistic, but are - I admit - based upon doing a fair amount of work in auditioning each track separately and applying the appropriate articulations. That is the main thing. It is not at all strange to have a change of articulation on EVERY SINGLE NOTE of a line, because even though many might seem generally applicable - such as "sustain" or "legato" - often in actual orchestral practice there is a huge amount of variation in articulations the player is automatically applying based upon his years of mastery of his own instrument. Also, dynamics - a composer will mark an entire section "f" or "pp" but if you acoustically analyze the individual part performance, you will hear constant crescndi, diminuendo, accent, etc. Becuase it is the musical phrasing implied by the content of the line. So the great difficulty one faces is in trying to match that vast amount of variability in all the lines of the orchestra. Something which I believe actually dwarfs what a conductor has to do, who relies upon his players doing all that, automatically, for him. Someone doing MIDI does not have that luxury.
I realize you probably know this, being a good musician, but I am stating it to give an idea of what I think needs more work.
A few other things are - what sample set are you using? On this piece, you need the entire symphonic cube because it is a heavy-duty piece of music. Also, do you have this in a standard MIDI file and are you using Vienna Ensemble Pro or what? Also, are you using the VSL presets for instruments, or did you create your own custom presets? ARe they selected by keyswitches? I can't do much more unless I had the actual MIDI file in front of me with the VE file, and then I could make further suggestions by examining exactlywhat the keyswitches are doing, as well as the controllers.
However, whatever the case this is a great project and I do admire your undertaking it in a serious way!
Hi William,
First let me thank you for your detailed response in my thread and let answer on some of your questions:
- You can expect nearly all and as extended as ever possible what vsl has ever produced (except "MIR", the dimension brass and the choir - I just have those womanchoirssamples they had once as a part of it as Christmas give away for Cube-owners before they sold the choir) I intend to upgrade to MIR in the near future and agree, that this will help me to improve the overall acoustical realism.
- I dont work with Keyswitches anymore but started with which expressionmaps I created when I was editing the music, so I did not want to have a kind of standard articulation set for all, automaticly applied but intended more a set that depends on the need of each track. But it turned out to be possible and useful to keep the articulations of Violins (I+II), Violas and Celli similar.Exept tremolo and staccato I avoided graphical articulation signs but preferred texts to be forced to decide each articulation conciously by myself (for me there is even no general standard staccato, marcato, legato etc. but having the extraordinary variety of vsl-Samples I want to chose the best articulationsample for every single note played.
After I experimented in the first Movement with the "Spielweisen"(I dont know the english term) of the expressionmaps which are not bound on the single event but act more like keyswitches for the following events, I decided to work only with "Attributs" in the second and thrid movement, that has the practical reason, that I dont need to open the articulation controlerchannel in the keyeditor which absorbs especially for the strings with their 30-articulation expressionmaps nearly the whole screen, but can see and control the articulation of each single event via the information line above the keyeditor.
- Meanwhile It sounds quite interesting, how you try to humanize trackwise, it differs in the way I try to become musiaclly convincing. First I dont like and dont want any automaticly humanization, since anything I have heard like that (for instance what Finale does) is in my eyes just a bad joke and not at all what a musician really could only do depending to his understanding what he plays. As you can see (in the second screenshot I posted in the other thread) at the very detailed editited tempotrack of the project, Humanization is for me (who would be surprised for a pianist) very much a question of breathing often very subtle rhythmical changes. I must confess I was that much disappointed by the automatical "humanization" in Finale, that I was never interested, if and what cubase might provide in this area. I dont know if I understood you right in this point: Did you use something like a humanisation funcion for single tracks in Cubase with more acceptable results?
Anyway, it is probably true, that I should think about more how to avoid to artificial parallelism of different tracks. Your concept seem to me of course at least very consequent if not to say rigid, but in order to stay efficient with my timemanagement for this project my way to work on editing a certain passage just always started with looking for the main/dominating musical idea/intention, trying to make it audible as good as possible and shape all occuring musical events in reasonable relation to that. That means beside shaping the timetrack in an appropriate way, to get the right articulationsample and the right dynamical development and proportion between for all notes in the different tracks.
Still I am sure I also did not in every passage answered those fundamental questions in the best possible way I found for instance for some passages already mentioned by others, that I could not only fix the appearance with other articulations but also with additing soloinstruments or chamberstringsections to section-tracks that did not sound convincing enough before.That might be in some cases also a solution to avoid the paralelism you critizised.
However, since I found most of passages already named here and from other listeners worth to improve, I will be happy for every further hint, on a concrete less convincing passage or perhaps even tipps, what might be improved at this certain passage. How about for instance with the work we (together with the woman of the IDG) invested in the most critizised beginning at least you currently already can hear a first improved version.
best
Steffen
I still have not had the chance to listen to the piece (but still intend to do so), nor am I in William's league, but still wanted to reply to your comment about humanization. If you wish to do so, VSL's humanization can be fully controlled via any controller you wish to use. If needed, both parameters can be controlled individually. Thus, more intonation "problems" can be created for extended fast passages, and fewer, for slower legato ones.
Another helpful option is instead of relying solely on VSL's humanization presets, create your own. If you create - say 40 or so fast in tune presets - you can assign a different one to each articulation with very little overlap, and have a great variety of presets for each instrument/section. Doing so ensures that the presets are different for each instrument/section for the same articulation. Having different humanization presets on marcato for violin 1, violin 2, etc., works, as opposed to using the same preset on each, as no two players make identical errors.
Also, as William has said, frequent changes in articulations is critical. Within Finale, one can create all the keyswitches on a separate staff for each instrument, but only display those staves in the piano roll view, not the actual score.
Hi Steffen,
after listening to a small portion of the 1st mvt, I' ve formed the following opinion:
there is no sense in trying to improve anything "performance"- or "musicwise" in this production until two major issues have been solved:
a) a convincing and well proportioned spatial disposition for the whole orchestra has been set up
b) at least roughly balanced eqing of instruments & sections in order to loose the strongly "artificial" spectral composition of the mix in its current state as well as to get a better (meaning more natural) sound of individual instruments
Only after this "technical" aspects have been solved properly can any further work on "interpretational/musical" improvement be done in a meaningful way - at least in my opinion.
BTW - congratulations on the stupendous production of Fallas Alborada - it has long been my favorite EW Pianos demo track ;-)
P Paul moved this topic from Orchestration & Composition on