Freiberg - G. Silbermann Organ Model progress

[zurek]

I feel that I owe some information on the progress of the work on the Freiberg Silbermann Organ Model. As I mentioned earlier, I had to dedicate my time in the last 2 weeks to work other than Hauptwerk. Neverthless, I did not stop working on the Freiberg entirely. First of all, I had to wait for a new agreement with the Freiberg authorities. Now, the new document has been signed and so I can start distributing the sample set. The owner of the organ, adviced by the organbuilder, required that the sample set is encrypted to prevent misuse by third hand. For this reason, I will distribute the sample set encrypted. Therefore, it will be usable only in conjunction with full version HAUPTWERK (with dongle), no other sampler will be able to read the files. Also, the demo version of Hauptwerk will not be able to load the sample set. This is an exception among the Sonus Paradisi sample sets which are normally distributed in plain wave format to allow for user editing. However, the wish of the organ owner is always the primary concern of us.

On the occasion of singing the document, we visited the Freiberg again. And being there, I could not resist the temptation to arrange a new recording session and record the short staccato release samples again to be sure that the organ model will have the right behavior with short releases. And I finally discovered, what was the problem with the releases. The tracker action of this particular instrument is quite a tricky one and it in fact does not allow to record very short key strokes (there are some "lazy" opening pallets, some late speaking pipes - especially the Fagot is very late, the tracker is quite sensitive and loud when quick tones are played, giving a series of tiny noise "ticks" when released). Of course, this is no obstacle when normal musical performance or normal recording is done on the instrument, but it is a big problem in our case. Therefore, on our first occasion, we opted for recording a bit "longer" staccato tones. These were good as the "portamento" releases, but not as good as the "staccato" releases. So, this was the problem. With great difficulty, we tried to record the shortest staccato possible on the instrument. Then, I spent endless hours denoising these samples removing the tracker action noise and other noise which the tracker produces when playing staccato. But it is done now. I hope to publish some short demo pieces with these new releases. At the end, the new visit of Freiberg was very beneficial, since now the sample set will have THREE levels of releases, adding a FOURTH level for the manual reed stops (for a good measure:-). I decided not to discard the originally recorded longer staccato, as these perform very well as the middle-length release, between the short staccato and long one. After this addition, Freiberg organ model will have now short staccato, longer staccato (portamento) and long steady tone release for each sample. The fourth level of the releases for the manual reeds will be a second short staccato release, and Hauptwerk will select one release at random when the key is released. In this way, the variability of the releases is even increased for reeds where this issue is heard the most.

I hope that I will be able to finish the preparation of the sample set during the weekend (the remaining work is that I have to prepare the encrypted installation media and to test them), and I hope to release the sample set during the next week (finally!). I have to prepare also a demo version of the sample set, so that interested users can have a demo before they decide to purchase the full version. I will not discard the demo pieces from my web pages for now, but these were recorded using the previous version of the Freiberg virtual organ model. I hope to add new demos later during the month to come.

[www.orgeljournal.de]

"... At the end, the new visit of Freiberg was very beneficial, since now the sample set will have THREE levels of releases, adding a FOURTH level for the manual reed stops (for a good measure:-). I decided not to discard the originally recorded longer staccato, as these perform very well as the middle-length release, between the short staccato and long one. After this addition, Freiberg organ model will have now short staccato, longer staccato (portamento) and long steady tone release for each sample. The fourth level of the releases for the manual reeds will be a second short staccato release, and Hauptwerk will select one release at random when the key is released. In this way, the variability of the releases is even increased for reeds where this issue is heard the most. ..."



Now I suppose this set will be a very attractive one. I'm curious.

http://www.bildergalerien.prinzengarde- ... 1982_3.jpg

Best regards,
Matthias


__________________________________



Hauptwerk-News for German readers
www.Orgeljournal.de
__________________________________

[prinzipal]

Dear Jiri,
I am also very happy that this problems with the staccato releases will now certainly be solved. Honestly I also have to say that I am happy about my ears, which were able to hear exactly this staccato problems.
Thank you also very much about your honest explanation, why this problems occurred and how you solved them , that is not self-evident.
The last months I had also some thoughts, that it has to be very difficult to record staccato releases without to get the resulting tracker noise into the recording - or how to denoise this recordings.

Best regards,
Johan

[bcollins]

The owner of the organ, adviced by the organbuilder, required that the sample set is encrypted to prevent misuse by third hand. For this reason, I will distribute the sample set encrypted.

This is very disappointing news for me, as I was looking forward to buying this [dry] sample set.

Oh well.

[deWaverley]

.
Could anyone explain what sort of "misuse" the 'organbuilder' (just how old is Herr Silbermann these days, anyway?!) is so afraid of, and why it could possibly matter to anybody?

I'm confused.

deW.

[Sander]

As Silbermann unfortunately passed away, somebody else has to take care of this instrument. I guess this person adviced the owner of the organ for the possibility of misuse. They gave permission for a recording for historical purposes and the possibility to make a Hauptwerk model. They didn't gave permission to make a Gigabyte or Kontakt set or to have it used by Allen or Johannus. Hence the encryption is needed.

If I understand correctly, if the set is encrypted you can still make full use of all Hauptwerk functions like CODM and tuning. However you can't export the samples for use with other programs. It's for use with Hauptwerk (2.0/3.0) only.

[Lougheed]

[quote="Sander"]
If I understand correctly, if the set is encrypted you can still make full use of all Hauptwerk functions like CODM and tuning.

Yes, but sadly not the all important public performance licence. That is a huge item lacking in a number of the sample sets, and does not exist in the "real" sample world, apart from some restrictions in production use for movie trailers (also known as previews).

I just don't get it. I can buy ANY set of samples and make music and release as a CD, except for MANY of the Hauptwerk sample sets - although not all of them. (And an additional sad fact is that the sample sets that I really want to use for public performance, are the ones that will not allow for this use. This really needs to be addressed.

Lawrence

[bcollins]

Some things one cannot do with an encrypted sample:

Fine tune the release tails to better fit the natural acoustic space the pipe/rank is being used in by clipping and shaping the release(s) in a practice that is counter to the HW method (whereby lower notes are clipped longer and higher notes are clipped shorter). This is good in very reverberant spaces but very unnatural sounding in large, but relatively dry spaces.

Load the .wav file into editors such as Sony Soundforge to shape (improve) the attack of slow speaking pipes.

Create new ranks from existing ones using the Pipetune software. This can be done in the CODM, but with much more thought and work.

Easily create more uniform ranks, by discarding "sour" pipes and replacing them with neighboring pipe samples. And increasing or decreasing the output volume of individual pipes samples prior to loading the organ.

[adri]

Some things one cannot do with an encrypted sample:

Fine tune the release tails to better fit the natural acoustic space the pipe/rank is being used in by clipping and shaping the release(s) in a practice that is counter to the HW method (whereby lower notes are clipped longer and higher notes are clipped shorter). This is good in very reverberant spaces but very unnatural sounding in large, but relatively dry spaces.

Load the .wav file into editors such as Sony Soundforge to shape (improve) the attack of slow speaking pipes.

Create new ranks from existing ones using the Pipetune software. This can be done in the CODM, but with much more thought and work.

Easily create more uniform ranks, by discarding "sour" pipes and replacing them with neighboring pipe samples. And increasing or decreasing the output volume of individual pipes samples prior to loading the organ.

The entire idea of playing a Silbermann is that I will be playing a Silbermann, and not a remake or re-voicing of the same organ for different and unintended purposes, which IMHO cannot be done convincingly for reasons I have outlined in other earlier posts. It is exactly in accepting the supposed and perceived limitations of an organ that you start to discover its real musical strengths, and also start to learn more about history and how things were done back then. I like these forays into the past to come to me as best and unaltered as they can.

[bcollins]

The entire idea of playing a Silbermann is that I will be playing a Silbermann, and not a remake or re-voicing of the same organ for different and unintended purposes

This is a valid (perhaps even prevalent), but different school of thought. The beauty of the Hauptwerk system is that it reaches out to all sorts of organ enthusiasts; organists, historians, "builders". Sometimes with extreme, but opposing points of view.

For an organist, an organ historian, the [wet] sample set - well documented - is a chance to appreciate an instrument of importance, a musical masterpiece, as close as possible to its original state, without having to travel or seek out special permissions. I can appreciate that.

But for my purposes (and not being an organist) a well recorded [dry] sample set is little more than crates of pipes, albeit crafted by a master, waiting to be assembled and voiced into something new - and hopefully fresh and wonderful.

Perhaps this is distasteful to the purists and maybe what the stewards of this great instrument mean by "misuse". But that's what gets me excited.

[telemanr]

I do wonder if Bach, for instance, would have put his foot down if a better wind supply system had become available and could be installed in an organ which had a noticeable unsteady system. Would he have said he thought it was charming just the way it was? I think not.
And a pipe or two which were noticeably slow to speak but could be improved...would he have said no to that too?
And do you leave a pipe which is obviously out of tune alone? Does this contribute to the experience? Just because it happened to be out of tune when the organ was sampled doesn't mean that Bach or anyone else who played it, would have been satisfied with it. After all, we do have organs tuned don't we? It's a bit like Glenn Gould and his humming. You could ignore it in a live performance since you only heard once. But on a recording it annoys you every single time you play the record.

And I really don't understand this love of noticeably unsteady wind. There is enough variation in the sound of an organ given the general physics of the whole system without worshipping something that the original players would have loved to have fixed if they could, I'm sure.
A little unsteadiness goes a long, long way.

Rob

[adri]

I do wonder if Bach, for instance, would have put his foot down if a better wind supply system had become available and could be installed in an organ which had a noticeable unsteady system. Would he have said he thought it was charming just the way it was? I think not.
And a pipe or two which were noticeably slow to speak but could be improved...would he have said no to that too?
And do you leave a pipe which is obviously out of tune alone? Does this contribute to the experience? Just because it happened to be out of tune when the organ was sampled doesn't mean that Bach or anyone else who played it, would have been satisfied with it. After all, we do have organs tuned don't we? It's a bit like Glenn Gould and his humming. You could ignore it in a live performance since you only heard once. But on a recording it annoys you every single time you play the record.

And I really don't understand this love of noticeably unsteady wind. There is enough variation in the sound of an organ given the general physics of the whole system without worshipping something that the original players would have loved to have fixed if they could, I'm sure.
A little unsteadiness goes a long, long way.

Rob

Yes, I believe that it is the responsibility of the sample set producer to make corrections in speech and tuning to the best of his ability. Having no wind model at all, as I live in the US, I must say that I sorely miss it. Yes, it can be easily over-applied, but totally steady wind in especially a baroque organ definitely doesn't sound totally satisfying. E.g. the Spanish Frechilla organ would greatly benefit from wind presence. But even a large C-Coll organ will have undulation during large chords in a plenum registration.

We know that Bach liked ample wind supply, but remember the bellows in his days were still wegde bellows, and one must not confuse their effect with the later reservoir-magazine bellows. Their effect upon the speech of pipes and specially the plenum is quite different.

[telemanr]

Adri you are making it sound like Bach only wished to have a better wind supply than wedge bellows could supply. For all we know he would have approved of any improvements. Even concussion bellows and any other modern improvements. I can't imagine players of that day complaining if a wind supply became as stable as current supplies are. I seriously doubt any of them would say it ruined any musical effect they were looking for.

That said, I think we want the variation today in those samples only because then it sounds exactly like that real organ, and not because it is in any sense musically important.

[jfziehlke]

I can't imagine players of that day complaining if a wind supply became as stable as current supplies are. I seriously doubt any of them would say it ruined any musical effect they were looking for.

Did you ever see one of those old Austins with the walk-in wind chests? Their wind supply was so stable you could not shake it with a tremulant! They had to fit a large fan-like contraption above the pipes to produce this effect.

Jerry

[zurek]

I can see that there are several issues discussed here:
first, the use of the encryption. As you can guess, the necessity to use the encryption was somewhat a disappointment for me. As you can read on my internet pages, I love the freedom of users to edit and possibly alter the sound files and therefore, I normally deliver my sample sets in plain wave format. So far, all the sample set of my offer are without any encryption. The Freiberg will be the first sample set using the encryption.
However, I had to subject to this wish of the owner, since otherwise I would not be able to create this sample set at all. The Freiberg authorities simply strictly required the encryption and it is written in the agreement which I signed with them. Without this, there would be no possibility to record the instrument. As Sander rightly pointed out, they allowed the use of the audio material only with Hauptwerk where it can be well protected against unauthorised use.

Believe me, the encryption is not so much a protection of me as the seller of the sample set, but it is the protection of the audio data to make the use by the third hand impossible outside Hauptwerk. This was the wish of the parish and of course, I respect it. The most important duty of the Sonus Paradisi is to be RELIABLE partner to the owners of organs, so that we can get the permission to record other instruments. Therefore, their wishes are the supreme law for us.

Now, there is the issue of the archival/documentary recording in opposition to the digital organ school. Here, Sonus Paradisi is firmly and deliberately rooted in the archival/documentary school. Anyone is free to do the job in different way, but for us - in the Academy of Sciences of Prague - the most important aspect of the documentation is to capture how the instrument sounds NOW, at the exact date of the recording. Having more recordings of the instrument (for example, the Tyn instrument has been recorded at least 5 times in the course of the last 20 years), you can very precisely track the "life" of the instrument in the course of the time. As you can read on my internet pages, the "practical" aspect of usability of such an organ model which reproduces also the imperfections of the original instrument is only secondary! Of course, I have users, who told me: "I will not buy the sample set because the low A# in the Bombard of the Forcalquier instrument sounds different than the rest of the rank". It is perfectly OK, but I think that I also have the freedom to produce the sample sets in this documentary way and I do not feel obliged by the market to make the instrument to sound more "perfect" than in reality. If you do not like our documentary approach, you can buy other sample sets from other producers who usually (according to my experience) opt for more even sound of the organ. I leave the sound "rough" to large extent.

Of course, there must be responsible judgement of what the error is in a way "characteristic" to the instrument in the given time, and what is simple casual error. These I do not leave in the sample set This is about 1-2 samples per sample set. In the case of Freiberg Organ Model, I remember to recreate only one pipe in the low octave of the Vox Humana, which was indeed absolutely unstable. Then, there was the middle F in the Principal 16 of the Hauptwerk, where the pipe produced intense resonances of the organ case (most probably the pipe is wrongly seated on the windchest). So, I repitched from the neighbour also this pipe. As far as I remember, all the other pipes are original, including the late speaking Fagot. The late speech is the very characteristics of this stop and therefore there is no intention to change this.

As far as the tuning of the organ is concerned, I LEAVE it as it is when recorded, so that the documentary purposes are matched. The information about the tuning of individual pipes is simply an information which is good to preserve in the sample set. The beauty of Hauptwerk is, that you can have this information present in the sample set, but if the user wants more in tune output, he can simply switch on any temperament chart and suddenly he gets what he wants! If we had hard repitched the pipes to be in perfect tune, the result would be the same, but the information about the original tuning of individual pipes would be lost. The preservation of the documentary information is also one of the primarily concerns of Sonus Paradisi, which takes precedence over the practical aspects.
Therefore, using the ORIGINAL ORGAN TEMPERAMENT in Hauptwerk is good only for these documentary purposes, since usually you get the organ slightly mistuned. However, the experienced user of Hauptwerk knows, that there are many other temperament charts available, and using these has one big advantage: Hauptwerk will repitch automatically the samples to sound in tune (except the random tuning error which may be turned off in the Hauptwerk general settings). Also, Sonus Paradisi almost always includes a proprietary temperament charts with the sample set, which recreate the original tuning of the organ with precision, but without the casual mistuned pipes. So, if you want to listen to Freiberg original Neidhardt II temperament, it is not good idea to turn the Original Organ Tuning on, but you should install and use the proprietary Freiberg St. Petri Neidhardt II temperament chart instead! Then, if you still do not like the random mistuned pipes by Hauptwerk, simply turn off the random tuning error in the General settings of Hauptwerk.

In the case of Freiberg, there is one more interesting aspect: the latest restoration of the instrument decided not to return to the very original temperament of the instrument (meantone) but to apply deliberately different temperament more suitable for these times (Neidhardt II for the small city was chosen). Therefore, we recreated the original meantone temperament digitally in a form of a temperament chart which is distributed with the sample set and therefore, whoever wants to listen to the temperament of the instrument as Silbermann made it originally, he has the chance to do this using the Hauptwerk Organ Model! Again, this is one of the marvelous functions of Hauptwerk.

There was also an issue about the wind model. Again, what is important, is the preservation of the information. It is absolutely no problem to create absolutely stable wind model in Hautpwerk (in fact, it is the most easy thing at all) but the same effect is achieved when the user simply switches the wind model off. It is very easy as well. On the other hand, I am extremely proud of my wind model using the unstable wedge bellows and invisible calcants. To tune this model to react in the way real organ does took several months of my life and several important organ experts were involved in this process (including for example Mr. Henk Kooikher from Netherlands, who takes care of the restoration process of the Zwolle organ). I am very sure that the model as it is now, is pretty good. Indeed, it is quite instable, but again, it is what the organ on old wind produces and it is the brilliance of the author of Hauptwerk that he build the wind model so that it can reproduce the behavior in such a convincing way! It is again an added information which can be preserved in Hauptwerk compared to other software. However, if you still do not like this information, you can simply get rid of it by switching the wind model off or by lessening its amount in the Hauptwerk settings.

Last, there is the question about the versatility of the sample set. Indeed, no creativity of the users is lost using the encrypted version, since I left all the voicing features enabled in the ODF. So, you loose nothing compared to the unencrypted samples. You can freely revoice the instrument using all the Hauptwerk functions, you can even combine the dry ranks of the instrument with other ones since they are referrenced in the installation package so that CODM can make use of them even if they are encrypted. The only restriction which comes with the encryption is that the samples are not usable outside Hauptwerk and they cannot be altered by the editing in conventional audio editors. If you need this functionality, I can still offer to you many different sample sets which are offered in the plain wave format.