It is currently Fri Mar 29, 2024 2:06 am


When will we reach this level?

Existing and forthcoming Hauptwerk instruments, recommendations, ...
  • Author
  • Message
Offline
User avatar

Jan Loosman

Member

  • Posts: 380
  • Joined: Sat Dec 20, 2008 4:33 pm
  • Location: The Hague, Netherlands

Re: When will we reach this level?

PostWed Jul 20, 2022 5:49 am

I don’t know if they do this already?
But if i was a sampleset producer i would make the recording of each pipe in each perspective to be recorded simultaniously with each note pressed so avoiding frequency shift due to temperature changes when diiferent perspectives would be recorded at different times.

Further more i would experiment placing the mics for the 3 front perspectives at exactly the same position in the church.
But i would use for the dry perspective highly directional mics thus filtering out the surrounding acoustcs.
For the mid front perspective i would use cardiod mics(adding acoustics) and for the distant front perspective i would use omnidirectional mics.(most acoustics)
Thus preserving the exact timing of each pipe in each perspective but with each sort off mic adding more or less direct sound vs acoustics. This way ther would be no Phase shift /timing problems between the different perspectives
Since i have no experience producing sampesets i don’t know if this is doable, but i would experiment with such a setup.

Jan
Last edited by Jan Loosman on Wed Jul 20, 2022 7:43 am, edited 3 times in total.
Offline

organsRgreat

Member

  • Posts: 561
  • Joined: Fri Nov 16, 2012 5:30 am
  • Location: England

Re: When will we reach this level?

PostWed Jul 20, 2022 6:34 am

There are a number of simple reasons why an electronically derived organ is unlikely to sound as good as one with the equivalent number of pipes. The first is intermodulation distortion in the loudspeakers, as explained by Dr. Colin Pykett:

https://www.colinpykett.org.uk/EndOfPip ... #Technical

As Colin points out, the cumulative effect of this distortion is a perceptible level of background noise, and a general veiling of the sound. I am fortunate in owning a pair of electrostatic speakers, which have exceptionally low levels of distortion (Quad ESL57s); but even so, whenever I hear a pipe organ, or a symphony orchestra, I’m amazed how natural and relaxed the sound is. Unless we have enough speakers for each pipe sample to sound through just one speaker each, we cannot hope to achieve the clarity and richness of a pipe organ. I believe the late Leo Christopherson used eighty speakers – exceptional for a home installation. His playing ability was limited, so he used MIDI files, very carefully registered. His recordings are still available on Contrebombarde:

https://www.contrebombarde.com/concerthall/user/82

and on Youtube:

https://www.youtube.com/user/leo037chris

Leo’s organ recordings do sound particularly natural and musical to me, despite the negative effects of digital compression.

Secondly - Colin Pykett did not seek to address the radiation pattern of most speakers, which are quite directional, whereas organ pipes function more as an omnidirectional point source, radiating sound in all directions. This helps the sound to blend and create a more interesting result. One reason why an orchestra generated electronically in a computer, sounds artificial when compared with the real thing, is that sounds mixed electronically do not sound as good as when they mix in the air. Not only do we need a LOT of speakers to approach the sound of a pipe organ; we need them to be omnidirectional, and such speakers are expensive. We then need more sound card outputs to drive those speakers, and sound cards with multiple outputs are also expensive. As Colin points out in his article, providing enough speaker channels to make the sounds as independent as pipes are, costs about as much as a pipe organ, so why not build a pipe organ in the first place?

Thirdly, have you ever heard a pipe organ in a small space? I have, on two occasions. During the 1960s I heard a Wurlitzer theatre organ, expertly installed in a house in a London suburb. I have no reason to doubt that the pipes were as well voiced and tuned as they had been the cinema; but it was obvious that we were much too close to the sound. UK cinemas typically seated two thousand people, so organs were voiced accordingly – for brightness as well as power. Twenty years later I attended an Open Day at the English Organ School in Dorset:

http://www.margaretphillips.org.uk/eos.htm

Concert organist Margaret Phillips has assembled seven small to medium-size organs in a former Congregational Chapel (much larger than a home!), and kindly demonstrated them. Because there was a larger volume of air the result was more satisfactory, but a larger church would provide more reverberation and space for more ranks of pipes. If a pipe organ sounds disappointing in a home, when compared to what we hear in a church or concert hall, it is not surprising that Hauptwerk is also less interesting - though Hauptwerk has the advantage of being able to add reverberation.

Finally, we need to be constantly aware that both Youtube and the Contrebombarde site use compression – Youtube uses aac and Contrebombarde uses mp3. The reduction in file size is useful, but compression inevitably degrades the sound.

Perhaps a more useful question would be “Is a good digital organ, installed in a home, a valid musical instrument?”. Whenever I play my grand piano after a session with Hauptwerk, I’m aware that the piano is musically expressive in a way Hauptwerk is not. My Hauptwerk installation uses twenty speakers – I’d like to double the number, but twenty is what I can afford and find space for. I use Hauptwerk mainly for its theatre organs, so perhaps I’m not comparing like with like. What I can say is that Hauptwerk is a lot of FUN! Linking it to a Digital Audio Workstation such as Reaper makes it possible to add sounds such as a realistic piano – theatre organs often had one, and Pianoteq works well for me. Theatre organs in the UK typically had eight ranks, whereas Hauptwerk sample sets can easily provide 30 to 40, thus including every family of sound ever included on a pipe theatre organ. For someone interested in arranging light and popular music, a Hauptwerk theatre organ, with a few additions, provides endless interesting opportunities.

It is worth remembering that quite early in the history of electronic organs, a number of popular organists established careers playing them. In the UK Jerry Allen and Harry Stoneham were outstanding, and there were similarly talented players in other countries. The overall sound quality of a drawbar Hammond or a Lowrey Holiday did not approach the digital instruments we have today, but these musicians were highly skilled at producing interesting arrangements from limited resources.

https://www.youtube.com/watch?v=A41myys6PIY
Offline

josq

Member

  • Posts: 912
  • Joined: Tue Jan 24, 2012 12:11 pm

Re: When will we reach this level?

PostWed Jul 20, 2022 8:39 am

I wonder if it is a correct theory that we need many speakers because pipe organs have many pipes (and because we need to reduce distortion).

The same theory would apply to orchestral music: we would need a separate channel for each instrument. No one is buying that. Why? If the distortion is so bad, then it would be bad too for orchestral recordings? And if 2 speakers are insufficient to represent all individual instruments in an orchestra, surely there would be a significant market for large speaker arrays?

I think there is a subtle but important misunderstanding about the goal of music reproduction. The goal is (usually) not to reproduce the orchestra or organ itself in our living room - that would not fit indeed. The goal is to reproduce the sound of an orchestra or organ, as it sounded (and has been recorded) in the original room, including the original acoustics.

And maybe (well, very probably), a simple stereo setup goes a long way in achieving that goal. And if you want more, there is a huge market for surround systems (as opposed to large speaker arrays), because the surround channels contribute to a sense of envelopment, or possibly (as Jan suggests) to add a height dimensions to the sound.

I am saying this for a reason: if people don't have room or budget or time to invest in a complicated speaker setup, they should not be dissatisfied. Smaller systems are perfectly valid.
Last edited by josq on Wed Jul 20, 2022 3:13 pm, edited 1 time in total.
Offline

mnailor

Member

  • Posts: 1602
  • Joined: Sat Nov 30, 2013 5:57 pm
  • Location: Atlanta, GA

Re: When will we reach this level?

PostWed Jul 20, 2022 9:19 am

Based on the original post, we're talking about what it takes to get Hauptwerk to sound as good as Pipedreams streaming audio, which isn't even as good quality as a stereo CD. So factors like IMD aren't really involved in the comparison (not that they aren't relevant to HW audio). What else is missed in samplesets that keeps some of them from sounding as good as a CD of the organ?

In addition to depth (relative pipe attack timing based on distance from ears), I'd consider the sound engineer's role. Polishing the mix and other good stuff to make each CD track sound its best is one of her/his important jobs. Polishing samples to sound their best for every feasible performance on HW is not even possible, so editing samples has to be a generic solution. CDs (performance recordings in general) will always have the big advantage of editing that can be specific to each piece and performer.

I honestly think some samplesets sound as good or better than CDs/recorded performances on similar organs or the same organ, after allowing for my much lower quality playing: Armley, Caen 2.66, Oloron, Goch, Nancy, and Doesburg (newest version) stand out to me. All of them require exploration and careful settings to get there.

I'm not so sure that one can conclude too much from playing a demo subset or hearing contrebombarde/youtube recordings, where the performer may have used apparently random settings. You have to set up the whole sampleset to fit your audio and room and live with playing it for a while to start getting good results.
Offline
User avatar

tf11972

Member

  • Posts: 338
  • Joined: Tue Jan 14, 2014 4:33 pm
  • Location: Bavaria, Germany

Re: When will we reach this level?

PostWed Jul 20, 2022 2:20 pm

organsRgreat wrote:
Finally, we need to be constantly aware that both Youtube and the Contrebombarde site use compression – Youtube uses aac and Contrebombarde uses mp3. The reduction in file size is useful, but compression inevitably degrades the sound.



There is a fundamental difference between reducing data and compressing the sound: The first omits everything what a human being is not able to hear: quiet sounds beside loud ones, sounds below the hearing threshold and so on. The second one is compression which significantly lowers dynamics.
Many who claim to hear a difference between CD and mp3 have failed mercilessly in blind tests. I have often uploaded recordings that I made with 24 bit and then converted to mp3 to Contrebombarde, and a sound difference has never been heard.
Best regards
Thomas

Forestpipes - Virtual Pipe Organs
https://forestpipes.de
Offline

larason2

Member

  • Posts: 752
  • Joined: Thu Feb 04, 2016 9:32 pm

Re: When will we reach this level?

PostWed Jul 20, 2022 5:50 pm

As Colin Pyket describes in his article, the amount of intermodulation distortion is directly proportional to the number of sound sources being played back. So a recorded orchestra, unless it’s some kind of multi track, will always have less IMD than a Hauptwerk organ, unless you have only one stop registered and only play one note at a time. It’s a function of more than one recording being digitally combined, not a function of more than one instrument (or pipe) being recorded. It’s probably that most of us don’t notice the difference.

There probably are a lot of other subtle factors that make the Hauptwerk sound different than a recording. For instance, I personally think how good the organist is makes a difference. In the online comparison tests between Hauptwerk and the original organ (or a digital copy by other organ builders), by the same organist, in my opinion the difference is pretty subtle. subjected to a blind test, many of us probably wouldn’t be able to tell the difference. However, when we’re comparing a recording to the high definition audio of our systems with us playing, it’s a lot easier to tell the difference. Also, with all simalcrums, there are things the simulation does well, and other things it does not. For instance, if you turn of the wind and hold a chord, Hauptwerk doesn’t sound as authentic as a real organ. Arguably it’s not designed for this, but it’s just an example. For the purposes of playing a piece of music, I think it does pretty well though, usually “can’t tell the difference from a recording” well. There are actually few recordings that I would consider show the realism gap between Hauptwerk and a recorded organ, and usually the organ is being used on the edge of it’s abilities, or it would be challenging to make the sample set sound the same. Usually smaller sets in smaller acoustics do better, and bigger sets in big acoustics are the hardest to pull off. That’s why I practice a lot of Bach on smaller Baroque style organs. They just don’t sound right on bigger organs to me, even from the time period.
Offline

josq

Member

  • Posts: 912
  • Joined: Tue Jan 24, 2012 12:11 pm

Re: When will we reach this level?

PostThu Jul 21, 2022 4:40 am

larason2 wrote:As Colin Pyket describes in his article, the amount of intermodulation distortion is directly proportional to the number of sound sources being played back. So a recorded orchestra, unless it’s some kind of multi track, will always have less IMD than a Hauptwerk organ, unless you have only one stop registered and only play one note at a time. It’s a function of more than one recording being digitally combined, not a function of more than one instrument (or pipe) being recorded. It’s probably that most of us don’t notice the difference.


There are multiple potential sources of (intermodulation) distortion.

* Distortion created in the digital domain/inside Haupterk, when combining multiple recordings/samples. I don't know about any mechanism that could create this kind of IMD. If it exists, it should be measurable by intercepting and analysing the audio output of Hauptwerk. Pykett mentions phase differences, but a phase difference does not introduce IMD, it only modulates amplitude (cancelling). However, repitching in Hauptwerk is a source of distortion, see https://www.sonusparadisi.cz/en/blog/au ... uptwerk-7/. This distortion is much less when using the high audio quality settings in HW 6/7, although I have to admit I have not yet been able to hear the difference in practice.

* Distortion created in the signal path from Hauptwerk to the speakers: during digital to analog conversion (the DAC unit of your audio interface) or during amplification. In the following review you can find a 32-tone measurement of a Topping DAC: https://www.audiosciencereview.com/foru ... amp.27148/. Distortion is around 120 db below the signal (for individual frequencies), way beyond audibility. Here is a Topping amplifier: https://www.audiosciencereview.com/foru ... ier.27079/. In the amplifier the distortion is about 100db below the signal, also practically inaudible. However, other brands might have much higher distortion levels.

* Distortion created when multiple tones are being played through the same speaker driver. At this stage, it does not matter if the tones are from a single or from multiple recordings. This is the distortion mechanism mentioned by Colin Pykett. He shows sum and difference tones that are individually 30-50 db below the signal, which certainly could be audible. IMD measurements of speakers are often a bit hard to get. However, IMD is closely related to harmonic distortion, which is often and easily measured. The percentage of distortion is not fixed (as suggested in many speaker specifications) but goes steeply up when the speaker is playing louder. So if you like to play loudly, get speakers that can play loudly without much distortion. If you always play at low volume levels, speaker distortion is not something to worry about. Pykett's theory that speaker IMD could be kept in check by using separate drivers for each of the 12 scale tones makes sense, except that in practice Hauptwerk uses a lot of detuning which can give rise to a lot of additional sum and difference tones.

* Distortion created between your ears. This is known as Tartini tones. Those sum and difference tones are created by your brain when your ears receive multiple tones simultaneously. They also occur with real instruments. Tartini tones are not measurable by a microphone, but yet I often hear them clearly, particularly when playing 2 or 3 voices with a single (flute) stop. I wonder if the IMD demo given by Pykett demonstrates Tartini tones, rather than anything of the above.

When I learned about Tartini tones, I had to conclude that any quest for 100% perfection is doomed to fail: our very own brain is spoiling the music. Sometimes you'll hear the music and enjoy, sometimes you'll notice the distortion and be dissatisfied, and this will often depend on our mood.
Offline

1961TC4ME

Member

  • Posts: 3144
  • Joined: Mon Feb 07, 2011 1:45 pm
  • Location: Lake Minnetonka, Minnesota

Re: When will we reach this level?

PostThu Jul 21, 2022 10:33 am

Hello all,

Good stuff here for sure! Thanks all. And thanks, Martin for your reply, input and explanation as well.

First off, listening to the recordings I referenced on Pipedreams, no it's not as good as being there in person, no recording is. However, put on a good set of headphones and listen to the recordings, it's very pleasing. Also compare the Pipedreams recordings to any sample set with your headphones on and you will definitely hear the difference.

Out of curiosity I ran a test this morning. First off just to get the sound in my head I played a portion of the Pipedreams program via blue tooth from my phone through just 2 channels of my HW sound system. It sounds very good with much clarity and spaciousness. I don't know if this is a good way to describe it but you can hear the 'air' or the space in the recordings. Is this what we call head room? Overall it was a cleaner sound and the lower notes really come through much stronger than any sample set I own as well. Just much more realism.

Here's what I'm thinking..... Rather than the sample set producer recording the instrument and each pipe in it's acoustic from various perspectives, instead record the instrument and each pipe as dry as possible, then make separate IR's of the room the instrument resides in from different perspectives, allowing the end user to apply the perspective of choice. You could even apply the various perspectives via IR's to side channels, rear channels, what have you.

I own the Dingelstaedt instrument from Sonus Paradisi. Overall I've never been too happy with this instrument because it's very mushy, no matter how you adjust the sliders for near or distant the thing just doesn't sound good. Some stops are very harsh, trumpets have an annoying buzz and distortion here and there, certain stops just get completely buried in the sound, etc., etc. So, figuring this one might be a good candidate for a test, I loaded it and set the perspective slider to as dry as the set gets. In this position you can still hear some of the acoustic, but the sound does become very 'near' to you. Keeping it in this driest position I then experimented and applied the various IR's in Hauptwerk to all 10 channels. The one I felt did best was Church04-Omni-16m WideLR 4.5s. What a difference! This really woke the set up, there was better clarity, the mushy mixing was mostly gone, and suddenly the majority of the harshness was gone as well along with the annoying trumpet distortion. It's still not as good as it probably could be due to the fact there's still some of the original acoustic and harshness in the mix, but the overall clarity and smoothness improved nicely. Perhaps a good voicing session would improve it further but I wanted to just do a head to head as is comparison for starters. This makes me wondering how things would work if again the instrument is recorded as dry as possible and it is then provided with the IR's of the actual space we can then apply. I mentioned earlier the recordings of each pipe in it's acoustic piled on top of each other likely causing some issues, this would eliminate that issue as again the sampled pipes would all be sounding in 1 acoustic.

I don't know, maybe I'm all wet here but this test I performed this morning was rather interesting.

Marc
Offline
User avatar

dkoschinski

Member

  • Posts: 89
  • Joined: Tue Jun 07, 2011 11:55 am
  • Location: Netherlands

Re: When will we reach this level?

PostFri Jul 22, 2022 5:00 pm

As a producer I achieve a proven realism of 95% (guaranteed). Nothing less, but with pleasure even more than this,

Regards,
Danny Koschinski
Virtual Pipes
Offline

1961TC4ME

Member

  • Posts: 3144
  • Joined: Mon Feb 07, 2011 1:45 pm
  • Location: Lake Minnetonka, Minnesota

Re: When will we reach this level?

PostFri Jul 22, 2022 9:40 pm

dkoschinski wrote:As a producer I achieve a proven realism of 95% (guaranteed). Nothing less, but with pleasure even more than this,

Regards,
Danny Koschinski
Virtual Pipes


Well, I can't thank you enough for all the time.

Regards,

Marc
Offline

münsterorganist

Member

  • Posts: 135
  • Joined: Sat Aug 27, 2016 2:24 am

Re: When will we reach this level?

PostSat Jul 23, 2022 8:42 am

my humble opinion: the stops are sampled individually, each for itself. That's why only a few stops sounded in HW very authentic. During the addition, despite wind modulation or detuning of the samples, there is no real mixing of the stops, nor can it be, since the stops are recorded individually and always played back individully. Only HW can offer a solution here by imitating the pipe mixing and coupling.. As I said: in my opinion.
Offline
User avatar

IainStinson

Member

  • Posts: 1391
  • Joined: Tue Dec 29, 2009 6:08 pm
  • Location: NW England, UK

Re: When will we reach this level?

PostSat Jul 23, 2022 8:46 am

Here's what I'm thinking..... Rather than the sample set producer recording the instrument and each pipe in it's acoustic from various perspectives, instead record the instrument and each pipe as dry as possible, then make separate IR's of the room the instrument resides in from different perspectives, allowing the end user to apply the perspective of choice. You could even apply the various perspectives via IR's to side channels, rear channels, what have you.


Do some of the digital organ manufacturers (such as Allen) take this approach (recording samples in a very dry environment such as an anechoic chamber ), then add reverberation?

I thought that a key feature of HW was that the samples included the ‘acoustic ‘of the room?

With HW5+ we can use dry sample sets plus good IR, I guess we can have a choice.

You may find this account of sample set production of interest - it is from a recent public newsletter from a large European digital organ company.

How do these recordings get made, anyway? "We use high-quality recording equipment that we generally set up in the church in the evening,” ….“We place a total of eight microphones in various positions in the church, with a few of them as close as possible to the front of the organ, so we capture the information in great detail. By recording at multiple locations within the church, we make sure that we can use the recordings for all applications, and that we are prepared for the future.

“We used to have to press the keys one by one ourselves, but we have a robot for that now. Each stop is recorded key by key, for ten seconds per pipe.”
 
It’s worth mentioning, ….that sound and reverb are recorded separately. “Since we record each key separately, we can decouple all the keys from the reverb individually later. This is a complicated process, but it has considerable advantages. The most important of these is that we retain the sharpness and purity of the original pipe organ sound. We can then use the authentic reverb to mimic the desired acoustic effect.
 
Thanks to this recording technique, organists can even determine their preferred ‘reverb position’ in the church from behind their digital organ, whether they prefer to hear the organ the way it would sound from their position on the bench, as it would sound from the front-row seats, from the middle of the church, or from a location beyond the scope of the reverb. “We record the reverb at four positions, and by thoroughly analysing the differences in these recordings, we can get a very accurate idea of the effect of the reverb on the original sound. This information enables us to reproduce the sound as faithfully as possible.”
 
Offline

josq

Member

  • Posts: 912
  • Joined: Tue Jan 24, 2012 12:11 pm

Re: When will we reach this level?

PostSat Jul 23, 2022 3:10 pm

I dont think recording stop combinations is a viable idea. Because there can be literally millions or trillions of unique stop combinations. If not all combinations are recorded, it will happen that by changing just one stop the entire sound character will change.

Modelling is the better approach. If a stop combination sounds different than expected, try to understand the effect, and then approach it in a model. With proper understanding and modelling, chances are that the model will be correct even for stop combinations that you have not tried yet.

I think the wind model is extremely important, but it is not the only model needed.
Offline

mnailor

Member

  • Posts: 1602
  • Joined: Sat Nov 30, 2013 5:57 pm
  • Location: Atlanta, GA

Re: When will we reach this level?

PostSat Jul 23, 2022 4:21 pm

The sampled stop combinations on one of the SP harpsichords pointed out a problem for me with that approach: If my 8' and 4' ranks are routed to different speaker groups, whichever group I pick for the 8 + 4 rank doesn't sound even close to the two ranks played together on two speaker groups. I had to route the instrument to one group (out of four).

Changing registrations can be jarring instead of a smooth add/subtract of one stop when it triggers a combination sample. It only makes sense to have sampled combinations and individual stops if they're all on the same bus group, which is. And even then, switching to a combination sample when a stop is added may not be as smooth as adding a sampled stop.
Offline

TubaFan

Member

  • Posts: 54
  • Joined: Fri May 15, 2020 8:57 am

Re: When will we reach this level?

PostFri Jul 29, 2022 7:35 pm

HW VI is much better than V in terms of sound, so I'm not sure it's fair to compare V to the Pipe Dreams recording.

I can't imagine that it is easy to achieve realism in a living room, but in a large building it can be done effectively. In terms of reducing distortion, I have found that using speaker groups of 4, mono, with tone matching mode 1 have made for a very clean sound. You also need to work out how you assign the stops to these groups.
PreviousNext

Return to Hauptwerk instruments

Who is online

Users browsing this forum: No registered users and 6 guests