More routing experimenting

Hello all,

Some time back I experimented with 4 different pipe routing schemes with my 8 channel set-up and reported the results here >> viewtopic.php?f=17&t=13679

After all this I had concluded, at least for my system, that my 'divisions routed' scheme outlined in the above link sounded overall the best, providing much realism along with the best stereo field and clarity. I have also pointed out in a few earlier posts that these results are best achieved with a 'stacked' left and right front speaker layout along with a rear pair of speakers. I got to thinking the other day that with my divisions routed configuration I am basically asking pipes of dissimilar timbre to sound with each other on the same pair(s) of speakers per division, trumpets along with flutes, mixtures with other principals, etc., etc., which got me wondering if perhaps an even better routing layout would be to route pipes to pairs of speakers in groups based more on their timbre and or what family of pipes they belong to.

Just for reference, here's how I have my audio laid out.

Channels 1-2 stereo amp to 2- T15 Polk bookshelf speakers
Channels 3-4 identical to above
Channels 5-6 stereo amp to larger 3- way Bose speakers, sub also runs off this amp.
Channels 7-8 stereo amp to rear bookshelf speakers used either for mix-down or for rear surround signal.

Now keep in mind, with a total of 8 channels to work with this isn't perfect, and a few more channels would likely be even better, but the results were quite remarkable none the less!

I started out with the Zurek v5 set which is a rather good free and sizeable set and is also a surround set, to me it's been a very good sounding set and I've enjoyed it much. Being it is a surround set we are actually limited to using only 6 of the total 8 channels for this routing arrangement as the rear pair receive all the rear surround signals of the Zurek. I went to the rank table and went down the list of stops, anything in the reed family or similar sounding like nazards, quints, etc., went to channels 1-2, anything foundation sounding / flute / etc., went to channels 3-4. One small dilemma came with channels 5-6 which are connected to the larger pair of 3-way speakers that is also connected to my main sub which I've used all along for the pedal division. Here I decided I would send the entire pedal division to channels 5-6 along with the mixtures figuring the pedal division is much smaller, only certain stops of the pedal are generally used at the same time depending on the registration, and since I really had no other choice to keep the mixtures separate from everything else, that this would be best as the mixtures seem to stand out pretty well on their own anyways.

Results? This routing scheme is by far the most realistic sounding arrangement I have come up with! Clarity alone jumped about 50%, no kidding! The reeds, especially when coupled from the pedal to the manuals just bark with authority now, the set just rumbles with realism. The stereo field remains nice and wide, but probably the biggest thing I noticed is how well you can now hear the individual stops and how well they work with each other, nothing seems to get buried in the mix and the thing really sounds now like you're playing a real organ, it literally transformed this set!

I next tried it on the St. Max from SP, although not as dramatic, I got very similar results here as well, more clarity, no harshness, great stereo field and more overall realism.

Unless I come up with another routing arrangement which is better (I now doubt it), this one will now be the one I go with. Anyone here with 8 or so channels should try this arrangement and see what you get, I'd be interested to hear if your results are the same.

Marc

Hello Marc,
Having started with two channels per department, I have been fiddling around with routing with varying degrees of success so I was very interested in your article. I too am limited to 8 channels (Cloud CX-A850 amp. 8 Bose 301 speakers plus a sub and two underseat Dayton tactile transducers).
I first tried your formula on Hereford and was very impressed with the improvement of sound quality especially the full swell. Now I am working my way through the rest of my sample sets Caen, Rotterdam, Zwolle, St Eucaire etc. with great success.
Like you I think that there is a good 50% improvement in quality with less clashing of harmonics, better reeds and generally more clarity. It seems to be making the best use of eight channels. Well done and thanks!
Best wishes
Julian.

Hello Julian,

Very glad to hear it's working for you as well! Like you, in a quest for better sound and overall realism I've fiddled around with several different routing schemes. After giving thought to this grouping of similar sounding stops per speaker pair, it seemed on paper at least to make the most sense. What got me thinking about it is I considered if we look inside an organ case, the different sounding families of pipes are all laid out together in groups, so I figured why not put them together in groups based on their sound / family per speaker pair?

I used to think that trying to just randomly distribute stops (and many of dissimilar sound) evenly between speaker pairs was best, but it never really panned out the best and I ended up with varying results, most I wasn't completely happy with. With this arrangement, even though you may end up with a pair or two of speakers handling more stops, just by the fact that they are handling similar sounding stops has made all the difference here for me and I continue to be impressed with this latest arrangement. What's really amazing though is how much of a difference in sound a person can make just in how you route things.

Marc

Hi all

I too have an 8 channel system plus subwoofer with two channels being surround channels. I've been routing divisions to a speaker pair but over the holidays decided to experiment.

I tried Marc's scheme on some of my smaller sample sets (St Eucaire for one) and found it worked quite well. For this smaller set I did try putting one of the flute stops and Viole de Gambe in with the channel carrying the reeds to separate them a bit when using quieter registrations and put the celeste type stop in with the pedal and mixture stops. I find separating the celeste stops from the other flue stops helps quite a bit and figured I wouldn't be using the mixture along with the celeste anyway. With this working out well I went to the larger Hereford set and did a similar routing scheme.

Overall, full registrations are much improved, especially when the reeds are added. The quieter registrations are not impaired at all. It did improve overall clarity and realism.

I did more experimenting over the past week and it comes down to this for me anyways..... If the set has a very large reverberant acoustic, perhaps a bit too mushy sounding, this latest scheme I lay out above cleans things up and adds clarity and may sound best, but it depends and is really a matter of preference. I do notice in my case that it does seem to narrow the stereo field some as the trade-off. I went back to my 'divisions routed' scheme for St. Max and like it better, but for the Zurek v5 it was no contest, the latest pairing of similar sounding stops to speaker pairs is waaaay better and added much realism, extra bite of the reeds and as I mentioned, really rumbles now. I just aquired Dinglestadt from SP and have tried both schemes on it (divisions routed and similar sounding stops paired to speakers), for this one I like divisions routed a bit better, mainly due to the fact the stereo field opens up a bit more. I guess it comes down to if you're the patient type and can wait while the set loads with the new configuration for the first time, try both and see which you prefer, both are very good.

One of those six to one, half dozen to another scenarios.

Marc

Good Morning Marc,

I emerge from the Holidays with pedal board repaired!!! Well almost -- Eflat wants to stick but can correct that today. As always I continue to follow your routing scheme threads with great interest. I gave a quick try last night to your latest (like sounding pipe pairs - LSPP) and was immediately impressed with the distinctive sound. I use Mt. Carmel Skinner set and am a fan of the warm, rich, registrations using the strings and reeds quite extensively. The separation given these "families" was very enjoyable with the LSPP routing you suggest. I have a bit of trouble deciding where to send those cousin-type pipes (Mixtures, 15th, Octave 4 and Principals). I believe I put the Principals into the Flute Family, and the Mixtures into the Pedal Family. I am running 6 Behringer B2031A's and a Velodyne Sub (12"). A pair of JBL tower 3 way speakers are attached to the sub.

Because of room shape I am unable to effectively "stack" the speakers and unfortunately speakers arranged elevated to nearly shoulder level with approximately 6 to 7 feet between each starting from left side of console in 4 quadrants around the room (basically nearly a circle of speakers with 6-7 feet between each). In actuality this does produce quite a nice full sound. I have routed Flute Family to 1&2, String Family to 3&4, Reed Family to 5&6, and Pedal + Mixturse to 7&8 which is the sub and its two JBL towers.

Any suggestions you might have regarding where to place my odd and end stops would be appreciated, but I believe your new LSPP routing system is a winner!!!!

Tom

Hi Tom,

First off I like your naming it LSPP idea! We gotta come with names for this stuff like my 'divisions routed' term, that way everyone knows what the method is. I basically treated anything I'd consider flute like sounding pipes, such as principals, montre, bourdon, prestant, ect., to basically be in the same family. Anything like mixtures or for the lack of a better word 'off sounding' stops depended on its timbre get grouped together. If it was more reed like sounding I threw it in with the reeds, more mixture sounding went with the mixtures. Part of it was even if two different but similar sounding pipes went to the same pair, I then considered if they would even be used together so didn't worry too much if a few ended up in the same speaker pair. There's certainly some room for experimenting here which is what is part of the fun!

Glad to hear you like the results!

Marc

Hello Marc:

I upgraded to HW Advanced in December and have been learning and experimenting with a few basic routing schemes. First I want to say thank you for taking the time to post your findings in great detail. I enjoy reading them, especially the earlier post where you outlined the 4 separate routing schemes along with a pros & cons critque! In fact I have forwarded that post to a few HW colleagues who rarely participate or read this forum to use as a reference.

For my setup, I have 5 stereo pairs: 3 Behringer 2031a monitors, 1 bookshelf, 1 Cervin-Vega D7. The D7s are very large, full range, with distributed mids and excellent bottom end down to 35Hz. The D7s by themselves in a straight stereo config can handle most sample sets just fine. When I purchased them many years ago, I did my comparison shopping by playing back several organ recordings. So I wasn't too surprised they can handle actual pipe organ sample play back (Hauptwerk).

What I wanted to share on this thread is my multi-channel initiation has started with St. Anne's since I am most familiar with this instrument. For a quick and easy start, I assigned the Pedal and Great to the D7s and the Swell to a group consisting of only 2 of the Behringer pairs (4 speakers in stereo mode). The Contra Faggoto 16' foot in the Swell I used the bass-split feature so the last octave is routed to the D7s. Using the CueMix software with my MOTU audio interface, I did some minor EQ adjustments on the Behringer's channels to better match them with the D7s so when Swell and Great are coupled or played hands separate it will sound a little more cohesive. After the new organ cache creation, St. Anne's indeed had more presence. I don't think it was a much noticeable improvement but a more fuller sound as you'd expect with a wider sonic spread.

Then for fun I took the 3rd Behringer pair and made the 2 separate mono channels. I created an audio group just for them with the OCTAVES CONSTANT (C/C# SPLIT) method and assigned these last 2 Behringers to it, thus having two separate mono channels. I then went to the Ranks Routing menu and assigned the Great Trumpet to this channel. In addition I raised the volume of the Trumpet rank as well. You can probably guess where I'm headed . . . a 'poor man's en chamade'! Just about every real, exposed, vertical Trumpet or Tuba I had the pleasure of playing is arranged on a diatonic chest so I wanted to experience that with my new setup. Whoo hoo, believe me I got to experience it I got chills playing some trumpet voluntaries with the added realism of a wide stereo field and how easy it was to configure! Not to mention how easy it would be to change it back or to something different. As expected, when the Trumpet is added into a full registration, the C/C# ping-pong effect is less noticeable since notes in the neighboring monitors are also moving between the other two Behringer pairs configured as stereo. With those, you always have sound on the left and right of the console.

I'll soon graduate to my larger sample sets for mult-ichannel routing and will sure to try your "divisions routed" design.

Danny B.

I recently had the opportunity of moving from 4 to 8 audio channels and have tried this arrangement with very positive results. There is a level of realism that I had not experienced previously.

I find it interesting that conventional wisdom for reproducing a pipe organ sound would seem to suggest that each division of the organ should be kept together like they would be in a real pipe organ installation (for example all of the great stops should sound in one part of the room, swell in another, and choir from yet another location.) This arrangement obviously works with real pipes where each pipe speaks individually giving a sense of "movement" within each rank and sense of "location" for each division. But I'm wondering if this wisdom does not apply as well when the sounds are produced through a limited number of speakers (rather than having one speaker per pipe - we could only wish!) It is also possible that the conventional placement of divisions works in a large space, but not so well in the smaller spaces available to most of us with Hauptwerk organs.

When I reviewed the proposed methods for spreading the sound around, it seemed like there were really two methods under consideration (which are sometimes in conflict with each other.) The first method was to keep similar sounds together (e.g. all flute sounds to one set of speaker, all reeds to another, strings to another etc. regardless of the pitch level for each sound). The second method was to keep similar pitched stops together (e.g. 32 and 16' pedals to one set of speakers with a sub-woofer, 8' to another, mutations to another, mixtures to another, etc.)

I'm no audio engineer, and not especially knowledgeable about the physics of sound, but I think the underlying need may be to keep stops with similar harmonics together and to send them to the speaker(s) that are best suited to those harmonics. I found a table of harmonics (fundamentals and overtones) in "An Introduction to Organ Registration" by James Engel on page 8 that would broadly suggest the following groupings:

Group A: stops with strong 1st, 2nd, 4th, 8th, and 16th partials (i.e. 16', 8', 4', 2', 1' stops) together, but send the 16' and 8' to a speaker that handles them best (including a sub-woofer), 8' and 4' to most any speaker(s), and 2' and 1' to the smallest speaker(s).

Group B: stops with strong 3rd, 6th, and 12th partials (i.e. 5 1/3', 2 2/3', 1 1/3') together, sending them to the best suited speaker(s) (i.e. good tweeters)

Group C: stops with strong 5th and 10th partials (i.e. 3 1/5', 1 3/5') together, sending them to the best suited speaker(s) (i.e. good tweeters)

Group D: stops with strong 7th and 14th partials (i.e. 1 1/7') together, sending them to the best suited speaker(s). (i.e. good tweeters)

These groups may make sense from a theoretical perspective, and maybe they make sense for what real speakers can handle simultaneously (I don't understand speaker dynamics enough to say for sure); but in the real world of stops, once one moves past the flutes with their relatively simple harmonics, it becomes much harder to categorize other stops, but some broad suggestions might be helpful:

Principles and flutes clearly belong together in Group A
Most mutations belong together in Group B
Reeds vary greatly, but clearly don't belong in Group A
String can also vary somewhat, and might span several groups depending on the stops actual sound
Mixtures have very complex harmonics, and clearly don't belong in Group A

A few assumptions might also be relevant:

Flutes and reeds aren't normally used together (since the reeds overwhelm the flutes)
Reeds, mixtures, and strings aren't normally used together (since the reeds and mixtures both have rich harmonics and either would overwhelm most strings)

All of the above sort of come together in three loose "rules":

1. Group by harmonics when possible
2. Group by pitch range when possible, sending each pitch range to the best suited speaker(s)
3. Allow for some sharing of speaker resources (which are always limited), especially between stops that aren't normally used together.

I had 4 stereo pairs of speakers to work with:

Speaker Pair 1: Behringer Truth B3031a with sub-woofer, near the console
Speaker Pair 2: Behringer Truth B3031a, stacked above pair #1
Speaker Pair 3: M-Audio BX5a, to the rear of the room
Speaker Pair 4: M-Audio BX5a, to the side of the room, nearer the console

With these speakers and the "rules"above, I set up my audio routings for each organ as follows:

Speaker Pair #1: Nearly all 32, 16', and 8' pedal stops without much thought to the harmonic groupings since these lower pitched stops tend to have simpler harmonics. Most 16' manual stops including 16' strings.

Speaker Pair #2: Nearly all 8 and 4' principles and flutes, most 8 and 4' reeds, and some 2' flutes and principles.

Speaker Pair #3: Lower pitched mutation-like stops including quints with a strong 5th, 2 2/3, and mixture-like stops, and 8' or higher strings, celeste-like stops, some 2' flutes and principles, 1' flutes and principles and special effects like blower and stop noise (since these are farthest from the console)

Speaker Pair #4: Higher pitched mutations, some mixtures, and most special effects like keyboard, and tracker noise, etc.

I suppose the ultimate test is how everything sounds in the available space (whether large or small) and through the available speakers (whether few or many.) In this case I found the sound to be surprisingly clear and well-spread around the room, clearly better than some other arrangements I had tried. I found the "rules" useful in quickly deciding how to route each organ, and I think they have some basis in the physics of sound and speaker dynamics, but I'm guessing that they could be varied considerably without greatly compromising the sound.

My thanks to those who have contributed to this topic. Clearly there are many good things to explore.

Thanks to Marc, and all who are contributing to this ongoing discussion. Having just finally figured out how to make multiple audio output groups work, and having tried this, I have to agree that in the tiny space I have and using exclusively wet sets, this sounds significantly better to me than did octave cycling through one output group (although that was a big improvement over just single channel stereo). It is currently set up as follows:

Channel 1 - Foundations: all manual 16, 8, and 4' principals and flutes, as these have essentially sinusoidal waveforms. They go into a pair of M-Audio Bx8a, aside and at ear height. I am happy to have some roll-off of the 16' stops in the bottom octave, as this seems less muddy and ponderous, especially against the pedal.

Channel 2 - "Coloration": All flue stops above 4' pitch on the manuals, plus any string stops plus manual reeds on surround sound sets. IOW, until the reeds are drawn, nothing that requires much excursion of the main speaker cones, which factor I would presume to potentially introduce distortion of higher partials sounding at the same time. These are served by a pair of Bx5a in front of the console, near the ceiling. This makes for a more dramatic contrast when upperwork is called upon, and the 2' seems especially to benefit from being separated from the foundations.

Channel 3 - Pedal: The entire pedal division. This seems to me to hold the pedal together better than splitting up the stops. This is currently served by a stereo amp driving the original pair of Allen HC10 (15" woofers only) and a pair of Polk T15s, alongside the console and beneath the Bx8a. With this arrangement, the upperwork sings out above the rest of the division and clearly defines the pedal line, without drawing attention to itself against the manuals. I am currently putting the celestes here too, to separate them from both the strings and flutes, and because the crossover arrangement essentially isolates them to the T15s alone if only the soft pedal foundations are drawn (which then have the HC10s mostly all to themselves), as is usually the case.

Channel 4 - Antiphonal: Any surround ranks; or, if the sample set is non-surround, this is where the manual reeds go. These are served by another set of Bx5a, placed behind the console.

Among other nice things about this setup is that having the upperwork on a separate speaker pair seems to give it better presence so that it can speak at lower volume, reducing the overall shrillness yet remaining assertive. While it may not so perfectly mimic the directionality of sound as it would be in a large space, the result seems, to me at least, to be much more pleasing when one has to make do in a more confined area.

Nothing so terribly new here; I offer this mostly as just another vote of confidence that you guys are on to something good. What's really amazing to me is that all this subtlety makes such a difference, considering that in the end it all has to be processed by only one tiny set of eardrums ...

Hi guys,

First, I'm glad to hear that so far everyone who has tried this routing scheme has come away feeling they've heard an improvement! It makes me feel good that I can contribute something that enhances everyone's experience with Hauptwerk! When it comes to sound I'm very picky, usually to the point I won't leave things alone until I figure it out, so I have spent countless hours just on routing and speaker placement attempting to figure out what gives the most realism. Whenever I get the chance to be at a church that has both a good organ and acoustics, I listen very carefully, get the sound in my head, then go home and attempt to recreate it. We've got two very good examples here in the Minneapolis area at the Basilica of St. Mary and the St. Paul Cathedral, two organs and acoustics to die for (well almost!). As Stuart mentions, there's what I'd also consider conventional wisdom here, but when I really started looking at things, due to the fact we are dealing with speakers (and in most cases not that many), it made sense to me that perhaps grouping ranks together based in their timbre and keeping them together in speaker pairs would likely be the best. Stuart and Sylvaticus both did a great job of going into detail, much greater detail than I did, but is exactly what I was thinking when I came up with this.

Just out of curiosity and if anyone is wondering what difference it would make, I did experiment a bit further and moved anything 2' and higher out of the 16', 8', 4' principal / flute sounding group into the mixture group and the results were still the same, I couldn't tell any difference of keeping them in the same speaker group or in with the mixtures, so it seems it is O.K. to have anything 2' and higher in with the mixtures if you are running out of channels to pair thing up.

The biggest thing I have noticed with this arrangement is how much it bumps up the clarity without adding harshness. If you have a set that just seems dull or that certain parts lack clarity or presence and no matter what else you have tried has not given satisfactory results, then this might just do the trick. It's also pretty remarkable that you can leave your speakers in the exact same position and depending on the way you route things how much of a difference it can make in the overall sound.

Also, what's equally important in all of this is the speaker positioning itself. Just turning them a little one way or the other, bringing them closer together or further apart, raising them, lowering them a little, all can really change things as well with some good some bad. I've written about this here as well in other posts, but if a person can stack speakers, starting with the lowest drivers at about head height and speakers going up from there when seated on the bench has by far sounded the best to me. Equally important is each stack's spacing and I've settled on about 5'- to 6' stack spacing to be ideal and you end up with a very nice stereo field in front of you. With wet sets it gives the illusion of maybe standing a quarter of the way out in the church with the organ in front of you. Then with rear speakers added and again up higher (6' or more off the floor) vs. lower and with proper volume adjustment really opens things up, it's like going from 2D to 3D.

Thanks again all!

Marc

Just to add a bit more here, I recently added a second sub to channels 3-4 for the manuals. I decided perhaps to look things over one more time and to take an approach to simplify the process in deciding which ranks go to which speaker pair. I first played each stop one at a time for both the St. Max and the Dingelstaed, then determined in what speaker pair / stop grouping they belong and made a list as I went along to keep track of things. I then proceeded to load both sets again and I did end up with a few minor changes compared to the previous way I had loaded both sets. It probably comes as no surprise, being foundation stops out-number most other stops on a given organ, one speaker pair is going to end up doing more duty with all the foundation stops, whereas the other speaker pairs don't end up with nearly as much to do, but it sure works and does not detract from things at all! I'd think even another pair of channels to split up duties for the foundations might work even better, but again things are very good in my arrangement as is.

Since I've found 2' and higher pitched ranks can be a bit on the shrill side (listen to them individually to determine this) and if anything may be the ones needing some voicing work, this can be accomplished in two ways, either by using the voicing facility or in my case which I can quickly accomplish using the EQ on my sub amp that is also attached to my larger pair of 3-way speakers. If you can't iron things out with EQ, then you can do some voicing as well. But in the end I decided since I do have the EQ to use, I kept anything 2' or higher (including mixtures) grouped together and sent to the sub / larger 3-way pair. I did find in the case of the St. Max though that although the nazards are off pitch, they are not shrill sounding and do have a bit more of the timbre of a foundation sound to them, so in this set's case I grouped them in with the foundation stops. After doing a little EQ work to settle the high pitched ranks down just a hair and bring them in line with the rest of both sets, all I can say is WOW! There is now an added level of realism and presence in both sets that I have never heard before. Sounds like you're sitting right there in the church listening to the real thing.

Marc

Marc,

I'm ready to give this a try, but not really sure how I would group ranks with this speaker set up:

Group 1 - Adam F5 - stereo
Group 2 - Adam F5 - stereo
Group 3 - Adam F5 - stereo
Group 4 - Spendor SA200 - stereo (currently used for the Solo, seems to handle reeds well)
Group 5 - Adam F7 - stereo with SVS SB2000 sub

All speakers stacked in two columns on each side of console (listed in order from top to bottom). Interested in your thoughts.

Michael

Marc,

I'm ready to give this a try, but not really sure how I would group ranks with this speaker set up:

Group 1 - Adam F5 - stereo
Group 2 - Adam F5 - stereo
Group 3 - Adam F5 - stereo
Group 4 - Spendor SA200 - stereo (currently used for the Solo, seems to handle reeds well)
Group 5 - Adam F7 - stereo with SVS SB2000 sub

All speakers stacked in two columns on each side of console (listed in order from top to bottom). Interested in your thoughts.

Michael

Hmmm.... well, it is subject to a bit of experimentation and will depend some on which set. From what you describe above I will assume you are not using any of the speakers as a rear pair for surround? That would possibly be the only thing I'd change would be to move one pair to the back for surround duty, but the plus here is you also have an extra front pair to divide up duties. If your stacks are getting a bit on the high side you can also stack the matching size speakers in two's side by side then the next pair on top and so on. I've tried both arrangements (directly on top of each other and the side by side in two's) and both seem about the same, the trick seems to be as long as all the speakers stay in-line with each other vertically.

For your current arrangement I'd probably try something like this....

Group 1- All foundation 16' and 8' flutes, bourdon, montre, etc.
Group 2 - All foundation 4' and 2' flutes, bourdon, montre, etc.
Group 3- All off pitch sounding stops such as nazards, quints, etc.
Group 4- All reeds, trumpets, clarions, bombards, etc.
Group 5- All pedal stops, 1' and mixtures.

Any stops you're not sure of, listen to them individually first to determine which group to put them in.

The only other change I might consider is adding a lesser expensive sub to attach to group 1, it really helps the 16's.

I've found you can include the 1' in with the foundations and it really doesn't make a difference either way, but keep the mixtures by themselves in one group, I've found they fit / sound best in the pedal group. If they end up too bright you can easily do a little voicing work to tone them down.

Let us know how it goes!

Marc