It is currently Fri Mar 29, 2024 7:27 am


Görlitz - Polyphony

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

ludu

Member

  • Posts: 999
  • Joined: Wed Nov 02, 2011 8:52 am
  • Location: Tournai (Belgium)

Görlitz - Polyphony

PostFri Sep 15, 2017 11:26 am

Hello,

I have installed the full edition of Görlitz (the 6 channels) on my PC with a ram of 64 GB. 19GB are free and my polyphony is set to 24,000, but when I play large chords on the tutti with trills, the sound stops a short time and comes back after a moment. Is this sample set really playable for Widor and Dupré? Or do I miss something in my settings?
Luc
Offline

student

Member

  • Posts: 25
  • Joined: Thu Jul 04, 2013 4:27 am
  • Location: Timisoara, Romania

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 12:49 pm

I guess your polyphony is set too high and the CPU can't process all those sounds in time! Try a more conservative setting (like 5000) and go up from there little by little, until the sound starts to stops, then lower below the breaking limit.
Offline

ludu

Member

  • Posts: 999
  • Joined: Wed Nov 02, 2011 8:52 am
  • Location: Tournai (Belgium)

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 1:37 pm

Sonus Paradisi recommends a polyphony of 9000 voices for the full organ. I started with 8000 and have increased gradually to 24,000 voices. I am not sure the problem is there.
Luc
Offline

Johannes Sørensen

Member

  • Posts: 228
  • Joined: Sat Jan 20, 2007 10:25 am
  • Location: Copenhagen, Denmark

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 2:02 pm

Hello Luc

My thought goes in the same direction.

The computer shall be extremely powerful to generate so high a polyphony, that you can set it to 24.000 for the organ, roughly the half of the static polyphony.
Have you measured the static polyphony of your system and what is it?
What is the specification of your computer in particular the processor(s)?

Best regards
Johannes
Offline

ludu

Member

  • Posts: 999
  • Joined: Wed Nov 02, 2011 8:52 am
  • Location: Tournai (Belgium)

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 4:43 pm

Hello Johannes,

The specifications of my computer are:
PC Cooler Master
Memory: 64 GB
Processor(s): Intel i7-3820 Bx, 4 x 3.6 GHz
Audio/MIDI: RME HDSPe AIO

Cordially.
Luc
Offline

josq

Member

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

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 4:52 pm

Does the CPU meter at the bottom of the Hauptwerk screen go into orange/red? That should be prevented.

Do you have the same problem if you set the polyphony to 8000? If so, 8000 may still be too high. Try 5000, as "student" suggested, or even 3000. Then increase to find an optimal setting: the polyphony meter should go into orange just before the CPU does.

A too high polyphony may cause the problems that you describe. A low polyphony merely results in a somewhat dryer sound, because Hauptwerk starts to cut off the releases of some samples when playing many large chords quickly with many stops enabled.

There may be other issues, but at this point, too high polyphony is definitely the most probable cause.
Offline

josq

Member

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

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 4:55 pm

I have an i7 3820 processor too, and on my pc, a polyphony of around 3000 seems to be the optimum (2800 to be really safe, 3300 often becomes quite risky).
Offline

1961TC4ME

Member

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

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 7:25 pm

I'll jump in with my 2-1/2 cents worth. Just because it says do it, that doesn't mean you can. I have the SP Dingelstaedt which is also a 6 channel set and quite wet if you want it to be. At around 4200 polyphony with my i5 2500k, the meter is in the yellow and if I go up any higher in polyphony the audio starts to break up. The Gorlitz is probably even wetter, is a larger instrument, and sure, it might be advised for optimal sound to set at 9000 polyphony, but that doesn't mean your computer is gonna get there and handle it. There's absolutely no reason to raise polyphony to 24,000, it's not going to do anything and all you've really done is found the number where audio breaks up and have then surpassed it by a very large amount. In other words you're not doing yourself any favors and 24,000 is waaaaaay too high and not needed anyways.

I too would dial things back to around 4000 or so polyphony and I'd bet the problem goes away.

Marc
Offline

jerrynazard

Member

  • Posts: 123
  • Joined: Wed Aug 21, 2013 7:02 pm
  • Location: Patagonia, AZ

Re: Görlitz - Polyphony

PostFri Sep 15, 2017 10:57 pm

I'm running Goerlitz on a i7-5820 with 128GB ram, loaded at 24-bit compressed. I initially started with 9000 polyphony as recommended by Juri, but occasionally ran my CPU meter into the red. 7100 seems to be the right level for my setup.
Offline

ludu

Member

  • Posts: 999
  • Joined: Wed Nov 02, 2011 8:52 am
  • Location: Tournai (Belgium)

Re: Görlitz - Polyphony

PostSat Sep 16, 2017 4:58 am

Many Thanks to All!

Yes, indeed, my CPU is in the red. Setting the polyphony to 5000, everything is OK now. I had never tried this value because of the recommendation of Sonus Paradisi. However this information on their site must be correct for a conventional configuration. In my case, I have cut the release of the front direct channels to 250 m/s and all the ranks are loaded without compression in order to increase the speed of the loading (a tip I have read recently on this forum).

Cordially.
Luc
Offline

Johannes Sørensen

Member

  • Posts: 228
  • Joined: Sat Jan 20, 2007 10:25 am
  • Location: Copenhagen, Denmark

Re: Görlitz - Polyphony

PostSat Sep 16, 2017 11:50 am

Hello Luc

The processing power of the computer determines how much polyphony the computer can deliver, and that level must not be exceeded by the setting for the sample set. It is that the recommendation of Jiri on the Sonus Paradisi website is about, how powerful the computer has to be to be sufficient for this huge organ.

It is fine that you have solved the problem so far, but unfortunately the processor Intel i7-3820 is not strong enough for this big sample set with 6 channels. That you “have cut the release of the front direct channels to 250 m/s and all the ranks are loaded without compression” helps some, but is probably not enough to avoid that Hauptwerk will fade out the releases of some samples when you play fast with heavy registrations, or even tutti.

I don’t know how much you do inside the computer, however the i7-3820 can be replaced with a much more powerful CPU without it needs to cost a fortune.
Intel Xeon e5-26xx processors socked LGA 2011 will fit in your motherboard, and although they were very expensive form new some years ago, occasionally they can now be found cheap on eBay and other second hand markets. Most Xeons are clocked low intended for servers, but some Xeons are made for workstations and have a relatively high clock frequency and do fine for Hauptwerk.
Good options could be an 8 core 16 threads e5-2687w, an e5-2687w v2 or a 10 core 20 threads e5-2690 v2.
Before buying of course make sure that the CPU is supported by the motherboard. If the BIOS has to be updated to support the Xeon CPU, do that before exchanging the processors.

Recently I got an e5-2687w, have clocked it to 3.57 GHz and with buffer size 1024 the computer delivers 20.000 static polyphony, 9-10.000 for the sample set.

Best regards
Johannes
Offline

ludu

Member

  • Posts: 999
  • Joined: Wed Nov 02, 2011 8:52 am
  • Location: Tournai (Belgium)

Re: Görlitz - Polyphony

PostSat Sep 16, 2017 12:06 pm

Hello Johannes,

Many thanks for your very interesting comment. I will buy a new console soon with a computer integrated. I will probably contact you when I receive my offer and see with you if everything is coherent, if you please.

Kind regards.
Luc
Offline

josq

Member

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

Re: Görlitz - Polyphony

PostSat Sep 16, 2017 1:38 pm

By the way, starting from HW 4.1, polyphony is not that critical anymore. Quoting from te 4.1 release notice:

Audio engine performance has been improved very significantly, so that there's much less noticeable
'drying up', much less CPU load when the polyphony limit is exceeded, significantly less polyphony needed
overall for wet sample sets for realistic results, better load-balancing, and polyphony should never run out
completely in times of continuous extreme polyphony overload. Overall, a significantly less powerful
computer (CPU) should be needed to achieve realistic results, especially for wet sample sets. Polyphony-restricted Hauptwerk editions (the Free and Basic Editions) should now be capable of much more realistic
results with wet sample sets, and with larger registrations. Important note: you might need to re-adjust
your polyphony limit settings to get optimal results for each organ (since release samples are now faded
more gently and progressively, in some cases you might find you even need to reduce polyphony limit
settings a little to prevent CPU overload, but you should still get very significantly better and more
realistic results overall).
Offline

1961TC4ME

Member

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

Re: Görlitz - Polyphony

PostSat Sep 16, 2017 2:46 pm

A few additional things to consider with these 6 channel sets (and I'm sure someone will correct me if I'm wrong), is the fact of how you use them. If you are adjusting the fader to mostly wet and like the sound that way, you might consider not even loading the direct portion of the set. I'd assume this alone would not only speed up load times, but would also take some strain off the computer and allow you to increase polyphony. If I were to truncate anything and still load all 6 channels, I'd probably be looking more at truncating either the diffuse or the rear surround portion of the set which would likely take more load off the CPU and allow you to increase polyphony vs. truncating the direct part of the set which is much dryer and requires less polyphony in the first place.

Overall, I'd prefer not to truncate at all and I never do (I regard it as cutting corners and cheating), and stray from the original sampling, as in my opinion the results are only more artificial.

Marc
Offline
User avatar

mdyde

Moderator

  • Posts: 15446
  • Joined: Fri Mar 14, 2003 1:19 pm
  • Location: UK

Re: Görlitz - Polyphony

PostSun Sep 17, 2017 4:48 am

josq wrote:By the way, starting from HW 4.1, polyphony is not that critical anymore. Quoting from te 4.1 release notice:


Just to clarify, it's still essential that the polyphony limit is set appropriately (not too high, so as to avoid the audio-CPU meter going into the red and audio breaking up), but when set appropriately v4.1+ is much better at giving a more realistic result than earlier versions.
Best regards, Martin.
Hauptwerk software designer/developer, Milan Digital Audio.
Next

Return to Hauptwerk instruments

Who is online

Users browsing this forum: No registered users and 12 guests