Unit impulse response for adding a delay?

Hauptwerk software technical support only. Please make sure you have read the manual, tutorials and FAQ pages before requesting support.
Post Reply
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Unit impulse response for adding a delay?

Post by mnailor »

Does anyone know where I can download (buy) a unit impulse response (IR) in Hauptwerk format, to pass the original signals unchanged?

This would be to add a configurable pre-delay (in the IR settings for the output bus) to some speakers without actually adding any reverb.

I think somebody mentioned doing this, but I can't find the thread.
Thank you.
User avatar
mdyde
Moderator
Posts: 15695
Joined: Fri Mar 14, 2003 1:19 pm

Re: Unit impulse response for adding a delay?

Post by mdyde »

Hello Mark,

If I recall correctly, Jerry Martin created one:

https://www.forum.hauptwerk.com/viewtop ... 69#p155268
https://www.forum.hauptwerk.com/viewtop ... 62#p157762

(You could then use the "Wet pre-delay (ms)" setting on the mixer to adjust the delay.)
Best regards, Martin.
Hauptwerk software designer/developer, Milan Digital Audio.
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

Thank you!

(Unfortunately, the minimum delay built into those IRs is 30 ms. I'd like to start with 0 delay and add 1 - 10 ms using the predelay setting.)
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

I did some testing with Jerry Martin's 30, 50, 80 ms delay IRs, adding delays to rear channels and front Pedal only. That opened up the room's spacial impression, even if it isn't exactly what I was after. Thanks!
User avatar
mdyde
Moderator
Posts: 15695
Joined: Fri Mar 14, 2003 1:19 pm

Re: Unit impulse response for adding a delay?

Post by mdyde »

Thanks, Mark.

If looking for a zero-length delay in the reverb file (so as to be able to adjust it to any delay value via the pre-delay setting on the mixer screen), you could perhaps copy/edit one of Jerry .wav files to remove its sample frames before the unit impulse, and also edit its ....ImpulseResponseReverbDefinition_Hauptwerk_xml file to set IncludedPredelayMs=0 and reduce the TotalLengthMs accordingly.
Best regards, Martin.
Hauptwerk software designer/developer, Milan Digital Audio.
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

I may try that. Audacity seems to be free. Thank you.
User avatar
mdyde
Moderator
Posts: 15695
Joined: Fri Mar 14, 2003 1:19 pm

Re: Unit impulse response for adding a delay?

Post by mdyde »

Thanks, Mark. Yes -- I think you should be able to edit them in Audacity (and it's free).
Best regards, Martin.
Hauptwerk software designer/developer, Milan Digital Audio.
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

It was easy to delete the initial delay before the impulse using Audacity. The only trick was not knowing how to make it export the same 4 track format. This works well, and the delay on each bus is set by Hauptwerk's IR predelay parameter. Thanks!
User avatar
mdyde
Moderator
Posts: 15695
Joined: Fri Mar 14, 2003 1:19 pm

Re: Unit impulse response for adding a delay?

Post by mdyde »

Thanks, Mark.

Excellent.
Best regards, Martin.
Hauptwerk software designer/developer, Milan Digital Audio.
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

Notes on usage:

Set the wetness scalar to 100% on organs for which you want the predelays. In my case, that's physically larger organs with a tall, deep, wide case or multiple chambers.

Set the wetness scalar to 0% on organs where you don't want delays.

Don't set the wetness scalar in between 0 and 100. That causes a double attack, which you can hear plainly by setting the predelay to 500 ms and wetness scalar to 50%.

In the front, I'm using a group of delayed buses and a group of non-delayed buses on the same speakers so some manual divisions can be more distant (slightly delayed) than others. For example, Great in your face, Swell farther away. Pedals and rear speakers are always delayed, if wetness scalar isn't 0.
Last edited by mnailor on Wed Oct 02, 2024 4:13 am, edited 1 time in total.
User avatar
jerrymartin
Member
Posts: 113
Joined: Sun Oct 30, 2005 1:01 pm

Re: Unit impulse response for adding a delay?

Post by jerrymartin »

Thank you for this great suggestion and development Mark and Martin!
I had not previously conceived that one could adjust this using wet pre-delay - hence having the baked-in delay of 30,50,80....

Here is a download link to ah updated set of IR files which includes my prior delay, but also one with the no delay, no reverb - which can be used in this way.

Super!

https://www.dropbox.com/scl/fi/x67un8vb ... 6FkFA&dl=0
Jeremiah Martin,
Portsmouth, Ohio
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

Thank you, Jerry!
mnailor
Member
Posts: 1721
Joined: Sat Nov 30, 2013 5:57 pm

Re: Unit impulse response for adding a delay?

Post by mnailor »

More notes on how to use this. Remember that, for organs that don't need delays added between divisions or ranks, you turn the delay IRs off by setting the organ's wetness scalar to 0 instead of 100. I keep that mini control panel visible. There's no significant convolution processing overhead at 0.

It's easy to go too far and create an unnatural effect. 20 ms delay on front channels between two divisions can be too much, producing an audible double attack on punchy stops like trumpets. But 20 ms is plausible for an Echo division.

I ended up with 4 front manual groups using the same 6 stereo speaker pairs, each with a delay range: 0-5 ms, 5-10 ms, 10-15 ms, and 15-20 ms. The stereo pairs' delays vary in 1 ms increments so they don't all speak exactly at once.

Edit: Later I changed that to only have a 1 ms range within the group, so 0 - 5 ms became 0 - 1 ms with alternating delays, to keep more separation between the distance ranges of the groups. Will tweak it more as I try it out over time. This is all just an example anyway.

The front pedal group has 20 - 25 ms delays. The rear manual group is 30 - 40 ms, and the rear pedal group is 50 ms.

For samplesets with some built in pipe delays, like Armley, Landau, and Goch, I send front manual ranks to the 0 - 5 ms and 5 - 10 ms groups, which keeps the effect subtle. Or turn it off.

For organs large enough to spread the pipes out along the listener's line of sight, so there should be delays between pipe attacks reaching your ears, I route the manuals to different front delay groups based on a guess about where the divisions are placed in the organ (some are documented). So the manuals may be placed at 0-5, 5-10, and 10-15 ms, spreading the attacks out over a 15 ms time interval, corresponding to about 5 m distance from the closest pipes to the farthest along the listener's line of sight.

A 10 ft or 3 m distance between pipes along your line of sight is about 9 ms of time delay from the closest pipe attack to the farthest. That is a 3D distance, not just along the floor. So a small case may only have distances in the 0-5 ms range from a given listening position, while a large or multi-chambered instrument may need 15-20 ms or more.

There's an enhancement request for user-settable pipe delays in Rank Voicing, which would make this workaround unnecessary in a future HW version.

The best thing would be for sampleset producers to estimate line of sight distances from their mic placements to the windchests and include the time delays, starting from the closest pipes' attacks, in samplesets. That would give a more accurate spatial picture of the organ. An option to disable pipe delays would be needed for some people. PG has this feature on some organs, so it can be done.

(The most accurate way would be to time the attack time from key press on every sample, then subtract the shortest time from all of them. That would require hardware and software support that probably isn't available in the sampling systems.)
Post Reply