Page 2 of 3

Re: LADSPA plugins

Posted: Sun Feb 17, 2008 11:45 pm
by dahnielson
Actually I'm having a perverse idea of creating a LADSPA convoluton-reverb-construction-kit that takes a jconv config file as input and builds a plugin with the response compiled into it that dynamically links to the zita-convolver. No more external config or response files.

There is a convolver called dssi_convolver implemented (as you might have guessed) as a DSSI plugin so it can have its own GUI. There's also a JACK version of it called jack_convolver (I guess). But I can't either of them to build at the moment.

Re: LADSPA plugins

Posted: Mon Feb 18, 2008 11:27 am
by Alex
Anders,

I couldn't build either of these either. I wonder if the jack convolver hits a problem with using jackdmp instead of vanilla jackd.
I'm not smart enough with linux to know where to go from here.

Alex.

Re: LADSPA plugins

Posted: Mon Feb 18, 2008 1:21 pm
by dahnielson
Alex wrote:Anders,

I couldn't build either of these either. I wonder if the jack convolver hits a problem with using jackdmp instead of vanilla jackd.
I'm not smart enough with linux to know where to go from here.

Alex.
I have received a patch for them but not yet applied it. Will report back when I've tried it.

Re: LADSPA plugins

Posted: Mon Feb 18, 2008 1:27 pm
by Alex
Thanks Anders. I would appreciate some feedback on this.

Alex.

:)

Re: LADSPA plugins

Posted: Tue Feb 19, 2008 9:14 pm
by dahnielson
Alex wrote:Thanks Anders. I would appreciate some feedback on this.

Alex.

:)
Ok I've managed to build dssi_convolve on my Gentoo Linux system as a regular. Forwarded the fixed ebuild to you in case the it's useful to you in some way.

In the meantime I've actually grown fond of JConv. Once you found an IR you like you usually stick to it and just leave JConv running once set up. And in case there are a couple different IRs to owns liking it doesn't hurt to set up a couple of channels instead of just a stereo pair.

Re: LADSPA plugins

Posted: Wed Feb 20, 2008 3:13 pm
by Alex
Anders, thank you for the ebuild!

Alex.
8-)

Re: Reverb

Posted: Fri Aug 01, 2008 5:39 am
by ccherrett
I use CPlate Reverb in ardour. It was recommended for vocals but we have since moved on to use it for orchestral instruments and are very happy with it!

Re: Reverb

Posted: Tue Sep 02, 2008 9:09 pm
by dahnielson
External hardware reverb
I just want to give a shout out to traditional external hardware reverbs. I think they're highly underrated by many in this convolving day and age, at least by orchestra users. Sure, if you write symphonic or chamber music and want your tracks to sound exactly like they were being played at Amsterdam's Concertgebouw or Vienna's Musikverein then convolution is the way to go.

But convolution can actually be a dead end for cinematic tracks. Most film scores are recorded fairly dry on recording stages, usually with multiple close microphones that are balanced in the mix -- unlike the lonely stereo pair in a concert hall that use the orchestras natural balance. The recording stage provide early reflections but the reverb tail is added in the mix by a reverb. Lexicon 960L (and it's predecessor 480L) is pretty much the standard for film scoring and mixing.

So there isn't anything inherently wrong about using external reverbs like Lexicon PCM91/PCM90 or T.C. Electronics M-One/M-OneXL if you can afford them or a budget reverb of good quality for your sample based orchestra music. That's why it's a good idea to have a low-latency audio interface with more in/out ports than a single stereo pair so that you can hook it up as an external send effect. (Most quality reverbs and effect processor have S/PDIF that you can use to interface with them.)

Of course, if you really want to get the sound of a 960L there's always IRs of it to buy (e.g. from Samplicity) but then you loose the big benefit of using the hardware equivalent: editing the reverb settings.

That's why I have dusted of and begun using my old budget hardware reverb again.

Mixing technique
I also want to share a mixing technique that I picked up from the VSL SE tutorial videos. If the instrument signals you receive from LinuxSampler are already well balanced in terms of gain level: Insert a pre-fader send to the reverb, the fader can then be used to position the instrument depth wise. At 0dB it will practically be in your face (or at least in your lap) while being distant at -10dB and even more so at -20dB ...well you get the idea.

Another good idea is to pan the instruments (if they aren't already correctly positioned) using the MIDI pan setting in your sequencer so that LinuxSampler will pan the instrument signal for you.

Re: Reverb

Posted: Sun Mar 01, 2009 7:15 pm
by mazarick
What is your opinion on how the GigaStudio Convolution Reverb stacks up with the open source solutions? Based on what you are seeing, is there an open source solution that is 'on the right track' or has the right aproach? Would we be better off sending the dry audio signals somewhere else and back again or is there something that plugs into jack that works well enough?

To me, the current state of linux audio is like the start of a Dickens novel... "it was the best of times, it was the worst of times". At this very second, things are looking pretty difficult (no usb midi, confusing/nonexistant realtime kernel options, etc.) but there is a lot of hope on the horizon that some very major breakthroughs are getting ready to happen.

-Mike

Re: Reverb

Posted: Sun Mar 01, 2009 7:27 pm
by dahnielson
mazarick wrote:What is your opinion on how the GigaStudio Convolution Reverb stacks up with the open source solutions? Based on what you are seeing, is there an open source solution that is 'on the right track' or has the right aproach? Would we be better off sending the dry audio signals somewhere else and back again or is there something that plugs into jack that works well enough?
I have never used GigaStudio Convolution Reverb so I can't compare. But I do work with Jconv and loving it.
mazarick wrote:To me, the current state of linux audio is like the start of a Dickens novel... "it was the best of times, it was the worst of times". At this very second, things are looking pretty difficult (no usb midi, confusing/nonexistant realtime kernel options, etc.) but there is a lot of hope on the horizon that some very major breakthroughs are getting ready to happen.
Actually, I do have both USB MIDI and a real time kernel in my studio. It works very well if you stay with the 2.6.24 series (and Ingo's patches) while later kernels isn't suitable for low latency audio work. Hopefully this will be sorted out in the 2.6.29/30 series.