IRDust (prototype convolver)

[Halma]

IRDust is a simple prototype low-latency (64 samples) true-stereo convolution plugin.

Wow! Low-latency & True Stereo? Now I am really interested...

Well, 128 samples currently to allow for the linear-phase resampling.

But.. yeah.. I'm not a big fan of low latency, and 2x2 matrix stereo allows some optimizations over running two copies of mono-to-stereo engines, so.. why not?

But that is low latency enough for me compared to the fixed latency of 8960 samples from free SIR1 (I know, that´s maybe a "BIGGER" reverb with lots of other features and quite good for being free - don´t wanna argue about quality or features but was only comparing it`s latency). Not quite sure about ReaVerb from Reaper which also has a low latency mode. Will check that over the weekend too. But REAVerb again is not True Stereo by default (iirc) and you have to set it up manually for that inside Reaper. Now I am interested in this topic.

Regards
Sebastian

[Halma]

Never thought about if it would be possible to have an option in having 3 different stereo inputs in one reverb each with its own predelay time and if the idea make sense in general...

I have thought about it. I have thought about it a lot, because it would be something that would make a lot of sense in a algorithmic reverb that allows more realistic positioning of individual sources, but can combine the processing cost of several sources in the same space. But the DAW setup mess is exactly why I've not really considered it for reverbs that only do levels and pre-delay. IMHO it's easier to just setup a few sends that do the delays and feed to the actual reverb.

Yeah that was "my fear" too. But what about "two workarounds":

1) a plugin that can be used at the end of a chain as an insert which acts as a "send channel" that communicates with the reverb and offers 3 different send modes: 1 2 3 like the 3 cue busses which then will be routed into the reverb. Problems: not sure where the insert channel in the different DAW sits in the signal flow (prefader, postfader etc).

or

2) we use the oldschool cuebus method BUT get rid of the delay andt insert a plugin which then routes the audio into the selected reverb input channel. A simple 1, 2 or 3 knob and the rest could be done within the reverb gui.

Dangit, some cool stuff could be done that way. Like having an additional small "specialized" spatial channelstrip after each input for some cool modulation. Or diffuse the different delayed signals differently...anyway I think I am going to get a little bit offtopic now.

Looking forward in testing this device a little bit.

Regards
Sebastian

[Lee702]

Thanks works great. Fills the void for me because I am switching to 64bit.

[mystran]

So any bugs found, or anything else I should add (besides pre-delay, I guess), before I declare this thing to be "stable" and call it 1.0?

[Halma]

So any bugs found, or anything else I should add (besides pre-delay, I guess), before I declare this thing to be "stable" and call it 1.0?

I tried it 32 bit inside Reaper and it ran stable and luckily pretty lighweight (maybe 1-2% cpu). The only thing I wish it could do too is also importing *.aiff files. Tbh the first thing I tried was loading some new IRs that I have found but I could not because they were *.aiff. Not that I need it that much but if I wanna use those IRs then I need to recode them to *.wav. So I ended up using REAverb for them.

Regards
Sebastian

[mystran]

So any bugs found, or anything else I should add (besides pre-delay, I guess), before I declare this thing to be "stable" and call it 1.0?

I tried it 32 bit inside Reaper and it ran stable and luckily pretty lighweight (maybe 1-2% cpu). The only thing I wish it could do too is also importing *.aiff files. Tbh the first thing I tried was loading some new IRs that I have found but I could not because they were *.aiff. Not that I need it that much but if I wanna use those IRs then I need to recode them to *.wav. So I ended up using REAverb for them.

Additional file formats should be possible, I'll add .aiff to the list of things to add.

[Halma]

[Lee702]

Ok in Studio One 2 64bit when I load the plugin or switch ir's I get a cpu spike of 100%.

[mystran]

Ok in Studio One 2 64bit when I load the plugin or switch ir's I get a cpu spike of 100%.

Yeah, part of the IR conversion is currently synchronous so the audio thread waits, as this seemed like the most reliable approach. It's a known issue that I'll try to reduce at some point; I'd certainly like to have smooth IR switching eventually, but at the same time I don't want to rush it and risk crash bugs. Resets can also take some time (though less than setting up a new IR) and fixing that is a slightly higher priority.

[Lee702]

Ok thanks. Any plans on adding length adjustment knob?

[mystran]

Ok thanks. Any plans on adding length adjustment knob?

Possibly.

[mystran]

0.0.8:

- added predelay, separate from convolution so adjustable on the fly
- hopefully fixed resets so they no longer should cause any spiking

[tico-tico]

For me in the dual-mono mode IRDust uses the same CPU amount as Reverberate LE does (2%). But Reverberate LE is zero-latency.

I use ancient Athlon 64 x2, win 7 x32

FR: please add FLAC support

[mystran]

For me in the dual-mono mode IRDust uses the same CPU amount as Reverberate LE does (2%). But Reverberate LE is zero-latency.

I use ancient Athlon 64 x2, win 7 x32

FR: please add FLAC support

Sure, will consider FLAC as a future format to support.

As for latency, linear-phase IR resampling and zero-latency are unfortunately incompatible and that's where half the current latency comes from (it's a latency vs. quality trade-off, and I found 64 for be a reasonable compromise). The other half is convolution latency that I might eliminate at some point.

[tico-tico]

Sure, will consider FLAC as a future format to support.

Thank you!

As for latency, linear-phase IR resampling and zero-latency are unfortunately incompatible and that's where half the current latency comes from

I don't understand why you need to resample IR in realtime, but i believe you. But it's only for if IR's samplerate != project samplerate. Right?

Edit:
btw maybe just use libsndfile for impulses? it supports many audio file formats and api is quite simple