Why is convolver different sounding ..then others ?

I compared convolver with other reverbs and the difference is well …hm rather big …
I used a test sample in reaper …using reaper ir reverb , cakewalk perfect space and sir 2 .
All reverbs have automatic impulse gain reduction set to off , Only wet sound set to -20 db …
Convolver setting gain to 0db ( so no ir gain reduction )
All three reverbs sounds nearly identical …except convolver .
I loaded the wave files in renoise , track
headroom 0db , and compared with convolver …hear for yourself …
Difficult to describe but in some examples it seems like the sound of convolver is closer to the source (ir) .less distant …missing that certain depth of space

I am not saying it ain’t usable …but I wonder why the difference is so prominent , considering the other reverbs sound nearly identical .
Maybe convolver is analyzing only a certain part of the ir file making is realtime playable , but so are the other reverbs .
Just guessing .
And cpu wise …convolver performs roughly the same as the the other three .

Kraken ? ( I am giving you a hard time ain’t I ?)

here’s the songfile …

some compromise has been taken to allow lower latencies with a moderate cpu usage.

will take a look at your example asap.

Why not offer both and let the user chose?

the initial version of the convolver allowed selecting the latency, but has been removed to simplify everything.

will experiment some more with more a dense output as soon as i have the time frame to do this.


convolver uses roughly the same amount of cpu as the other real time convolution reverbs …
Sir2 is realtime …
cakewalk perfect space ,512 samples same as convolver …

Of course I meant 64 samples instead of 512.

forgive me if this is a dumb questions. i am pretty new to using convolution, only read about it in the past.

Why do these sound so different from each other?
here is a render of a guitar using SIR1 : http://ge.tt/9pSHYrW1/v/0
and here is the same sample using Renoise Convolver : http://ge.tt/2A2QYrW1/v/0

The IR I used for this test is in this pack http://www.voxengo.com/files/impulses/IMreverbs1.rar

I used all the same settings in each, and matched the volumes (-12db in renoise convolver)
Here is the XRNS for you to look at, in case I’ve missed some setting. : http://ge.tt/6SWdZrW1/v/0

Shouldn’t all convolvers at their basic level sound the same? Otherwise what’s the point in capturing an impulse response at all :P

I can only assume that this has been fixed in 3.0 final, right?

No it has not been fixed , and I don’t think this is a bug ,but a design choice (flaw ?)
I am verry happy the multitapdelay bug is fixed
I’ve done all the tests before

Reaper built in convolution ,knuffinke sir 2 , cakewalk perfect space …al these 3 sound 99´% identitical .
Convolver does not
All running at 64 samples latency ( near realtime )

Bottom line , convolver needs improvement

Very strange…I am really wondering why the final release has been so rushed…

Because they want to sell a shitload copies of redux …

Well, if that was true, they would have worked at least more on the sampler (which will be Redux). It must have some other reason.

Nopes, there simply is a tendency to hold Beta periods for not not too long. I don’t know the exact Beta period but i suspect something like 3 months. As Beta started December last year we were already up to four months.
I remember a discussion for a few years back where we had a very long beta period and there actually went in a long time where Taktik was quite exhausted from the debugging rampage (and meanwhile no new development).
The devs got to make a point somewhere, wrap it up and continue with a fresh start from there. Even if you don’t hear or notice anything, in the background they definitely don’t sit still aside from the few weeks of holiday every human needs once in a while.