Is there a such thing as the "best" IR Loader?
- Thread starter dboy1612
- Start date
atoragon
Member
DullElysium
Member
You will not notice difference between SIR2 and Voxengo Pristine Space, because difference lies beyond 24 bits
KeFIR has zero latency due to direct convolution. More efficient fast convolution, on other hand, requires more intricate algorithms to provide zero latency, that`s why LeCAB has some latency due to simplistic approach, while NadIR (SIR2, Pristine Space) has no latency.
RiF
Member
Not necessarily "the best", but I am using mixIR2 from Redwirez because:
- you can easily mix multiple IRs, parallel (e.g. cabs) or in series (e.g. eq curves)
- it supports AAX-64 for Pro Tools 11+
- good enough IR-browser
- it simply works without any crashes or hickups
- you can easily mix multiple IRs, parallel (e.g. cabs) or in series (e.g. eq curves)
- it supports AAX-64 for Pro Tools 11+
- good enough IR-browser
- it simply works without any crashes or hickups
The only thing that I not like in Pristine Space - it`s Zero Latency mechanism, Zero Latency setting can work bad if interface buffer length is not power of 2, plugin buffer shorter than host buffer is not very efficient method because you will need adjust buffer closer to host buffer length for maximum efficiency.
KeFIR is also good in terms of precision, but eats CPU with moderate impulse length.
My old test table:
My terms about good convolver:
1) Precision comparable to mathematical software
2) Zero latency
3) "Reverberation-grade" i.e. low consumption with impulses of several second length
4) Multirate i.e. quality automatic impulse resampling to different project sample rate.
I`m solve 1) and 2) in my own convolver (PFF Convolver) but currently have troubles with CPU consumption with large impulses (around 1 second is comparable to SIR2 or Pristine Space) and working on quality resampling
KeFIR is also good in terms of precision, but eats CPU with moderate impulse length.
My old test table:
My terms about good convolver:
1) Precision comparable to mathematical software
2) Zero latency
3) "Reverberation-grade" i.e. low consumption with impulses of several second length
4) Multirate i.e. quality automatic impulse resampling to different project sample rate.
I`m solve 1) and 2) in my own convolver (PFF Convolver) but currently have troubles with CPU consumption with large impulses (around 1 second is comparable to SIR2 or Pristine Space) and working on quality resampling
The only thing that I not like in Pristine Space - it`s Zero Latency mechanism, Zero Latency setting can work bad if interface buffer length is not power of 2, plugin buffer shorter than host buffer is not very efficient method because you will need adjust buffer closer to host buffer length for maximum efficiency.
KeFIR is also good in terms of precision, but eats CPU with moderate impulse length.
My old test table:
My terms about good convolver:
1) Precision comparable to mathematical software
2) Zero latency
3) "Reverberation-grade" i.e. low consumption with impulses of several second length
4) Multirate i.e. quality automatic impulse resampling to different project sample rate.
I`m solve 1) and 2) in my own convolver (PFF Convolver) but currently have troubles with CPU consumption with large impulses (around 1 second is comparable to SIR2 or Pristine Space) and working on quality resampling
Cool, very useful data!
Have you tried the Twonotes product? http://www.two-notes.com/en/software/torpedo-blendir/ http://www.two-notes.com/en/software/torpedo-wall-of-sound-3/
There is one problem in this table - Overloud TH2 always have active filters in convolver module (hi- and low-pass), so it gives average result because of difference due to filtering. It is also the reason why I`m not included NadIR in this table when it come - you can not compare it directly to quality convolution (without additional processing). Also LeCab (both version) were compared with account to introduced delay (delay was removed before comparison).
For the time being I`m mostly developing my own software than testing other
For example, instead of blendir approach, I`m prefer matching, but with precision beyond Voxengo CurveEQ or Ozone.
(look for MatcherCL topic there, but new version with GUI is currently under development).
For the time being I`m mostly developing my own software than testing other
For example, instead of blendir approach, I`m prefer matching, but with precision beyond Voxengo CurveEQ or Ozone.
(look for MatcherCL topic there, but new version with GUI is currently under development).
I like SIR2 a lot of for finalizing new IRs, particularly when I'm trying to use multiple cabinets and I want to widen them a bit or widen a reverb effect or something. It seems to do more than simply widen, but also balances the EQ curve for my left and right ear so if I was a little off due to ear strain, I can go back and fix it sort of, debatable as to retaining dynamics, but has some useful bonus features like that.
XxSicRokerxX
Gabriel R.
Similar threads
