Do different impulse loaders yield different results?

[NSGUITAR]

In the past couple weeks, I've been testing out different impulse loaders. I used to use Voxengo Boogex to load my cab and verb impulses, but recently tried out Sir2.


To me, Sir2 sounds 10x better, more realistic! I was surprised.. Isn't the loader's job JUST to load the impulse and nothing else? I'll post clips later today, but this is a honest question I have: Do different impulse loaders yield different results?

 

[Ascendant]

Could be that you had the wrong settings in boogex.

I haven't tried sir2 myself, only sir and it had a lot of delay so I didn't bother. Neiter have I tried to A/B different ir loaders, I used to think that they all yielded the same result (except kefir).

 

[Emdprodukt]

depends on your boogex settings. it has ampsimulation in it that needs to be turned off.

 

[AntonioPetrole]

I believe different impulse loaders have different algorithims, so some are made higher quality than others. Dont quote me on that though

 

[Pxz]

Pristine Space do sound different than Boogex for some reason ( with the ampsim control turned off)

 

[006]

In my experience they do all sound different, some really interesting results sometimes.

 

[schust]

i just ran through a test of a few different impulse loaders i have and boogex came out on top, for me.

here's what i tested, granted i don't have v2 of Sir - still only 1.

Sir v1, Le Cab, KeFir

definitely check the settings under 'Amplifier' on the boogex. by default i believe tone and dynamics are cranked. turn everything to 0 - off.

 

[espartaco]

mmm interesting , i used to use boogex ( with the ampsim control turned off too) and then i change to le cab, i do note a different though never try sir 2 for load guitar impulses. i though that only the paid version has zero latency??

 

[espartaco]

In my experience they do all sound different, some really interesting results sometimes.

but is there a different in quality??

 

[NSGUITAR]

I realized what my problem with boogex was:


There is a filter built in that I've never worried about fixing. It puts an automatic hipass and lowpass filter on the impulse.


Regarding Sir2.. I use the demo version and have no latency..

 

[Seth Munson]

but is there a different in quality??

there could be, my suggestion is trial and error and see what sounds best to you

 

[espartaco]

there could be, my suggestion is trial and error and see what sounds best to you

yeah definitly going to make some tests , this thread made me paranoid

 

[006]

Yes, for example, the ReValver loader sounds inferior to everything else I have tried.

 

[deLuther]

I`m checked LePou LeCab, Cubase 5 ReVerence, SIR2 and Voxengo Pristine Space against true mathematical convolution (with help of Wolfram Mathematica).
SIR2 and Pristine Space were almost exactly like mathematical convolution.
Difference with convolution in Mathematica was (maximum absolute difference value in db):
Pristine Space -143.755 (6.5e-8)
SIR2 -136.461 (1.5e-7)
ReVerence -43.7351 (6.5e-3)
LeCab -29.1458 (3.5e-2)

ps. Sample rate of impulse and audio was the same, I`m found that resampling algorithm in PS is slightly better than SIR2, so, for example, with 96k impulse and material with 44k PS sounded slightly better than SIR2 (but I`m not tested this case with Mathematica).

 

[XnaySaturo]

Yes, for example, the ReValver loader sounds inferior to everything else I have tried.

Does it straight up load impulses or it's like Guitar Rig with different settings and stuff? I honestly don't know, I've never used ReValver.

 

[kev]

Definitely, I had a mix going a few years ago with RiR- I tested pristine space which I suspect is much better and my carefully crafted tone fell apart.

 

[John_C]

I`m checked LePou LeCab, Cubase 5 ReVerence, SIR2 and Voxengo Pristine Space against true mathematical convolution (with help of Wolfram Mathematica).
SIR2 and Pristine Space were almost exactly like mathematical convolution.
Difference with convolution in Mathematica was (maximum absolute difference value in db):
Pristine Space -143.755 (6.5e-8)
SIR2 -136.461 (1.5e-7)
ReVerence -43.7351 (6.5e-3)
LeCab -29.1458 (3.5e-2)

ps. Sample rate of impulse and audio was the same, I`m found that resampling algorithm in PS is slightly better than SIR2, so, for example, with 96k impulse and material with 44k PS sounded slightly better than SIR2 (but I`m not tested this case with Mathematica).

How about a standard deviation? Max difference is ofc interesting but not the whole story.
p.s. an absolute difference can't be a negative number :nerdface:

edit: what's in the brackets? do you mean Pristine Space was out by a maxinum of 6.5x10^-8 dB? In which case what is number outside of the brackets.....

 

[deLuther]

I take Abs[a-b], where a, say, mathematical convolution, b - sample under testing, then for resulting array I found maximum value.
Value of difference is in brackets, negative number because I then compute 20*Log[10,Abs[a-b]].
So I posted 20*Log[10,Abs[a-b]] (Abs[a-b])

Standard Deviation will be
Pristine Space 9.6e-9
SIR2 2.5e-8
ReVerence 8.8e-4
LeCab 9.6e-3

Of course, LeCab was shifted to properly align with others (it has delay of around 200 samples).

 

[006]

Does it straight up load impulses or it's like Guitar Rig with different settings and stuff? I honestly don't know, I've never used ReValver.

You can do either way, there are tonal options in it for high and low end, "distortion" and "crunch," as well as "resample" and "hi-res." The thing is with them all set to "0" it sounds different, and shitty, in comparison to other loaders.

 

[darthjujuu]

kefir since day one, here

i think people just all too often forget to crank the mix knob to 100% since it defaults to 50%