Hi great maker of great software! ;-)
I just purchased R8Brain Pro for converting my new Lex 960 IR library (48/32FP) to 44.1 and 24 bit.
Do you recommend to use the minimum phase option? I mean, IRs are actually kind of filters/delay "scripts" (for lack of a better word). It's often hard to decide what the effect is of specific EQ-ing or filtering on IR's themselves.
For instance, I applied a lot of Waves LinEQ low cuts to get rid of a lot of low energy. The result sometimes makes the IRs "look" weird (negative slow ripples around positive peaks), but this is actually adding FIR filtering to the IR. The music mix indeed also gets rid of low freq energy this way. (Glad I read some articles on how FIR EQ filters look!)
a) Do I have a risk that the minimum phase ("DA/AD") option for my IRs will affect the music when convolution is applied?
b) Today I read some notes from Bob Katz, saying that he often upsamples to 96 Khz for mastering and then downsamples again, saying that DSP in 96 gives smoother results, even with the additional up/down sampling. Can you imagine this is true? Would this also have worked for processing my raw IRs? They often have low levels when coming out of the Deconvolver, even with 30 second sweeps... I already found out that increasing the volume before EQ-ing gives a lot better result than first EQ-ing these low level IRs. So, the DSP order can be quite important on editing and processing IRs.
Thanks again for your great products!
Peter Emanuel Roos
www.PeterRoos.com / www.Samplicity.com (IR libs)
Minimum-phase should sound OK. It may only start to sound a bit 'adjusted', but not in a bad way for sure.
Any usual DSP works better at higher sample rates - distortion works better, EQs work more precisely, chorusing sounds finer, etc. However, if you are applying a linear-phase filter, sample rate is usually irrelevant because FIR filters are usually designed in 'ideal' form (straight from analog domain, for example).
I also do not understand why your IRs sound better if you normalize their level first. Maybe EQ you are using is a bit 'analog' in behavior (i.e. it has non-linear components) - if so, I can understand this tactic.
I just did a 32-bit float 48KHz to 32-bit float 44.1K conversion using minimum phase and ultra steep, and I think I can hear some strange artifacts. Sounds like a third harmonic/melody or something added the track at around 2:50 when all instruments come in.
Familiar with this ?
Resampling cannot produce harmonics, so I'm not sure what you are talking about. If you think minimum-phase mode does not work well for your track, simply switch to linear-phase mode.
Or, can you post an original track as well?
I am not sure what I am hearing, but can't harmonics be created by aliasing?
Of course, it could be the mp3 conversion too, allthough fraunhofer 320kbs is usually pretty safe.
Anyway, r8brain pro is pretty neat.