Hi Jörn, list
On 05/05/2016 13:32, Jörn Nettingsmeier wrote:
On 05/04/2016 10:00 PM, Alessio Degani wrote:
[CUT]
If I have understood the previous posts correctly, a tilt EQ is even
simpler than two shelves: just use a single one at the desired tilt
frequency and adjust the overall gain, like Fons mentioned: if you
dial 3dB treble boost, applying -1.5dB of total gain should give you
the exact equivalent of a tilt EQ.
more ore less yes
But I doubt this is what you want. Instead, I'm
sure you'd want to
match the perceived loudness to be the same as before the correction.
A tilt EQ is very unlikely to achieve that, so manual gain correction
will be needed in any case.
I do not need loudness/gain/energy correction
Hence, why bother with a tilt EQ?
I can see the appeal for DJs, but not for post production
My use of tilt EQ does
not match the "DJ case" nor the post production case.
Let me explain. I've used tilt EQ once, or twice in studio (with Mac
OSX) only beacuse an EQ plugin that I've used (I can't remember the name
of the plugin right now) has a "tilt" option among the other classic
parametric choices. I've used the "tilt" only just for try it :) That's
why I know the existence of this EQ "type".
At the moment, my use case is different from he DJ case or
post-production (in the "music" sense) case.
I'm using some custom made contact microphone that shows different
spectral response characteristics depending on different factors like
mounting surface, mounting method (i.e. adhesive foil, clamps, ...). The
contact microphone are used for measuring purposes, and not, for
example, for recording an acoustic instruments.
For what I've seen, a tilt EQ is sufficient for a "good" correction
without loosing to much time.
The correct way to proceed in my case, should be a calibration with a
matched EQ curve (i.e. whitening the response). The problem is that this
matched curve can be very different from one measure to another. A
simple way to control the balance between low and high frequency, such
as shelving or tilt is the way to go (for now... maybe in the future I
will find a simple way to do this automatically :) ).
The loudness is no my concern, because I apply a feature normalization.
Cheers
--
a.