splitting bassline

[regretfullySaid]

Any dynamics plugin by nature is going to color the signal.

[djspirit]

wow, I never realised just how much the Multiband Dynamics plugin colors the signal! I always presumed that it would be flat.

Did you watch this video?

http://www.quantizecourses.com/pages.php/?p=895

Presumably, if you turn the dynamics "amount" knob to 0, the signal is flat. (see around 1:20 in the video)

[jpga]

This is all you need!
http://rhythminmind.net/1313/?p=43
4 bands transparent splitter rack
The genius of your rack is that you can use your own favorite compresorer on the different bands. I have made a multi-band compressor with the glue from Cytomic and it sounds absolutely wonderful....Thank's again

All the best..

Jp

[jpga]

Don't miss this.

[djspirit]

Don't miss this.

Thanks JPGA!

I'm still waiting for a reply to this post to Rhythm Mind:
"I'm a newbie and don't know what you mean by transparent. Please provide a quick explanation and what advantage it offers? Also, someone wrote a comment on this page on your web site identifying a bug, "in the 24 db freq dividers the low mid and low x-over chains are swapped at the output, means only wrong named but causes little bit confusion which is the low mid and the low". Are you going to release a fix for that? Also, he asks you to explain the advantage of the 24db dividers."

Can you help with any of this?

Thanks!

[jpga]

Don't miss this.

Thanks JPGA!

I'm still waiting for a reply to this post to Rhythm Mind:
"I'm a newbie and don't know what you mean by transparent. Please provide a quick explanation and what advantage it offers? Also, someone wrote a comment on this page on your web site identifying a bug, "in the 24 db freq dividers the low mid and low x-over chains are swapped at the output, means only wrong named but causes little bit confusion which is the low mid and the low". Are you going to release a fix for that? Also, he asks you to explain the advantage of the 24db dividers."

Can you help with any of this?

Thanks!

When it is transparent there is no sound change when the rack is on or off you can color your sound as you want with your favorit tools.. I made multiband compressor with the glue that rocks...

Hope it help you.

All the best...

Jp

[Vios]

One thing to note is that a frequency crossover is never transparent. In the analog world, a capacitor or inductor that make up a crossover cause phase shift starting about an octave above the crossover point. In the digital world, the algorithms that cause a separation of the frequencies have a similar effect. Try putting an EQ3 after a limiter and you'll immediately see what I mean - it probably won't sound much or any different, but you'll see your signal spiking above the limiting point of the limiter.

[djspirit]

One thing to note is that a frequency crossover is never transparent. In the analog world, a capacitor or inductor that make up a crossover cause phase shift starting about an octave above the crossover point. In the digital world, the algorithms that cause a separation of the frequencies have a similar effect. Try putting an EQ3 after a limiter and you'll immediately see what I mean - it probably won't sound much or any different, but you'll see your signal spiking above the limiting point of the limiter.

I ran an experiment using this approach to splitting frequencies to check your claim:
http://www.quantizecourses.com/pages.php/?p=895

I set up three cases of Multiband Dynamics in a rack and turned the "amount" on all of them to "0" (which turns off all compression and other effects but leaves the crossovers intact). I solo'd one case for low, one for mid, and one for high. Then I dropped in a limiter after the rack. I loaded up an uncompressed clip (some glitch hop) to get good peaks. Low and behold, the limiter kicks in at the same exact gain point whether the frequency splitting rack is on or off.

How do you explain that? I'm really just trying to understand whether this method for applying effects to a limited frequency band is better or worse than other approaches.

[rhythminmind]

One thing to note is that a frequency crossover is never transparent. In the analog world, a capacitor or inductor that make up a crossover cause phase shift starting about an octave above the crossover point. In the digital world, the algorithms that cause a separation of the frequencies have a similar effect. Try putting an EQ3 after a limiter and you'll immediately see what I mean - it probably won't sound much or any different, but you'll see your signal spiking above the limiting point of the limiter.

Not completely true. You are correct that the filtering creates phase shifts, but they can easily be reversed/compensated for. This crossover divider is completely transparent & is actually based on an analog technique. Speaker crossovers often use this method. It's a really simple concept, You use a Low pass (or highpass) to create a signal band. In this case a low pass, & you end up with a "low" filtered band. You then mix a polarity flipped mult (copy) of that signal with original source. This cancels out Low information & gives you the "High" band. Thats it, repeat the process for as many bands as you need.
Most multiband processing will use a irr or fir filtering method, they have advantages for steep filtering, but in a mastering situation or any situation you want 100% transparency my example is the way to go.

@djspirit Audio transparency = A signal not being modified in any way. You might enjoy polarity null testing signals to learn what processors change. Just send the output of the signal in question to a bus, & send the original unmodified signal to the same bus but reverse the polarity before hand. With live use the utility tool for this task. This will show you the differences of the two signals. (I stopped using Live after ver 7/8 so I'm not sure if they have added a polarity flop to the mixer or not. If so you can use that over the utility tool)

[djspirit]

@djspirit Audio transparency = A signal not being modified in any way. You might enjoy polarity null testing signals to learn what processors change. Just send the output of the signal in question to a bus, & send the original unmodified signal to the same bus but reverse the polarity before hand. With live use the utility tool for this task. This will show you the differences of the two signals. (I stopped using Live after ver 7/8 so I'm not sure if they have added a polarity flop to the mixer or not. If so you can use that over the utility tool)

Thanks Rhythminmind! That helps a lot. I will see if I can do that testing.

Now, would you answer the queries on your web site: "in the 24 db freq dividers the low mid and low x-over chains are swapped at the output, means only wrong named but causes little bit confusion which is the low mid and the low". Are you going to release a fix for that? Also, he asks you to explain the advantage of the 24db dividers."

Thanks!

[rhythminmind]

I no longer use or have a Live installation to check it out. I created these in 2007/8 or so. It's easy enough for people to check/fix if needed. 24db filters are just stepper then the 12db.

[djspirit]

You might enjoy polarity null testing signals to learn what processors change. Just send the output of the signal in question to a bus, & send the original unmodified signal to the same bus but reverse the polarity before hand. With live use the utility tool for this task. This will show you the differences of the two signals. (I stopped using Live after ver 7/8 so I'm not sure if they have added a polarity flop to the mixer or not. If so you can use that over the utility tool)

I just tested this out using the multiband dynamics approach to creating a three-way frequency band (see http://www.quantizecourses.com/pages.php/?p=895)

The signals cancelled each other completely. I guess that means this approach is transparent! (and for me less complex than using the racks you built).

DJ Spirit

[hdrpbx]

i did little test aswell and the multiband compressor splitter doesnt really seem to cancel out with a dry inverted phase signal

just take a look at this project

http://ge.tt/8W77fOC?c

i set up two channels with the same drumloop

on the first you have a audio effect rack with the multiband comp splitter (i used the same method explained in the video) on one chain and a phase inverter on the other set to mute. When you activate the inv phase chain you will see, or better hear, that the two signals dont cancel each other out, whereas if you activate the inv phase chain on the other audio channel,where i used the linear 3 band splitter provided in the rhytminmind package, you get total silence

and honestly it's not even necessary to run a test this like to demonstrate it, you can actually hear it.
you even said yourself on your first post that a multibandcomp splitter colors the sound a bit, wich it's not necesarilly a bad thing ..in some situations it could be useful but if you need a linear splitter the multiband approach is not really the best imo

[mharris]

Any dynamics plugin by nature is going to color the signal.

We are talking about using the multiband dynamics plugin with all dynamics processing turned off for the purposes of creating a 3 band crossover.

Did you watch this video?

http://www.quantizecourses.com/pages.php/?p=895

Presumably, if you turn the dynamics "amount" knob to 0, the signal is flat. (see around 1:20 in the video)

It still doesn't phase cancel with an un-processed signal even with the Amount set to 0%

[djspirit]

i did little test aswell and the multiband compressor splitter doesnt really seem to cancel out with a dry inverted phase signal

just take a look at this project

http://ge.tt/8W77fOC?c

i set up two channels with the same drumloop

on the first you have a audio effect rack with the multiband comp splitter (i used the same method explained in the video) on one chain and a phase inverter on the other set to mute. When you activate the inv phase chain you will see, or better hear, that the two signals dont cancel each other out, whereas if you activate the inv phase chain on the other audio channel,where i used the linear 3 band splitter provided in the rhytminmind package, you get total silence

and honestly it's not even necessary to run a test this like to demonstrate it, you can actually hear it.
you even said yourself on your first post that a multibandcomp splitter colors the sound a bit, wich it's not necesarilly a bad thing ..in some situations it could be useful but if you need a linear splitter the multiband approach is not really the best imo

Thanks for the help, but I am as confused as ever. I admit that I did say in my first post that the multiband dynamics approach seems to color the sound, but it was subtle and my ears play tricks sometimes, so I'm no longer sure.

I downloaded an installed your test project. (Please keep in mind I am an Ableton newbie.) The signal difference on your multiband dynamics version seems way much too me (between the dry and inverted signal). It's hotter than the dry signal alone--making the master meter peak at a higher point. Can you explain this? Are you sure you can just drop in an inversion on a rack like that to get a perfect inversion of the other chain?

I set up my test differently, by taking audio in from the multiband dynamics channel and applying the inversion to a separate channel. By toggling the inversion channel on and off I get silence. Can you explain why this cancels the signal?

Here's my project:
http://www.yousendit.com/download/T2dkM ... bUs1aWNUQw

Thanks for the help!