I am playing around with the multi-band capability (in MXXX), trying to understand the behavior of the various band-types. They all make sense and perform as I would have expected except for Spectrum. I cannot find a meaningful demo situation where the behavior makes sense when solo-ing bands. Can anyone suggest a testing scenario that might make it clear how this thing is supposed to work.
Also, can anyone suggest a real-life scenario where this type of band separation might be practically applied?
Questions on Spectrum Multi-band
-
- KVRAF
- 2702 posts since 9 Jul, 2015 from UK
Basically imagine hundreds of bands, one for each frequency. Each of these bands is then ordered from quitest (left) to loudest (right)
Frequencies in noise normally don't have much of a sustained level dues to it's random nature. So the quiter frequencies (left) are normally the noisey part of the signal. Whereas the louder frequencies (right) are usually part of more meaningful part of the signal like the tone of a vocal for example.
Having the bands soloed sounds very unnatural and apart from creative stuff doesn't really have much use.
Due to the unnatural sounding separation i find it best to use it quite subtly.
One great use for this type of separation is for detection. For example an autotune would better detect the pitch of the vocal if it wasn't listening to all the noise also.
I made some presets called 'Multi-Mode' they are in almost every single plugin (easy screen). In these presets you can select the crossover type as spectrum. Then use the A/B parameter at the top, which basically distributes how much of the effect is present in either noise or tonal.
Please check them out on some tracks and you will get a quick understanding of how the spectrum crossovers can be used.
Highlights for me are reverbs and distortions.
Frequencies in noise normally don't have much of a sustained level dues to it's random nature. So the quiter frequencies (left) are normally the noisey part of the signal. Whereas the louder frequencies (right) are usually part of more meaningful part of the signal like the tone of a vocal for example.
Having the bands soloed sounds very unnatural and apart from creative stuff doesn't really have much use.
Due to the unnatural sounding separation i find it best to use it quite subtly.
One great use for this type of separation is for detection. For example an autotune would better detect the pitch of the vocal if it wasn't listening to all the noise also.
I made some presets called 'Multi-Mode' they are in almost every single plugin (easy screen). In these presets you can select the crossover type as spectrum. Then use the A/B parameter at the top, which basically distributes how much of the effect is present in either noise or tonal.
Please check them out on some tracks and you will get a quick understanding of how the spectrum crossovers can be used.
Highlights for me are reverbs and distortions.
Jason @ Melda Production
-
- KVRist
- 460 posts since 25 Jan, 2016
jmg8 - Could the Multi Mode preset be applied in MPS? I would imagine (but could very well be wrong) that it could be interesting to use for sound design. If it could be, and you are willing to do so, could you also show us how to use it, how to create ground-breaking sounds while using it and go through the marketing steps to make boatloads of cash from sales of them as well? I'll appreciate it greatly once the checks start to roll in!
-
- KVRAF
- 2702 posts since 9 Jul, 2015 from UK
Ha ha, Nice.
Well its a bit yes and no for this one. Firstly there is no easy screen in MPS so these is no existing preset. Although it can be done manually by using a crossover in the FX section. I will do a template for you to get you started.
It is great to use with the fx, I like to distort the noise more than the tonal for example. Have a play around.
But in my opinion the application for ground breaking sound design comes when you use the technique with samples rather than synth.
You see if you are sound designing with a synth you can pretty much make ANY sound from scratch and have full control over the noise and tonal elements.
But samples are already made and it is normally tricky to split that sample up into separate parts to process them individually. Commonly splitting is done just by frequency (normal multiband) but with these new crossovers spectrum and transient/tonal, we can get really creative!
Imagine using the noise of a car driving past and blending that with the sound of a man screaming! that would make for really interesting sounds for film or music sfx. Then throw Morph into the mix and we have some VERY unique sounds! I love playing around getting weird textures from samples like the noise of a match being struck and the transient of a penny dropping and the tone of a bottle being blown and then I subtly mix these in with my synth sounds to make things sound more interesting and organic.
Well its a bit yes and no for this one. Firstly there is no easy screen in MPS so these is no existing preset. Although it can be done manually by using a crossover in the FX section. I will do a template for you to get you started.
It is great to use with the fx, I like to distort the noise more than the tonal for example. Have a play around.
But in my opinion the application for ground breaking sound design comes when you use the technique with samples rather than synth.
You see if you are sound designing with a synth you can pretty much make ANY sound from scratch and have full control over the noise and tonal elements.
But samples are already made and it is normally tricky to split that sample up into separate parts to process them individually. Commonly splitting is done just by frequency (normal multiband) but with these new crossovers spectrum and transient/tonal, we can get really creative!
Imagine using the noise of a car driving past and blending that with the sound of a man screaming! that would make for really interesting sounds for film or music sfx. Then throw Morph into the mix and we have some VERY unique sounds! I love playing around getting weird textures from samples like the noise of a match being struck and the transient of a penny dropping and the tone of a bottle being blown and then I subtly mix these in with my synth sounds to make things sound more interesting and organic.
Jason @ Melda Production
-
- KVRAF
- Topic Starter
- 1759 posts since 11 Nov, 2009 from Northern CA
Thanks for the detailed response.jmg8 wrote: Having the bands soloed sounds very unnatural and apart from creative stuff doesn't really have much use.
Due to the unnatural sounding separation i find it best to use it quite subtly.
My reason for solo-ing the sounds the bands was purely to observe the effect of muting all but the band of interest.
Here's what I did for testing. I created an MPowerSynth oscillator that had a fundamental, another harmonic about two and a half octaves up, and another cluster another two and a half (or so) octaves above that. I left one of these at full volume, another down about 40 dB and another down about 60 dB. I then viewed the output in a spectrum display.
Solo-ing the respective bands resulted in behavior that basically made no logical sense. Solo-ing bands did result in a change to the spectrum display, but it was far from the expected result of just showing the band being solo-ed.
Thus I was asking for a suggestion of a scenario I could use to actually see what was going on with respect to the isolation this type of band separation was actually causing.
-
- KVRAF
- Topic Starter
- 1759 posts since 11 Nov, 2009 from Northern CA
Sorry, I'm not following you. I thought to have this band separator do its thing, we needed a) different frequencies in the signal, and b) different levels in those separate frequencies so that the band separator could assign them to a particular band.jmg8 wrote:OK, so simply create a sine wave with noise oscillator in MPS. You should be able to use the spectrum crossover to separate them.
I'm not following how "simply create a sine wave" will satisfy what's needed for band separation here. Sorry for being dense.
By the way, in case you're wondering why I'm jumping through all these hoops, it's because I'm hoping to write an article for SoundBytes Magazine entitled something like "MeldaProduction Redefines Multi-band". To do that, I need to rather thoroughly understand all the band types, which I think I do now except for Spectrum (and that other mystery about MultiAnalyzer behavior in MXXX for Panorama, which is the subject of a different thread).
-
- KVRist
- 460 posts since 25 Jan, 2016
I'm not following either...but also trying to understand.
1. In MPS I created a sine wave AND activated the noise generator so that a sine and some noise played simultaneously when a key was pressed.
2. I then added an MB ending effect to that mixer channel.
Note: I tried EVERY effect that ended in MB. MPhaserMB was the ONLY effect that showed any signal in the display. Upon deeper inspection, the display settings were set to Analog in MPhaserMB in the MultiMode preset.
3. When selecting MultiMode, I thought (rightly or not) that soloing A or B would yield either the tone or the noise. A yielded silence and B yielded the combined signal (in all MB effects except for MPhaserMB). In MPhaserMB I could recognize the separation was happening, but it actually sounded as if it was just a High/ Low filter has been applied. (Possibly due to having the Analog applied in the settings?)
In the end, I'm confused. I wasn't able to SEE anything happen and when I could HEAR it, I wasn't sure if it was as intended. I knew that I was over my head in trying, but curiosity got the better of me.
I numbered my steps so that everyone could point out where I so obviously went wrong and misunderstood and endangered the universe's balance.
1. In MPS I created a sine wave AND activated the noise generator so that a sine and some noise played simultaneously when a key was pressed.
2. I then added an MB ending effect to that mixer channel.
Note: I tried EVERY effect that ended in MB. MPhaserMB was the ONLY effect that showed any signal in the display. Upon deeper inspection, the display settings were set to Analog in MPhaserMB in the MultiMode preset.
3. When selecting MultiMode, I thought (rightly or not) that soloing A or B would yield either the tone or the noise. A yielded silence and B yielded the combined signal (in all MB effects except for MPhaserMB). In MPhaserMB I could recognize the separation was happening, but it actually sounded as if it was just a High/ Low filter has been applied. (Possibly due to having the Analog applied in the settings?)
In the end, I'm confused. I wasn't able to SEE anything happen and when I could HEAR it, I wasn't sure if it was as intended. I knew that I was over my head in trying, but curiosity got the better of me.
I numbered my steps so that everyone could point out where I so obviously went wrong and misunderstood and endangered the universe's balance.
-
- KVRAF
- 2702 posts since 9 Jul, 2015 from UK
dmbaer:
If you were to set 2 sine wave at different frequencies then yes technically you could probably seperate them using the spectrum crossover. However this seems more of an experiment and is not how the crossover will actually be used normally, as most signals will contain hundreds of frequencies. I think you are thinking of this the wrong way, I will try to explain.
The simplist way to put this, is that the spectrum crossover could also be called NOISE/TONAL.
This is why I suggested an example for you to try and seperate a single sine wave and the noise osc. Even though they are mixed together the spectrum crossover will do a pretty good job at seperating them.
If you were to set 2 sine wave at different frequencies then yes technically you could probably seperate them using the spectrum crossover. However this seems more of an experiment and is not how the crossover will actually be used normally, as most signals will contain hundreds of frequencies. I think you are thinking of this the wrong way, I will try to explain.
The simplist way to put this, is that the spectrum crossover could also be called NOISE/TONAL.
This is why I suggested an example for you to try and seperate a single sine wave and the noise osc. Even though they are mixed together the spectrum crossover will do a pretty good job at seperating them.
Jason @ Melda Production
-
- KVRAF
- 2702 posts since 9 Jul, 2015 from UK
dmbaer continued (sorry for some reason my computer is limiting how many characters I can put in one reply)
Basically noise is random, it has the frequencies jumping up and down in level randomly. So if you were to look at just 1 frequency on its own, it would probably be just mostly silent with the occasional jump of level. If you measured the loudness of this frequency it would be low, so would all the other frequencies that contain noise. So these frequencies will be set towards the left of the crossover.
Opposite to this will be our sine wave, which will have a constant level in one single frequency so it will register as having a much louder level and will appear to the right of the crossover.
Have a go at separating them. Then try other sounds, you will find you can get good results separating tonal/noise.
Basically noise is random, it has the frequencies jumping up and down in level randomly. So if you were to look at just 1 frequency on its own, it would probably be just mostly silent with the occasional jump of level. If you measured the loudness of this frequency it would be low, so would all the other frequencies that contain noise. So these frequencies will be set towards the left of the crossover.
Opposite to this will be our sine wave, which will have a constant level in one single frequency so it will register as having a much louder level and will appear to the right of the crossover.
Have a go at separating them. Then try other sounds, you will find you can get good results separating tonal/noise.
Jason @ Melda Production
-
- KVRAF
- 2702 posts since 9 Jul, 2015 from UK
werzel:
OK so first, you were doing the experiment correctly, however the multi-mode preset cannot know automatically what kind of signal it will receive. Therefore you need to fine tune the separation. In this instance just try to lower the smoothing to 0% that should work. Also the separation slider will act like the balance of how much will fall into A or B, so solo A and use the separation until you have a nice balance in each side.
Thank you so much for spotting the mistake in phaser, I will get that fixed. It is not supposed to be set to analog.
OK so first, you were doing the experiment correctly, however the multi-mode preset cannot know automatically what kind of signal it will receive. Therefore you need to fine tune the separation. In this instance just try to lower the smoothing to 0% that should work. Also the separation slider will act like the balance of how much will fall into A or B, so solo A and use the separation until you have a nice balance in each side.
Thank you so much for spotting the mistake in phaser, I will get that fixed. It is not supposed to be set to analog.
Jason @ Melda Production
-
- KVRAF
- Topic Starter
- 1759 posts since 11 Nov, 2009 from Northern CA
jmg8 wrote:The simplist way to put this, is that the spectrum crossover could also be called NOISE/TONAL.
This is why I suggested an example for you to try and seperate a single sine wave and the noise osc. Even though they are mixed together the spectrum crossover will do a pretty good job at seperating them.
I read your original response too fast and missed the "with noise" part. Now you are making perfect sense. Also, I like the noise/tonal view of this. It now starts to become something that anyone could get their head around. Many thanks.