Simple questions on synth tech basics...

[camsr]

The LP filter you mention is interesting as I noticed that the lower the cutoff frequency of the synth bass patch, the more solid the bass sounds. The best results I obtained were for a saw/pulse with cutoff so low it almost sounds like a sine. Maybe the volume ratio between low and high frequency content is screwed on headphones because the power is simply missing with those tiny membranes...
On speakers the higher frequencies don't hurt the low frequencies, they just add character.

No, it's because the bass frequencies are present at both ears. The human head will block sounds from the side that might have gone to the other side's ear. It's called Diffraction

[fluffy_little_something]

Strange, I have heard repeatedly that the human hearing/mind can't even located low frequencies, which is also the reason why people only have one sub-woofer, I take it?

[camsr]

It's not about localization, it's about intensity. The method I described earlier adds +6dB of bass to each headphone channel.

A funny thing about that is listening to out-of-phase stereo bass in headphones. One might assume it's inaudible, but there's just more to it.

[mystran]

The reason one can't really localize bass very well probably has to do with the fact that low frequencies have long wave lengths so a little time delay doesn't really add meaningful phase-shift (and if you add a bit of room resonance to mess with the phases, then it's even harder to make anything out of it) and small objects (like human head, not to even mention the ear which normally helps localization by providing direction dependent filtering) have little to no effect on the wave-fronts moving in air so you always kinda hear it at the same level in both ears .. so your brain basically has very little to use for localization. Also, given the higher frequencies of harmonic series typically present, the brain will use the more accurate information from the higher frequencies (and just assume the bass comes from the same direction), so the subwoofer approach works fine.

If you directly feed (only) low frequencies to each ear separately (via headphones), you can actually localize them (to one of the ears) just fine. Try it, it'll work. It actually feels rather weird, probably because the brain (well my brain anyway) isn't really used to that kind of stimulus, but it'll work.

[JCJR]

The headphone settings almost certainly apply a similar filter/blend effect to the output.

Unfortunately there are so many variations I'd never be able to tell you exactly what they do, and it is difficult to measure even given the devices themselves. Sometimes you're lucky enough that they'll publish what they do in the manual and sometimes not.

Yes, while you're listening to anything other than technical material the filter will almost certainly have a beneficial effect. As he describes on the webpage I linked, the localization of the sounds positioned too far left/right will be improved a ton even though the stereo image is actually reduced in width. This is because of the fact our brain doesn't measure the difference in amplitude, but measures instead the frequency dependent delay time.

So, having 100% width is detrimental to stereo image on headphones and I'd also say in a mix in general. The sound will not be localized "left" or "right", it will be localized directly to the speaker it comes from.

Using some VST plugins which implement this or similar filters is a great idea if you're having any trouble with wide stereo mixes.

Thanks Aciddose. Interesting.

If the passive matrix illustrated in your linked page works "good enough", or if some other/better passive matrix would work "good enough", am wondering whether the matrix would most properly be built-in to headphones, with a switch to toggle the matrix in or out.

Got me wondering about portable players such as iPod. If the matrixing is such an audible benefit, is it built-in to the hardware or driver software in such as iPod? A brief google didn't yield any relevant hits. My 32 GB ipod is a few years old, but doesn't show any relevant setting in the audio preferences page, so if it has headphone matrixing, then the feature is presumably always-on. Which would be less than optimal when I connect it to studio monitors or the old Sentry 100a's in the bedroom. But if matrixing is absent, less than optimal when used with headphones.

Perhaps the presence or characteristics of matrixing in a device or driver could be probed with a test audio file. Perhaps the diagnostic audio file could contain two clean unfiltered bass ramp waves, one wave hard-panned left and the other hard-panned right. Perhaps a proper tuning offset would be a flatted-fifth, so that the harmonic frequencies between the waves would be "as distant as possible".

By playing the test file thru a device and recording the output-- If each wave stays hard-panned, it would indicate no matrixing. If there is mixing between the channels in the bass, one could probably deduce the nature of the filtering based on the observed rolloff of the higher harmonics of the ramp waves.

Or maybe some other test method would be better.

Edit: Well, maybe the simplest test would be a single sine sweep hard panned to one side. The nature of any matrixing possibly simply analyzed based on whatever audio appears on the silent channel.

[aciddose]

Yes it is trivial to identify if a cross blend is applied or not.

Simply play a stereo clip with low frequency and high frequency tones in only one side.

100hz, 1000hz, 2000hz, 4000hz, 8000hz.

Use your own ears to measure the relative level in the left/right phone by placing the opposite on your cheek, whatever is most comfortable.

You can then identify whether a delay or cross blending filter is used, as well as which sort of filter is applied.

The standard filter should have no audible delay, but there should be significant content from one channel passed to the other for 100hz, 1000hz. 2000hz should be less, and 4000hz a lot less if even audible at all.

For the more "advanced" filter he mentions, you should also hear 2000hz blended nearly as much as 1000hz.

For a simple delay there should be less frequency dependency, so 4000hz and 8000hz should also be blended. You might also be able to hear the delay. If not, this could also simply be an unfiltered cross blend between the channels.

If 100hz is not blended, you can be fairly certain there is no effect.



re: inline application of this effect

A lot of filters can be applied inline, yes. Unfortunately this changes the impedance and can produce a much more heavy filtering effect than you want. I'm not certain, but I'm leaning immediately toward it not making much sense to apply this effect without it being behind a buffer, just like any usual EQ or similar.

You can always put a 100u capacitor across your speaker terminals, but the amp will have a hell of a time driving into that if it doesn't light on fire.

Remember the speakers are typically 8ohm, while the filter circuit uses 1k inline and 2k between. So your amp will suddenly see 1.008k impedance rather than 0.008k for low frequencies. The 2k doesn't have much effect, it changes the 8 ohm into 7.968 ohm total.

So this will severely high-pass filter the signal.

If we just ignore the 1k and 2k resistors and calculate the CR high-pass frequency, this is 1 / (2 * pi * 470n * 8) hz.

1 / (2 * pi * 0.000000470 * 8) = 42328hz.

Probably won't hear much :)

[fluffy_little_something]

Regarding the creation of sound, would it not make sense to approach it the other way round? It seems to me that usually what happens is that oscillators produce amplitude peaks across the whole frequency spectrum, right? They do that one way or another, for instance in subtractive or additive ways.
By the other way round I mean, why not start with a dense white noise floor (whose level one could set with the volume control, so no risk of dangerous peaks) and then using the filter as an "inverse oscillator" that carves out the valleys from the noise floor? This way that inverse oscillator and the filter would be one and the same, which might make things easier when programming sounds. It would be a bit similar to how different light colors can be extracted from white light ^^

[mystran]

Regarding the creation of sound, would it not make sense to approach it the other way round? It seems to me that usually what happens is that oscillators produce amplitude peaks across the whole frequency spectrum, right? They do that one way or another, for instance in subtractive or additive ways.
By the other way round I mean, why not start with a dense white noise floor (whose level one could set with the volume control, so no risk of dangerous peaks) and then using the filter as an "inverse oscillator" that carves out the valleys from the noise floor? This way that inverse oscillator and the filter would be one and the same, which might make things easier when programming sounds. It would be a bit similar to how different light colors can be extracted from white light ^^

Remember waveform produced by the oscillators are pitched with a harmonic series (sometimes not exactly harmonic, but still close in most cases). To get this from white-noise, you can either use one resonator per harmonic (which I suppose would be a form of "modal synthesis") or you can get the whole series directly with a comb filter (which is essentially what traditional physical models do).

With either approach, if you excite it with an impulse (or some impulsive sound), you get some sort of a "plucked" or "struck" sound, and if you excite it with continuous noise you get some sort of a "blown" type sound.

[aciddose]

It is a very similar idea to additive synthesis or physical modelling. These are almost definitely far better options for synthesizing most of the sounds possible, with the exception of subtractive sounds.

This is a bit obvious though. You wouldn't use another type of synthesizer to try to produce "FM" timbres, it would be far more easy to just use an "FM" synthesizer.

The real reason we don't use additive synthesis is the complexity. A physical model is already bad, a full additive system is far, far more expensive. A subtractive synthesizer by contrast is extremely simple. You only get more simple by using full samples at which point you're not even really a synthesizer anymore.

[fluffy_little_something]

Almost all synths have a key tracking knob for the filter. In Retrologue, however, one can also set the turning point for that knob. And I was wondering to which note that parameter is set with all those synths that are hardwired in that respect. Is there a "natural" note for that or is it a matter of taste?

[aciddose]

It's an issue of the range of the cutoff knob. If you can transpose the base pitch of the filter far enough it makes the zero point of the key tracking inconsequential.

Adjustment of that value is simply a 2nd cutoff knob.

[fluffy_little_something]

- Wouldn't it make sense to program the filter keytracking control so that the volume is automatically adjusted? As the filter opens on higher octaves, the overall volume gets louder and I always have to waste a modulation slot just in order to undo that volume increase.

- I was playing around with Hive, which has two filter sections one can run in series. So when I combine a 12db LP and a 24db LP, I get a 36db LP, right? But what is the result when combining a 12db LP with a 24db HP? A BR filter? And how many poles?

[aciddose]

You can't always combine filters one after the other. Two 12db filters in series does not produce a 24db filter with the same properties, as the spectra are multiplied together. If you had a 5db boost from the resonance before, it is now 10db. Likewise for phase.

Adjusting volume would not make sense, a filter working correctly does not change the amplitude of the signal at DC. If you'd like to add modulation to do this (to adjust changes in RMS at a particular frequency) you need a synthesizer with a capable enough mod-matrix, although I'd expect it to be extremely difficult, very complex and you could never compensate correctly for dynamic signals. Single frequencies yes, mixed frequencies very difficult, dynamic frequencies impossible.

[JCJR]

I haven't used many soft synths, or more accurately haven't patch programmed many softsynths.

With some of the hardware synths, one modulator was keyboard scaling, which could have various slopes, and could be applied either negative or positive modulation.

So if a patch got too loud in a high pitch range, it was pretty easily adjustable applying some KS negative mod to tha amp, and if a patch got too dark or bright over the keyboard, could usually be adequately fixed with some negative or positive KS on the filter.

I'd guess that would also be a common feature in soft synths? Haven't a clue really.

[fluffy_little_something]

I don't think that a dedicated amp key follow control is common with soft synths, I only remember one among everything I have used so far.