Frustration with Gates

[aciddose]

perhaps you've used analog expanders and havent ever known the difference?

[IIRs]

I've used both, and know the difference well enough to earn my living by it.

[aciddose]

well i've got an analog gate right here and the signal will be cut to -90db with only a single stage. using two stages i can get -112db, with the side effect of the envelope being n^2. now, this is with a single npn/jfet used for each scaling stage.

with a differential amplifier set up with transconductance, applying ground to the control voltage input instantly gives -110db, and two stages give -180db, or perfect isolation since the pcb will allow greater leakage. actually, the pcb should at high frequencies (20khz) allow -100db leakage anyway so anything greater than that is a bit silly without a complex layout designed to eliminate leakages by using ground planes.

these are two very simple designs, using less than 10 components each including passive components. scientific systems are far more accurate and far more complex. do you really think any of the communications systems in use today would work with only a -80db floor?

you've used very poorly designed gates or expanders i suppose, since the gates i build myself perform far better.

making the claim "analog gates only allow -80db cut anyway" really just proves the person making the statement has absolutely no idea.

[Jason Brian Merrill]

apparently all other gates besides GAG are mediocre..

strange how software uses hardware limitations....

saying this is like a strange crutch -- "find me a hardware gate that does more than -80 -- " uhhh, dude -- this is SOFTWARE. those limitations can be gone, if you want them to be.

for me? I cant see using more than -80db, but ok, i see where people are coming from to an extent.

[aciddose]

"GAG is not a gate"

i'm not arguing about is -80 is acceptable, for a gate, no it isnt, but i honestly dont care since i've written my own gate.

i'm saying GAG is not a gate, it is an expander, something radically different.

before making rediculous claims about 'analog only gets -80db anyway' anymore, read the lm13700 datasheet. this is the most simple chip you can use for a vca, and the typical crosstalk BETWEEN AMPLIFIERS while turned on to full scale is -110db. the leakage THROUGH a single amplifier while turned fully off is far less.

[xander]

http://media.zzounds.com/media/3630_Man ... pressor%22

I don't see where it says closed state is -10 dbv. That wouldn't be very closed at all. Oh I see...page 8: "Threshold (no gating to -10dBV)" No, that's referring to the switch on the back that picks between pro and consumer audio levels. Not sure what they mean, maybe the gate feature doesn't even work when its in -10 mode???

I don't have THAT button on the rear -- but I do have a -10dbu to +4dbv I/O Level slide switch. Is that it? Maybe the difference is that mine is a UK model?

Ayway, here's what my manual says:

1.6 FRONT PANEL - NOISE GATE CONTROLS
The noise gate will not pass signals below the specified threshold setting, but signals above the threshold will pass unaffected. In a typical application, the threshold would be set just above any residual hiss or noise. Thus with no signal going through the 3630, no hiss would be audible. A signal going through the 3630 would pass through unimpeded.
The hiss would also be present under these conditions, but since a strong signal is present it would tend to "mask" the hiss.

To use the 3630 solely as a noise gate, set the Ratio control to 1:1. This passes the input signal to the noise gate without processing.

Threshold (no gating to -10dBV)
Sets the level above which the noise gate will allow signals to pass.
Turning the Threshold fully counter-clockwise disables the noise gate and lets the signal through unaltered.

To set the proper gate level for most applications, turn the Gate Threshold to OFF. Turn on all instruments to be gated but do not pass program material through them (For example, turn your microphone on but do not sing through it). Increase the Threshold (clockwise) until the red
CLOSE LED comes on. Any background noise should now be eliminated.

Sounds good to me!

Actually, cheap or not, I really like this little single-space unit & it really does a man's job for the price -- and it's 14 years ago since I bought it and still use it nearly every day!

In fact, I've used it on all the initial vocal tracks (pre-master) you can hear on my website!

*BTW -- I especially like the side chain I/Os on the back and EQ it through me little Behringer mixer -- as I think you also mentioned.

Cheers,
Alex

[Jason Brian Merrill]

-10dbu to +4dbv I/O

this just means unbalanced or balanced I/O doesnt it?

[aciddose]

it means 500mv or 5v output. it really is a useless switch on most gear. it is there for the ninnies who have to use the volume control in full range at all times. most gear works at 500mvpp, i tend to modify that gear to work at the full range. at 500mv, you run in to all kinds of difficulty with noise.

[Torben]

[xander]

-10dbu to +4dbv I/O

this just means unbalanced or balanced I/O doesnt it?

Not impedance specifically -- just signal levels.

To quote the manual:

+4 dBu or -10dBV switch Use this switch to match each channel of the 3630 to the rest of your gear. Generally equipment designed for larger pro studios tend to use +4 dBu nominal signal levels. Equipment designed for smaller project and home studios and electronic musical instruments
tend to use -10dBV nominal signal levels. If the input signal doesn't provide enough drive to the LED meters, use the -10dBV position. If distortion occurs, use the +4 dBu position.

Changing the switch position modifies the way the input/output meter references signals. When set to the +4 dBu position, a 4 dBu input signal reads 0 dB on the input/output meter. When set to the -10dBV position, a
-10dBV input signal reads 0 dB on the input/output meter. When the unit is set up for unity gain (no compression, output at 0) the input level will be equal to the output level regardless of the position of the +4 dBu/-10 dBV
switch.

Cheers,
Alex

[IIRs]

well i've got an analog gate right here and the signal will be cut to -90db with only a single stage. using two stages i can get -112db, with the side effect of the envelope being n^2. now, this is with a single npn/jfet used for each scaling stage.

So use 2 GAG1s then! 160 dB of attenuation enough for you?

(You could even key the second GAG1 with the same signal as the first to ensure they open & close at exactly the same time )

i'm saying GAG is not a gate, it is an expander, something radically different.

GAG1 can be a gate OR an expander. You are correct that they are not the same thing, but you are totally incorrect in your description of how a 'true' gate should operate: a gate applies a fixed gain change when the signal drops below the threshold, while an expander applies a variable gain change. I have never known any analogue gate to offer more than 80 or 90 dB of gain change, and even that much is rarely desirable in my experience.

[OMNIFEX]

That's very wrong I'm afraid. Gates 'close' all frequencies: its just a gain change basically, like a compressor. The sidechain filters tune it to respond only to certain frequencies, but the gate still affects all frequencies. Clicks arise from very fast attack times regardless of the frequency content of the program material.

I'm well aware gates are designed with all
frequencies in mind. Possibly, I worded my explaination wrong.

The Fast Attack is based on Higher Frequencies. Having the gate close @ 25Hertz, will not have a fast Attack compared to 100 Hz. Any 1/24 Spectrum Analyzer will show the lag between the two.

A Dynamic Expander would be a better solution for Attack, than a Gate.

Gates were designed to reduce/eliminate noise which happens on either end of the Spectrum.

Many don't understand how to use a Gate properly, and end up with clicking sounds.

[IIRs]

That's very wrong I'm afraid. Gates 'close' all frequencies: its just a gain change basically, like a compressor. The sidechain filters tune it to respond only to certain frequencies, but the gate still affects all frequencies. Clicks arise from very fast attack times regardless of the frequency content of the program material.

I'm well aware gates are designed with all
frequencies. Possibly, I worded my explaination wrong.

The Fast Attack is based on Higher Frequencies. Having the gate close @ 25Hertz, will not have a fast Attack compared to 100 Hz. Any 1/24 Spectrum Analyzer will show the lag between the two.

A Dynamic Expander would be a better solution for Attack, than a Gate.

Gates were designed to reduce/eliminate noise which happens on either end of the Spectrum.



In the case of Side chaining. Pending on how you have your system configured, you

[OMNIFEX]

Read the edited post.

I'm more Pro Audio, and when I stopped myself, it had nothing to do with Recording Audio.

[xander]

Well, ta hell with it. I like GAG and what it does for my sound and I use it quite frequently now in post production. In fact, it's one of the best I have ever used, so Kjaerhus Audio was apparently paying attention to 'pleasing sound & performance enhancement' rather than text-book specs, which is way fine by moi. Listen to my songs and then decide...

Because, regardless of how perfectly the pie looks and how well the chef followed the time-tested recipe, it's the taste that is the only thing with any meaning!

Cheers & Let's All Party & Be Friends!
Alex