Hi

How do I know what my input gain should be set to? I have Sound Devices Mixpre-3. I know where to set it I just don't know what I need to set it to.
Also, I have fethead preamp attached to the mic. It gives me extra 18db, but I hear shhhhhh sound coming from it in my headphones before I even connect the mic. Is this normal?

The convention in digital recording is to set the gain so the signal averages around -18 dBFS (or some other value near that number, depending on the converters). That means it's spending about half the time above and half the time below that value when the source is in its normal/loud volume range.

Typically, a 0 VU level will translate to +4 dBu, which lines up to -18 dBFS for many converters, but sometimes a slightly different value. The exact number isn't critical as long as the analog side of things is in its optimum range (above noise, below distortion) and the digital level is well below 0 dBFS.

Percussive sources are a special case since they have so little sustained energy. For those, you have to set it by peak level, being sure to stay well below 0 dBFS. Peaks at -6 dBFS are okay if you're reasonably sure it's not going to go higher, but you can keep the peaks at -12 dBFS if you want a bit of safety margin.

Note that a signal recorded with an average level of -18 dBFS will sound low compared to finished audio. That's normal. The final levels for delivery should be achieved in post.

I'm not sure about the noise you're getting from the Fethead/preamp. I'd say don't use the Fethead unless you really need to to get a mic's signal up to the required level.

A very simple answer set gain so you're peaking -18db to -6db with fader at 12 position. This allows plenty of headroom so you won't clip. In post you can adjust it to your liking. The amount of headroom varies based on what you are recording. All of this assumes you have proper mic to subject distance. If you have a dynamic mic 6ft away then increasing the gain to achieve proper levels, you'll end up with a lot of noise.

If the MixPre has enough gain, (which it should in most scenarios), do not use the fethead preamp. If you are recording at conservative levels, use 24 bit depth .. or .. 32 bits if you have the MixPre II, and you will not have to worry about digital clipping either.

Thank you all. I will remove the fethead

A very simple answer set gain so you're peaking -18db to -6db with fader at 12 position. This allows plenty of headroom so you won't clip. In post you can adjust it to your liking. The amount of headroom varies based on what you are recording. All of this assumes you have proper mic to subject distance. If you have a dynamic mic 6ft away then increasing the gain to achieve proper levels, you'll end up with a lot of noise.
Can you explain more how to do this? What do you mean by fader at 12 position? Here is a photo of my recorder. What should the fader be set to on my recorder?

12 (o'clock) on a channel fader is usually "unity gain", unless specified or marked otherwise..

A very simple answer set gain so you're peaking -18db to -6db with fader at 12 position. This allows plenty of headroom so you won't clip. In post you can adjust it to your liking. The amount of headroom varies based on what you are recording. All of this assumes you have proper mic to subject distance. If you have a dynamic mic 6ft away then increasing the gain to achieve proper levels, you'll end up with a lot of noise.

12 (o'clock) on a fader is usually "unity gain", unless marked otherwise.

So set it to roughly where the number 1 is written and then adjust the input gain?

So set it to roughly where the number 1 is written and then adjust the input gain?
Yes. When the fader is in the middle (12 o'clock) you have room to adjust the levels up or down while recording.

Yes. When the fader is in the middle (12 o'clock) you have room to adjust the levels up or down while recording.
Do I need to be speaking to the mic while I'm setting the input gain?
I watched this video and I can't tell if he is talking to the mic or not:
https://youtu.be/aJRoZ_joUdU

It is better to have the talent speaking in a way that simulates the coming live shoot to get the most accurate levels.

If no one is speaking there is no way to set the levels meaningfully.

It is better to have the talent speaking in a way that simulates the coming live shoot to get the most accurate levels.

If no one is speaking there is no way to set the levels meaningfully.

I'm not talking about just setting levels. Obviously someone has to speak. I'm asking when setting the channel input gain.

I'm not talking about just setting levels. Obviously someone has to speak. I'm asking when setting the channel input gain.
Like I said before the fader should be left at the 12 o'clock position, while you set your levels by adjusting the gain. Gain adjustments are done before recording when you do your audio check.

Hi Kathy. There’s a lot of good help here and my 2 cents is, you should go back to the start and follow the UTube video introductions for your MixPre 3.

I didn’t post any here because I think the secret is to keep looking for the videos that answer all your questions, if one is confusing don’t worry just move on. It can be frustrating I know, but take it slow we all started like this. Keep your mixer with you and try what’s suggested, make notes, you’ll get there.

Cheers.

Yeah if you're unfamiliar with audio and are a visual person it might be easier to watch mixpre video tutorials.

And if you’re not a visual or audio person sell me your MixPre 3. Those fancy coloured leds around the gain controls are the bees knees.

What a fantastic idea, one trick the others can’t copy without industry ridicule and a nice bonus for the Sound Devices wallah who came up with that idea.

Cheers.

Like I said before the fader should be left at the 12 o'clock position, while you set your levels by adjusting the gain. Gain adjustments are done before recording when you do your audio check.
Yes, I got that. The fader is at 12 o'clock position. Then I adjust the gain until it peaks at about -18db to -6db. Got that. The question is whether while I'm adjusting the GAIN am I speaking to the mic? on all the videos I watched they are just talking to the mic they are using for recording the video and not the mic they are adjusting the input gain for, so my question again is am I speaking to the mic while adjusting the gain not recording levels (that's after the gain is set)

This is to everyone not just Pete.
Thanks for the helpful responses so far. I don't need to know that I need to watch more videos or to sell my recorder.

And if you’re not a visual or audio person sell me your MixPre 3.
I am both.

Yes, I got that. The fader is at 12 o'clock position. Then I adjust the gain until it peaks at about -18db to -6db. Got that. The question is whether while I'm adjusting the GAIN am I speaking to the mic? on all the videos I watched they are just talking to the mic they are using for recording the video and not the mic they are adjusting the input gain for, so my question again is am I speaking to the mic while adjusting the gain not recording levels (that's after the gain is set)

This is to everyone not just Pete.
Thanks for the helpful responses so far. I don't need to know that I need to watch more videos or to sell my recorder.
The terminology can get a little confusing because the same thing can be referred to by different names. I call the amplification as "gain" but it can also be referred to as trim and I call fader gain as "fader"

You won't be able to set your gain unless someone is talking into the mic. Before recording an interview you'll ask the subject to talk how the normally do. This is when you set the gain.

The question is whether while I'm adjusting the GAIN am I speaking to the mic? on all the videos I watched they are just talking to the mic they are using for recording the video and not the mic they are adjusting the input gain for, so my question again is am I speaking to the mic while adjusting the gain not recording levels (that's after the gain is set)

You set levels using the actual source you're going to record, or as close to that as you can get.

Things labeled "gain" usually control preamplifier input gain. Preamplifiers are there to accommodate a variety of input types and levels and bring them into a "standard" range for use by the device. That's why you need to set it using the actual source you're going to record.

One pitfall is that people often don't speak (or sing, or play) at performance level during soundcheck. I try to get them to give me their loudest sound. Having the gain a little low is much better than having it a little too high, especially with modern quiet digital devices. If I don't think they're giving me a true representation of their loudest sound, I run the gain a few dB low to leave myself a safety margin.

Technically, any change in level is gain (including negative gain, a reduction in level), so adjusting the fader affects the overall gain. But the label is mostly used on preamps.

Years ago on mixing boards the preamp gain was often marked "gain trim" or just "trim." That made a little clearer distinction between the gain of the preamp, and the attenuation of the channel fader.

Also, some channel faders were actually marked in dB, with maximum level being marked as 0dB (which means unity gain, no loss). Minimum level was marked as "infinity symbol" (meaning infinite attenuation, infinite loss, so no signal gets through). Intermediate settings were marked as negative numbers (for example -6 would mean -6dB, meaning that the attenuator passed 50% of the original signal voltage). IMHO, lack of intelligent markings does not make it easier for anyone to understand how the system works. (OTOH, I tend to look at things from an engineering perspective, which is not necessarily the viewpoint of the average appliance user.)

All of this would probably be easier to comprehend the first time, if you could see a block diagram with some numbers on it (or if you had a good mentor to explain this in person). But at any rate, Kathy, I think all the answers given here have been correct, even if your lack of experience made you somewhat uncertain about what was meant.

It would be nice if they'd at least mark a 0 gain position on the fader knob. The halfway point is just one possibility. The 70% position is also a pretty common default setting.

This contains a lot of history which may be rather arcane. If you're uninterested or impatient, just skip to the last paragraph (or to the next post).

Back when an amp (or mixer or whatever) had an analog input and analog output, the dB levels could actually mean something. I recall old radio control boards which used dBm as a reference level. Zero dBm was 1 milliwatt (0.001 watt) of power, and at the standard impedance of 600 ohms, that would be 0.775 volts RMS. Confusingly zero VU was NOT the same as zero dBm, because of internal wiring of the VU meter, related to some complex stuff affecting distortion.

Anyway, dBm and VU both express actual power levels, while dB expresses gain (or attenuation, if the number is negative). Gain is the ratio between two voltages, or between two power levels. Both numbers have to be in the same units. e.g. (measured output voltage) / (measured input voltage) = numerical gain. Then you convert the numerical gain to dB. If the two voltages are equal, gain is 1.0, and that translates to 0 dB. If output voltage is 2X the input voltage, then system gain is +6 dB. Etc.

If the input power of one of these antique boards was 0 dBm, and the channel pot was at maximum rotation (corresponding to 0 dB attenuation), and the output level control was at the position marked 0 dB, then the output power would be 0 dBm. If you turned up the output level control, you could get more level to feed to the transmitter. Or if you turned down the channel fader, then you'd have less than 0 dBm output. The dB markings were actually meaningful in terms of measured power or voltage.

If the board operator knew that -6 was the correct setting for a turntable, he would use that as a starting setting when using that turntable. After doing a level check with talent in "studio B" if he found that -12 was a good starting setting, he would use that. Of course one would ride the gain up and down from that starting position as needed.

Remember that gain is the ratio between two voltages, or between two powers. Both numbers have to be in the same units. Now that we're converting from input voltage to some binary number in a file, the concept of "zero dB" or "unity gain" becomes somewhat moot. There is no simple numerical system gain. Gain would be related to a ratio of (volts) / (dB FS) and that would become pretty well unmanageable.

I agree with Tracy: it would be useful to have some numerical scale on the faders. If common practice is to start with the fader at mid-rotation, then set gain trim accordingly, it makes sense to put "0 dB" at the midpoint of rotation. But does everyone start with the fader at mid-point? If every fader was marked like that, what about the people who are in the habit of starting with fader at 70% rotation?

Also, most audio potentiometers aren't linear. The gain difference between mid-point and maximum might be 10 dB. But the change between mid-point and minimum has to be an infinite number of dB, because "full off" is an infinite amount of attenuation or "negative infinity" gain. So if you mark the scale in dB, the number of dB per degree of rotation will vary from one end of the scale to the other. So some manufacturers have arbitrarily adopted some non-dB scale, like 1 to 10 (or, in the case of Spinal Tap, 1 to 11).

Personally, I think the best and most universal solution is to mark dB of attenuation, with the full maximum rotation (full volume, no attenuation) marked as zero, and everything else as negative numbers indicating actual dB. That's less arbitrary than putting "zero" at 50% rotation, or 70% rotation, or some other arbitrary place. I like using a dB scale like this because I understand dB and am used to working with that unit. But for a newbie, it might be confusing, since the numbers would in a way be "backward" with zero at the maximum end, rather than minimum. Should we use mathematically meaningful numbers that are useful to pros? Or arbitrary numbers that are "newbie friendly"? I personally do not like the idea of no numbers at all.

At any rate, there can be different names for "mid-rotation" depending on what equipment you're used to and whom you're talking to. If you say "twelve" or "twelve o'clock" most experienced audio people will know that means mid-scale. A newbie (on their way to becoming an experienced audio person) might need a bit more explanation. All of this can be learned from books (if you find the right ones) or from videos (if you find the right ones ... and there are a lot of really wrong ones!) or from a mentor (if you find the right one). All are valid sources, and all are valid suggestions to make (and to receive). As Socrates said, "A wise man knows what he does not know." And my own corollary to that: be gracious and don't lash out at people who offer helpful suggestions.