why is recorderd volume so low?

[misko]

Complete newbie here ... recording some sound for work, using Zoom h4n together with Rode NTG1. Zoom settings fairly out of the box, but recording levels are turned up to 100. Ni-mh battery, 48 V phantom power. And the recorded volume is so low!

I must position the mic about 15 cm from my face to get the levels up to -6 and a decent volume. In Audacity the file looks like this:


The mic is positioned above my forehead, pointing downwards to my chest, have followed this video).

Any suggestions - I am thankful for all the help I can get.

 

[Bobbsy]

Well, a few thoughts...

First, I disagree slightly with the guy in the video. Yes, put the mic on a boom but you want it to be slightly in front of you and aimed roughly at your mouth, not your chest. Even with a short shotgun like the NTG1, the closer the better. (Well, not SM58 close but every time you double the distance from your sound source, your levels become a quarter what they were before. You don't mention whether you're shooting video with this or just audio recording but get your mic in as close as circumstances allow. Also, depending on your layout, sometimes it can work to boom out of shot below your face rather than above.

Second, your levels aren't as bad as you think for an original recording. As has been explained many a time on these forums, there are two different standards for the metering of a recording. Ordinary analogue recorders/meters use dBu or dB(VU) where the zero level which is determined by an arbitrary voltage. In that system, you can allow peaks in your recording go above this zero mark, by how much depends on the quality of your gear but between +4 and +8 is fairly normal, allowing lots of overhead for safety. On your digital gear though, they use dB(FS) which stands for "Full Scale". Zero level on the Zoom or on any meters in Audacity is the absolute maximum you can record without clipping distortion.

The norm is to say that 0dB(VU) is equal to -18dB(FS). This means that your -6 level is, in fact, like recording at +12 in analogue which is actually a very hot signal. Normally you want to aim for the "average" of your waveform to be around -18 with peaks up to -10 or so. Trouble is, DAW waveform displays make this look really tiny--you have to go by the meters not what the waveform looks like.

Once you've done your recording, you can then (sometimes) boost these levels depending on what you're doing with the recording. Ending up on a CD, it's pretty common to boost things up to 0dB(FS) (using a limiter to avoid going above that level). For some broadcast purposes, you work to a standard specified by the broadcaster...for example our satellite transmissions, we demanded that audio be limited to +8 (i.e. -10 dB(FS)) to avoid cross talk between satellite channels.

Just as an FYI, I've used an NT1G in the past--I used to carry it as a back up to the Sennheiser 416 which was our primary mic--and results tended to be very similar between the mics so (barring a faulty unit) that shouldn't be the problem.

 

[misko]

Thanks, Bobbsy, great explanation to a newbie like me

I managed to get a good recording with my setup, but what concerns me is that I had expected that I could either stand at least 1 meter away from the mic and still get a good signal, or that I - closer to the mic - could set the recording level much lower than max (70-ish, maybe). But that is perhaps not the way things work ...?

 

[arcaxis]

Had you seen this video and are you close in sound quality.... https://www.youtube.com/watch?v=qoMQeAc6EU4

I have the H4n, but not the NTG1. I do however have a phantom powered cardioid condenser that is close to the same sensitivity as the NTG1 and I hooked it up to see how the H4n did.
I also had to set the record level up around 90-100 to get the meters on the Zoom to get to about slightly above -12. The specs on the H4n show that its preamps don't have a bunch of gain, but just about adequate for most condensers.

When I've recorded into Audacity I've usually "normalized" the audio to bring it up.

 

[rob aylestone]

1m and soft speech means that the levels will be rather low, but the real problem then is that is sounds like it's a metre away and the sound of the room becomes and appreciable component in the signal - which could be good, or very bad. If you experiment, you will find that as you add more gain, you get more noise. What usually happens is that the noise increase and the volume increase from the subject being closer do NOT cancel out. Sound follows the inverse square law - if you halve the distance, it does not double in sound, increases to 4 times the level. So usually the gain setting and the distance need adjusting to get the sound you want.

In the 'good old days' you adjusted the gain until the needle in the meter didn't wander into the red section - or on tape, maybe a bit into the red, and you got to understand where distortion started to creep in.

Dynamic range was quite limited so quiet to maximum was quite a narrow range, and a whisper tickled the meter at the bottom, while a shout pushed it into the red. Digital meters have a huge range, and always just look less busy. If you expand the scale you can see what's going on. Many DAWs have an expand function to make the waveforms fill more or less of the display area. Usually, your ears do the work - they can hear the hiss and spot the distortion better than a visual meter in many cases.

 

[misko]

Thanks!

Had you seen this video and are you close in sound quality.... https://www.youtube.com/watch?v=qoMQeAc6EU4

Yes, I wathced that too, it's a good one The volume in my recording is close to that, and after normalizing it is at least that "good".

I also had to set the record level up around 90-100 to get the meters on the Zoom to get to about slightly above -12. The specs on the H4n show that its preamps don't have a bunch of gain, but just about adequate for most condensers.

Ok, so I suppose my recording is what one could expect from the set up?

When I've recorded into Audacity I've usually "normalized" the audio to bring it up.

That does the job for me, although I dindn't think I would need to.

1m and soft speech means that the levels will be rather low, but the real problem then is that is sounds like it's a metre away and the sound of the room becomes and appreciable component in the signal - which could be good, or very bad. If you experiment, you will find that as you add more gain, you get more noise.

I am suprised I didn't get more noise than I did on record level 100, and Audacity does a good job filtering it. I think most of the noise actually was background noise from not being able to sound isolate the room enough (although I did turn off the ventilation for the entire office - and the coffee machine in the kitchen ...)

Digital meters have a huge range, and always just look less busy. If you expand the scale you can see what's going on. Many DAWs have an expand function to make the waveforms fill more or less of the display area.

Thanks, I'll try that.

Usually, your ears do the work - they can hear the hiss and spot the distortion better than a visual meter in many cases.

I suppose they do - however, as a noob one gets a little preoccupied by the waves, thinking "what if i hi fi enthusiast or sound engineer will listen to this on a much better speaker than I have ..."