Zoom Gear & Home Recording Forum

Hello,

I'm brand new to this forum and really brand new to recording as well. I have just purchased a H2N and I'm learning how to use it. So far it seems easy and I'm liking the results but I'm hoping to learn more.

I'm not a musician but I am married to one and I got the H2N so that I could record her gigs. The recordings are just for me to listen to. I'm not trying to do any production. I started out borrowing my daughters H2 and it did an ok job but I had a devil of a time with the recording levels. The gain dial on the H2N has made a world of difference already in the quality of what I can capture.

I've read a bunch on this forum already and it's been a big help. The key things I've picked up are:

Set the gain so that the average levels are around -12db.
Keep the recorder closer to the band to minimize room noise.
Use the MS and adjust the width.

Here are my questions:

Can you reduce the room noise by recording 4CH with the MS pointed at the band and the XY capturing the room and then subtracting the XY from the MS?
Is it better to record at 48Khz and resample to 44.1 if burning a CD or should I record at 44.1 right from the start?
Are there any standard post processing tasks other than normalizing that I should be using? EQ? Compression?
Any other standard practices I should be using?

I have a sample and would really appreciate feedback and advice on what I could do better.
The Jane Mutiny live at Sticky Lips

Thank you for all the great info already posted on this forum.

Scott

No, the cancellation idea won't work I'm afraid. But you should try it anyway. You can use Audacity for the mixing. You'll invert the audience side then try adjusting the level to see if you can get some cancellation. I hope I'm wrong about it not working.

It's better to record at 44.1 and skip sample rate conversion. There has been (so far as I know, and I try to keep up) any objective evidence that anyone can hear the difference between 44.1 and any higher sampling frequency.

There are no standard practices I can think of in audio, everything is context dependent.

My one suggestion, and it's kinda mean because it adds a chunk of complication - you can record from the board and the mics at the same time. The line input replaces the front mics, you'll need some kind of level control going in, I would recommend taking your direct signal from the headphone output of the mixer. This all assumes that the band is miked up and mixed. If all the instruments are miked you are golden, if some are acoustic only (like the drums and bass maybe?) the mixing might get a bit tricky, but this should deliver a more solid vocal.

You're doing very well already with good levels and no clipping and a good instrument balance, but it's always interesting to try to kick it up a bit.

Fran

48kHz is the default rate for /video/ PCM formats (DV-AVI) as it fits evenly into "cinema" frame rates (24fps -- 2000 audio samples per video frame). But any compressed video format is using some form of MP3 type encoding, which is no longer a confirmed match to any video frame rate.

Nice blues and nice recording And I like to hear the audience!
Nothing more to do...
Well, maybe... Forget EQ, maybe a little compression or limiter? RMS Power is about -19dB and it's OK to listen at home with good hi-fi, but sometime there's a need to have more loudness - for radio broadcasting or just to give it somebody to listen - in that case I try to achieve -16dB.

Fran,

Thanks for the reply. I thought I would be able to record to the MS and Line In simultaneously but didn't know how to begin. The posts I read suggested that Line-In has higher noise. I also wondered about impedance matching and levels. If the Gain dial controls both the mic gain and line-in gain then I assume I'd need an amp or attenuator before feeding into the H2n. I'm clueless about these.

My other obstacle is that I don't have access to a mixer to play with this. I'll have to wait until the next gig and then hope that the sound guy is willing to work with me.

I'll give the multichannel recording a test too. My goal isn't to eliminate the room noise, just reduce it some. I made one recording where the room noise was so bad I just scrapped the whole thing. I have also tried putting the recorder right in with the band near the drums and relying on the monitors for sound. That was also a bust. Drums were way too loud and monitors way to quiet. There must be a real art to selecting the right location for the recorder.

Thanks for your help.

Wulfraed,

I'm not quite sure what you were telling me. Do you feel that it makes sense to record to 48K and resample for other uses?

I'm wondering about the sweet spot for settings on the H2n. I know it's capable of recording at higher sample frequency and bit depth but if the quality of the source has to warrant the extra info. So far I don't get a sense that a live room benefits from higher settings. I'm very willing to explore it though (with guidance).

Scott

TomPL,

Thanks for the suggestions. That's the kind of thing I was hoping for. I get the basic idea of what normalizing does but there are so many parameters to adjust. In Soundforge you can select RMS or Peak. From your comment I would use RMS as my target. Am I right in assuming that normalizing the RMS could result in pushing the peaks to clip? I guess this is why there are plug-in chains. Would I use a compressor to tame the peaks and would I do that before or after?

Scott

Line in does not have higher noise. Impedance matching is basically a non-issue at line level. Levels are an issue, that's one reason I suggested using the headphone out on the mixer. Otherwise you do need an external attenuator in order to control the line in level. As you say, there is no way to control the line in gain separate from the mic gain. It's also the case that line level from the average mixer is too hot for the H2n line input.



You may also need a fairly long cable if you want flexibility in placing the recorder while it's connected to the mixer.



You're certainly welcome and good luck with your recordings.

Fran

I mention it only as -- IF you are using it for ambient or other VIDEO sound, the 48kHz fits the video frame rates easier.

"In Soundforge" - I don't know, I use Adobe Audition and do it separately in 2 ways:
1. after "normal" normalization use compression (you have to set the ratio and the threshold, eg. 6 and -3dB) and then normalization again. You may repeat this process if RMS is not as what you want, but be careful to not overdo...
or
2. just use Limiter (as a Vst Plug-in) with threshold you want - limiter is dooing likely generally the same as (1) but is written by a "specialist"

A compressor approaches a limiter when the /ratio/ approaches infinity -- that is, the above-threshold slope essentially turns flat. You lose all dynamics for the audio that is above the threshold.

As for applying them.. The sequence I'd suggest is:

Normalize (this puts the peaks at or just below 0.0dB)
Compressor (ratio and threshold adjusted to cover the range you wish to compress)
Normalize (if the compressor doesn't do an automatic or manual post-compression gain boost)

A fairly fast/deep compressor setting would be something like: 6:1 @ -24dB. In the source, a signal at 0.0dB (24dB above threshold) will be reduced to -20dB (24/6 + threshold). -6 => -21dB, -12 => -22dB, -18 => -23dB, -24 => -24dB. You've squeezed the top 24dB of dynamics into a mere 4dB range. You can now apply 20dB of post compression gain to move the peak back to 0.0dB (and audio that was at the -24dB point is now at -4dB, and ALL signal below the threshold has been raised by 20dB).

As stated, that's a fairly heavy compression... 3:1 @ -18dB is probably more common. inf:1 @ -3dB would be a limiter.

I would NOT recommend "redoing" a compressor. Instead, "undo" the first try, change threshold and/or ratio, and try again. In the example of 6:1 @ -3dB, the top 3dB is compressed to take up a range of half a dB... If you then reapply the compressor (assuming renormalized) what had started life as the first 3dB and is now 0.5dB range becomes a 0.08 dB range.