After reading posts, tutorials and wiki entries all day, I’ve learned enough to know that I’m still not sure what I’m doing. Here’s my situation: I’m recording rehearsals for a community chorus, which I then podcast. Until very recently we were recording with a FirePod, a laptop computer, and a pair of Samson Audio C03 condenser mics (and Audacity, of course). Yesterday we invested in a Zoom H4 (everything but the laptop was borrowed and the owner wanted his gear back).
We take the recorded file (consider it either a file recorded by Audacity using the old gear, or a .WAV file recorded by the H4, 44 KHz, 16 bit), and applying effects in this order:
- Leveller, set to Moderate and -60 dB
- Compressor, with what I think are the default settings:
Threshold: -12 dB
Attack time: 0.2 secs.
Decay time: 1.0 secs.
Normalize to 0dB after compressing
We export the file to MP3 and upload that to an RSS feed into iTunes (and also linked from our web site). We need to have a clear recording of the rehearsal, so we can clearly hear the conductor’s comments and corrections and the singing and accompaniment.
Is all this what we ought to be doing? Are there other effects we should be using? I inherited this setup from a previous volunteer and know almost nothing about it except what he told me; and he didn’t explain why on anything. Your advice greatly appreciated.
Which Audacity are you using? and what kind of computer?
First, top marks for going with what works, and it’s a very good thing that you wrote down all the settings.
The “leveler” is a brute force volume compressor. It greatly (depending on settings) restricts the differences between quiet passages and loud ones. It does create distortion while it’s doing it, so you can’t use this for a production deliverable. The compressor is a more gentle and graceful way to do similar jobs. I wonder if the compressor is doing much of anything given the order of the tools.
There is another way to deal with this. One of the other helpers found a very nice volume compressor that compares favorably with the ones used at the radio station. The software designer looked at the sucky software tools available, made a list of shortcomings, and then solved the list.
Here’s the whole thread. I open with comparing a radio broadcast with the podcast version of the same show. The podcast version has your problem. It’s really expressive and has wild volume changes during the performance. The radio version doesn’t, and this is how we did it.
Since you are processing the audio after recording, you should record in WAV format rather than MP3.
Rather than the leveler and compressor, I would recommend using Chris’s dynamic compressor: http://pdf23ds.net/software/dynamic-compressor/
Full instructions are included on his web page.
To install it, download the plugin source, and put it in your audacity plugins directory (which should be under the main install directory on windows, and under /usr/[local/]share/audacity/plugins on unix).
Ah ha, Koz beat me to the post - well it’s the same recommendation
Sorry, forgot the basics: Audacity 1.3.4 beta on a Windows Vista Home Premium laptop (Dell, if it matters).
Oddly enough, the Compressor does seem to be doing something: after I run it, but not after I run the Leveller, the waveforms are much broader and spikier and the sound levels are up. I’m noticing on the .WAV file recorded on the H4 that the wave forms are very narrow (vertically) and the sound is quite soft (fine in headphones, but I know that some of our singers listen on their computers from the podcast in our web site…) - I’m used to seeing much wider wave forms from the FirePod recordings (on which I control gain by turning the knobs on the front). I think this means I haven’t yet figured out how to adjust the gain on the H4 for recording in .WAV; the recording I did last week as an MP3 was at quite reasonable volume.
Actually, steve, when I record in MP3 I don’t process afterward - I just post that! I was experimenting to see if the sound would be better recording .WAV and processing on the computer.
I have to get rolling right now (another rehearsal tonight! performance in 2 weeks) - but when I get time I’ll review that thread, thanks very much for the reference.
You may have automatic gain control (AGC) switched on (time to read the manual )
AGC is fine for quick MP3 recordings, but for higher quality, turn it off, set the recording levels carefully (avoid clipping!) and record WAV, then try Chris’s dynamic compressor. The default settings should work quite well, but if you want to even out the volume more, increase the compression amount a little (stay within the range 0.0 to 1.0 or you will get very strange results).
No, I know that I don’t have Automatic Gain Control turned on. (I did read the manual.) I had the gain on the H4 manually increased from the default 100 to 113 or so - recording levels were still very low. I’ll certainly give Chris’ compressor a try but I think tonight I’m just going to record an MP3 again!
I’ve not tried the H4, but I have a Zoom H2, and on that there are 2 level controls - one is a high/medium/low sensitivity switch (a kind of coarse level control) and then an up/down recording level (offering fine adjustment). On the H2, setting the switch to the most sensitive setting it can record pretty low sound sources (such as quiet speech) without having to turn the recording level above 100. There may be something similar on the H4 (same manufacturer, same model series, just that the H2 is the little brother).
The H4 has exactly the same switches and recording level adjustment. Based on the recommendations in the manual, I’ve been using the L switch setting, since the description of the use seems to match my situation. The manual implies that the M setting is for low-volume situations like recording a single acoustic guitar. I’m not very clear on what the H setting is for; I’ve only had the thing 3 days!
For last night’s .WAV recording (I changed my mind about the MP3, I’m going to make this work), I jiggled the fine tuning to a recording level of 115, a little above what I used the night before. The problem with the rehearsal situation is that it has to handle one person speaking (and keep it clear and audible), and also 100 singers singing double forte - also clear; audible isn’t a problem here. 100 singers can make a lot of noise. It seems as though I ought to be able to leave the level at 100, but it comes across so soft when I do. Maybe Chris’ compressor will help that. What does happen when the recording level is >100, anyhow?
I’ve now processed one of my recordings from last night with the dynamic compressor and the difference is amazing! Taking all the defaults, the volume was still low enough that I had to increase the gain on the tracks by +10 dB; but I recompressed taking stevethefiddle’s suggestion of setting compression to .75 and the sound was excellent. In fact, it was a little too loud in headphones; I’m trying the second file with compression at .70 to see how it sounds. Thanks so much for taking the time to help me figure this out!!
You have just stepped into one of the problems with the Zooms. You can’t “ride the levels” during the performance to make up the difference between someone talking into his pint, and the weapons-grade singers. The choices are using the auto gain controller, or adjusting so the loudest performance doesn’t overload. Overload is permanent.
You’re so right, koz. Well, you pay for everything in either time or money, and the Zoom is cheap enough that I guess it’ll cost me some time. I finally decided I was being oversensitive. I think I’ll continue to go with recording levels that aren’t overloading, and play with the dynamic compressor and (if necessary) the gain meter in the Audacity editor to get the volume the way I want it.
Some folks who hang out at the Zoom forum http://zoomforum.us/viewforum.php?f=15 suggest that only the three way L M H switch is actually useful on the H2/H4. Basically, the variable control is after the a/d conversion, so turning up the record level is exactly the same as adding gain in post processing.
If the original poster is willing to spend a bit more, two of these units could be used to record the session, with one set for the low volume sections and one set for the high volume recordings. Then the engineer could edit between the two tracks in post as necessary.
Sounds like the post processing in software is doing a good job though.
Thanks for the information, Fran, I’ll have to check out that forum myself. Unfortunately, the actual owner of the equipment is a non-profit arts organization, so it’s very unlikely we’ll be buying a second H4; I’ll just have to do what I can with Audacity.
You can be the engineer. I’ll watch.
There is a fuzzy rule to allow ten times more editing and production time than the length of the show. So an hour show consumes ten hours of editing. I suspect using the two recorder process would result in that number being much worse. You would be able to hear each cut so you’d need to construct a dissolve or fade at each transition.
You let me know how that works out, 'K?
Sure, I’ll let you know. Just don’t hold your breath, huh?