Just installed Audacity 2.0.3 on XP Home, Service Pack 3. See the attached images for the volume adjustments I’m able to make.
Under Volume Control, the Input Monitor option is grayed out. Is that normal? It’s stuck around the 80% level, which seems kind of arbitrary. (It’s still grayed out even if Mute is unchecked.)
However, if I go to Options → Properties and switch to Recording Control, I can slide the Volume level for Line In no problem. This slider is also controlled by the Input Volume slider in Audacity, and vice-versa. What’s weird about this adjustment, though, is that sliding it all the way down doesn’t silence the input completely. I can still hear sound at about half the volume of when the slider’s at the top. Is that normal as well?
Not exactly “normal”. Rather than a single “input monitor” one would more often expect a separate output slider in Volume Control for each of the inputs in the Recording Control.
When you say “hear”, do you mean hear it while you are recording (without enabling Transport > Software Playthrough in Audacity)?
Or do you mean when the Audacity and Windows input slider are both zeroed, you still record a half wave height in Audacity? The latter would be odd, but if your motherboard sound drivers are not correct, anything can happen. See http://wiki.audacityteam.org/wiki/Updating_Sound_Device_Drivers for help checking if you have appropriate audio drivers (supplied by the motherboard or computer manufacturer).
I still hear sound when recording (without enabling Transport > Software Playthrough in Audacity). I’m pretty sure my drivers are correct, but I’m not an expert.
See the images below. In the first one, I’m not recording yet, just monitoring the Input Level. Notice the Input Volume is at zero, but the red bars are still showing sound.
In the second one, I’m recording with the Input Volume at zero, and you see the waveform is very small. But when I play the recording back, you can still hear the sound.
The third one shows what a recording from the same source looks like with the Input Volume all the way up to one.
However, none of this may actually matter. I am able to successfully make recordings, I just want to be sure that the input level is set correctly. Is it really more of an art than a science? The manual says to adjust it so that that peaks are around -6 db, but I have six 30-minute audio cassettes (of identical quality) I’m trying to record, so it’s really tough to know whether or not there is somewhere in those six hours that will go above zero.
Perhaps this is because the input monitor in Windows is not zeroed (despite being muted) .
You could check the link I gave if you have not already done so.
It’s the “noise floor”. You get that noise in the background when you record.
The recorded level looks high when the input slider is zeroed, considering that the level is not that high when the slider is maximised. That said, the level you are achieving when the slider is maximised seems a good, safe level.
The received wisdom is that it is better to record (and play) at a high level when using a digital volume control in a computer, and attenuate when needed using an analogue control (such as the output knob of a tape deck if you are recording from its headphones output, or the output knob of an amplifier if you are sending computer playback to a HiFi system).
Probably this does not matter too much if you use the default 32-bit float quality in Audacity as you are doing. Also on Vista and later, Windows I believe always upsamples all audio to 32-bit float before processing it (with the Windows audio programming interfaces that Audacity supports). But it may still be worth considering.
In the absence of an automatic input attenuator somewhere in the chain when 0 dB was approached, yes. But an attenuator is not very “artistic”, more perhaps for broadcasting or other applications.
Can I feel the resident sound engineers waiting to criticise?
I hadn’t considered that – definitely a possibility. As long as my source (a cassette deck I have yet to take possession of) doesn’t go into clipping territory when the Input Volume is at zero, I should be in good shape, right?
Thanks for giving such a detailed response. Considering your above two statements, I would guess I’m in OK shape to proceed with recording. Then, of course, the only remaining question would seem to be, “Do I normalize the recordings afterwards?” Are there pros/cons to this? I was unfamiliar with the process before downloading Audacity, but the manuals/guides seem to say this is a positive step.
A few non-continuous clipped samples are usually not audible.
If one song or section does record clipped then there is always the possibility to record that section again and blend it into the correct position.
There should be no cons if you use 32-bit float quality. Normalize (including DC offset removal) as soon as the recording is finished, and normalize again as the last step ( between -1 dB and -3 dB is preferred for this final normalization).
Do you mean I should normalize first and then after I apply any processing? I don’t plan to alter the file in any way, save for maybe trimming beginning/ending silences. The recordings I have are in pretty good shape to begin with.
Personally I would recommend using the Normalize too immediately after capture - but just with the DC Offset Removal and no amplification (uncheck the Normalization check-box in the dialog box).
Your Left & Right channels look pretty balanced - if theye weren’t that could be a reason for using Normalization on independent channels to balance them out.
Then do all your processing/editing.
As the final step before exporting production files (WAV, MP3, AAC, whatever) the use Nornalize (or Amplify) to bring the recording to th erequired level. I wouldn’t go all the waqy to the theoretical maximum of 0 dB (the default for Normalize IIRC) as some players don’t like that - topping out at -3 to -1 dB will be plenty loud enough, as Gale suggests.
In that image I would suggest that your recording level is set a little high. We normally recommend aiming for a peak level odf around 0.5 on the waveform display (which corresponds to around -6 dB on the meters).
Try turning on the show clipping display with View > Show Clipping - this will display red lines when your recording touches or passes 0 dB and goes into clipping (as your image seems to show).
Thanks for this follow-up info. I think the manuals are pretty clear on using amplify vs normalize, but I shouldn’t have to do it twice if I’m not doing any effects processing, right? As I said, maybe just trim a few seconds of silence at the beginning/end of the recording, but that’s it.