a total mess

I recorded a story for Librivox, and downloaded it to my computer. Then I imported it back to Audacity, so that I could start making corrections. What came back was not the same as what I put in. The wave pattern was all smushed together, with no gaps for breaths or silences. Without those gaps, I can make no further changes. So there’s that.
Secondly, I was told I have all kinds of background noise. I’ve already seen how I am SUPPOSED to reduce the noise. I go to what seems to be the end of my recording (though in its present state, that’s difficult to tell). I isolate a section at the end for a noise sample. I go to Effect/Noise Reduction, and I click the “take a noise sample” button. And every time I get an error message, telling me my sample is “too short”. Assuming I am correctly taking the sample at all, it appears to be a good ten seconds or so, which should be more than enough. (I’m assuming ten seconds, if the distance between 1:20 and 1:30 is ten seconds.

Possible explanations for smooshing …
#1. Your recording device may be in voice-activated mode, so it does not record “silent” gaps.
#2. (less likely option) You’ve used Audacity’s “truncate silence” feature.

I don’t know anything about LibriVox…

Audacity has a [u]Recommended audiobook mastering process[/u] and a there is an [u]ACX plug-in[/u] for checking the levels, including the noise level. But if you don’t have silence (just noise) it’s just going to measure the quietest sound and fail.

It’s intended for to meet the [u]ACX requirements[/u] but I assume the LibriVox requirements are “similar”.

Then I imported it back to Audacity, so that I could start making corrections. What came back was not the same as what I put in. The wave pattern was all smushed together, with no gaps for breaths or silences.

Are you staying LibriVox messed it up? And you didn’t keep a copy?

You should keep a copy in WAV format (probably multiple WAV files) and you may also want to make an Audacity Project depending on your “workflow”. It’s also a good idea to keep your original “unprocessed” files, although some simple editing (cutting and pasting, etc.) is OK. And it’s OK to delete something that’s “trash” if you’re sure you’re never going to want it.

ACX has a strict requirement for the noise floor, and they require some “room tone” at the beginning so they can measure it. They say it should be no more than 5 seconds but they don’t actually give a minimum. I assume 1 or 2 seconds is good. The room noise has to be below -60dB but if you have “dead silence” they will reject you for too much artificial processing. LibriVox MIGHT not complain if the background noise is too quiet.

Noise is the 2nd biggest challenge, after “talent”. :wink: -60dB is a ratio of 1:1000 so it’s “not easy”. If you’re not recording in a soundproof studio, you’ll probably need some noise reduction. There is some information about noise reduction in the recommended procedure. But if the noise is bad, the processing can create artifacts/side effects and, “The cure can be worse than the disease.”

It’s best to prevent noise as much as possible. The biggest source of noise is acoustic background noise in the room but you’ll also get some noise from your preamp/interface. You’ll get the best signal-to-noise ratio by speaking loudly and confidently, fairly close the mic (for a strong signal).

Of course most home recording artists use a quiet room and record during a quiet part of the day. People often turn-off the heating & air conditioning and any other noisy appliances. Sound absorption is fairly cheap and easy to add and that can help a little with noise but it mostly helps if you have a “live” room with too much sound reflection. Actual soundproofing is expensive and it involves reconstruction! :frowning:

ACX has a “strange requirement” that the peaks shouldn’t exceed -3dB. With most audio production it’s OK to hit the “digital maximum” of 0dB as long as you don’t “try” to go over and clip (distort). If LibriVox doesn’t have that -3dB limit you can alter the procedure.

[quote}I go to Effect/Noise Reduction, and I click the “take a noise sample” button. And every time I get an error message, telling me my sample is “too short”.[/quote]A few seconds should be enough. You do have to select/highlight the noise sample.

[bb]P.S.[/b]
This stuff is time consuming and your 1st project is probably going to be a nightmare and even more time-consuming. :frowning:

The makers of home microphones make you think you can buy their microphone, connect it, announce an audiobook, and retire to a nice cottage.

It may be more involved than that. If you are going to record your work on the computer, you may need to make sure that Audacity or the audio program of choice is the only program running. There are a lot of forum help postings with funny sounding voice complaints because of two or more computer programs are fighting.

Also, you are three people. The Producer, with the important decisions and the checkbook, the Recording Engineer, who knows what all those screen patterns and colors in Audacity mean, and the Performance Artist, who we hope is talented enough that people will pay to listen. Large studios have a lot more, but that’s minimum.

Koz