Well, I did record section 8 and sent it to you under
CS_raw_test, before I read your response here. Without amplifying the volume, I can't hear it well enough to edit out errors
It really is the strangest thing. My headset is clearly picking up the sound quite well, but it registers it as SOOO quiet.
59 dB !!! I've never encountered such a low volume before.
Please check the microphone input icon on top of your Audacity waves window. Is it all the way up to the + sign ? If not, do that, it may boost your volume from the start.
I use Audacity to record, on my Windows 10 desktop, via a USB Corsair headset mic which unfortunately seems to have a factory defect that picks up sound very clearly yet VERY quiet, so I used the MP3Gain app to amplify my reading to an acceptable dB level.
Aha ! Windows 10 is most probably the reason for the low volume. I heard the sound for Audacity is very low with the new upgrade. That's one of the main reasons why I still refuse to upgrade to it
I am not sure if anybody found a solution to it, but lots of people complained about it after the upgrade.
So first I amplified your file with 30 dB to get to 89 dB. Of course the noise was amplified as well in the process. But I did the normal reduction, taking a sample at the end (like you correctly do), and run the cleaning. The result after only one round of cleaning is good in my opinion. I don't hear the rustling noise anymore, do you ?
https://librivox.org/uploads/kitty/raw_test2.mp3 I am on a different computer today today though, so maybe that explains the difference in sound, but for me it's ok.
You know what, I just realized that my computer itself has a very fine microphone built in, which picks up the volume sufficiently, so no headset necessary
my own in-built microphone is not of a very good quality, but maybe yours is better. You can always try. But
make sure there is no "noise cancelling" box ticked, because that makes the sound very tinny.
I am also beginning to see where the sound can start distorting in strange ways if I don't know where to stop editing and nitpicking on my own recording.
yes too agressive noise cleaning can distort the sound. Your settings in Audacity are quite ok I would say.
Things like breathing sounds, lip smacks, rustles, strong fricatives, variations on tone and volume... Yikes. Eventually I'll learn how to leave these out of the reading in the first place.
at least a good hint to get the plosives our of your recording is putting the microphone a bit to the side of your face/mouth, because then you don't "blow into it", but the air passes alongside it. Other than that, it's practice. For me it helped a lot watching the audacity waves appear while I read, I can immediately see spikes when I am too loud or there is a plosive. Soon you will notice a pattern, what looks how
If this occurs, just let the recording run on, and redo the sentence with less loud voice for example. Later you can cut out all the "rubbish". My raw recordings are always some 30% longer than the final version where everything is cleaned out.
I have seen the Librivox recommendations to amplify via the Effect>Amplify... tab option within Audacity, but that doesn't seem to want to let me amplify it so far. It just greys out the OK button if I try to adjust the numbers or sliding scale.
that is because you don't have the box ticked "allow clipping". But it's recommended not to tick it, because distortion may occur if the clipping happens too often. Most likely though it's only because of a few peaks/spikes, where you spoke louder, maybe the first letter of a word only. If you see those spikes, you can zoom in and manually decrease them, so the waves all look a bit more uniformly. Then you can amplify it all in effects without clipping. I only ever use that function to amplify. But I am quite good at keeping my voice level by now even when I speak excitedly, it comes with practice and watching the waves appear.
But what are the "oops" you consider definitely worth clipping out, and are there any that you find less important, or even possibly add to the reading (do sighs ever enhance the storytelling, for example?)
sighs and such special effects are more done in our drama projects where we can act out as much as we want (I love it there, you should try it). In "normal" books we usually try to keep it to a minimum as we should only "read what the author wrote". That said, many readers (including me) love to read in character, especially the dialogues, so the occasional sigh my creep in. But tastes vary, what one person loves another person will hate. In the end you cannot please them all, so read the way you like your text, that's the best advice.
So, step-by-step, my procedure so far has been to open a new track (should I be opening a new Project for every Section? I thought at first the Projects were more of a Folder, but now I think I'm wrong)
oh definitely I would recommend having a special file and folder for every project. I create neat folders for each of my projects and then save my audacity files in the correct ones, so I always find back everything. For example, if I record 3 sections in the rover, I will have a folder with 3 separate files then. One for each section.
record the Intro, Section, and Ending, (is this where I should be doing Save lossless copy of Project?)
I save regularly, because I fear that I may lose the work if the computer crashes in mid-recording
I don't know about lossless. I always save the audacity wav files and when I have edited it all I export it to mp3 and send this. Once the project is catalogued, I delete the wav.files, because they take up a lot of space.
I select the whole track, go to Effect>Amplify>Manage>Factory Presets>Defaults, then OK. This bumps my volume to nearly-sufficient dB levels.
there is a plug-in for audacity that calculates the volume you need to amplify with:
The easiest way is to install an extra plugin in Audacity that will tell you by how much the recording needs to be amplified to reach the desired volume of 89 dB.
Replay Gain plugin for Audacity: - http://forum.audacityteam.org/download/file.php?id=4668
1. Save the file to your computer - your desktop or somewhere where you can easily find it.
2. Now browse to the folder in which Audacity is installed. It's probably in C:\Program Files\Audacity) and then open the Plug-Ins folder.
3. Drag the file you saved into the Plug-Ins folder.
4. When next you start up Audacity, you will be able to select the whole recording (Ctrl+A on Windows), then go to Analyze | ReplayGain, and it'll tell you how much to amplify your file to get to about 89 dB.
From there, I select the last five seconds of silence, go to Effect>Noise Reduction>Get Noise Profile, then select the whole track and return to Effect>Noise Reduction: it sets the factors automatically at Noise reduction: 12, Sensitivity: 6.00, Frequency smoothing: 3, and Noise: Reduce is selected. I leave those settings as they are, and hit OK.
this is correct. I have Frequency smoothing 0 but yours doesn't distort as far as I can tell, so it's ok
I open MP3Gain and upload the MP3 file in there, hit Track Analysis to amplify the volume to the last bit of parameters, and Track Gain to save the changes to the file. Then I open the File folder and double-check the MP3 file in Checker, before uploading.
ah I don't use mp3gain so I cannot comment on that one. I always use replaygain straight in audacity before I export
Sonia