Kiadawel wrote: ↑July 7th, 2021, 3:33 pm
adrianstephens wrote: ↑July 4th, 2021, 1:47 am
PL OK. Clearly read with great animation!
Advisory: The checker reports "a moderate level of background noise", which sounds like a computer fan, and is at about -50dB.
You don't need to take any action here, because it doesn't interfere with comprehension. But if you do want to polish your technique,
you might try adding noise reduction to the next recording you do.
Here's a video explaining the technique:
https://youtu.be/KsFkmvF-9d0
Thank you for your notes (both on polishing technique and pointing me in the right direction regarding how ProofListening works)! I truly appreciate it! I haven't been successful so far in finding the same level of detail you were able to see with the Checker program (I don't have a Background Noise reading in my version, for some reason) but I will watch the video a few times through and play along with my audio and see if I can't get it to clean up its act a bit. If you didn't cover this in the video -- do you happen to know if it's better to run Noise Cleaning before or after Normalizing with the "Replay Gain" plugin? I've been using that to keep my louder spots in check.
Here's what I do:
1. Noise reduction - as shown in my video.
2. Noise gating.
3. Editing mistakes. I also flatten the audio between lines to zero as I'm going through it.
4. Truncation of silence
5. Compression (factor of two) - usually. Certainly for dramatic readings, less certainly so for other works.
6. Normalization (you can use ReplayGain, I've now switched to Loudness Normalization)
7. Export
It might see a lot, but most of these steps only take a few seconds to perform.
The only truly necessary step is the editing. My setup is a shed in the garden hung with blankets to reduce echo.
It's not entirely quiet, and I have to remove road noise and the children's schoolyard sounds, plus the occasional
overhead light aircraft, the neighbour's lawnmower.
The flatten to zero step is necessary only because I use the automatic truncation of silence. When I'm reading a DR,
there's usually a lot of silence between lines as I'm searching for the next line to read. I like to truncate that silence
to a predictable value of 4s. This only really works well if there are no spikes between lines - hence the zeroing.
I use compression particularly when I put a lot of "acting" into my voice to reduce the level difference between
shouting and whispering some.
Now you've got the link to my videos, you can see explained some of these steps. If you go to the Librivox Wiki page,
you can see many excellent videos by PhilC, which are generally shorter and easier to digest than mine.
Silverquill has given you additional pointers to using checker. Don't believe what he says about expertise,
he's been around the block, whereas I'm still wet behind the ears.
Hope this is of help to you. If you're interested in developing your technique, then go for it. If you find it a chore
and an impediment to reading, then carry on doing what you're now doing. We care most about the reading, and
least about its technical quality.
If you're any questions, as a retiree, I have all the time in the world to talk to you.