I would first like to thank all the programmers that have supplied plugins like RMS Normalize and ACX Check for their hard work in making everyone’s ACX, and other spoken word, recordings better. I recently watched a YouTube video describing how to copy a segment of quiet room sound and use the Punch Copy/Paste plugins to replace segments of not-so-quiet room sound or other extraneous noise. This process was touted as sounding more natural than a noise gate that can easily replace such segments with silence. It occurred to me that a sort of combination of the two would be ideal. I can imagine a plugin that would replace the low decibel segments (mouth noise, chair squeaks, page turns) between sentences or words with a pre-selected segment of quiet room sound, possibly with the looping function of Punch Copy/Paste. Does anyone know of any work done in this area? Thanks
I expect the answer will be “no”.
For producing high quality audiobook recordings, you need to have a very quiet room to record in. If you find that you need to replace all of the silences between sentences, then you’re doing something wrong (probably your recording room is too noisy).
This patch system would have the same problems that Noise Gate does. High room noise behind your voice. Not between words, but during words. The design would give you chopped high noise words with low-noise, well-behaved interstitials.
It would also be a nightmare to set up. Part of the attraction of a quiet room is capturing the quiet expression and theater of your voice. Noise Gate usually whacks those off because they’re quieter than the full voice.
Don’t move around while you’re announcing. Any change in your voice volume will throw the keying off.
It’s not hard to produce cellphone voice. What’s hard is getting ACX to accept it.
I produced a “Kitchen Table Sound Studio.”
https://forum.audacityteam.org/t/too-compressed-rejection/52825/22
It helps with environment noises and will kill room echoes.
Koz
Thanks for the replies but I am not ready to give up on this idea. First of all, I am not suggesting that the function I envision would be a substitute for a noisy recording space. In fact, I have found that acquiring such a space in the Los Angeles area is one of the most challenging and most important parts of voice over work. However, I am suggesting, as mentioned in my first post, a way to find and replace anomalous noises such a page turns, chair squeaks, etc. with a clean piece of room sound. To that end I am thinking of some adaptation of the Silence Finder or Sound Finder plugins combined with the Punch Copy/Paste plugin to find and replace sections between phrases with a pre-selected segment of audio, probably clean room sound. In fact, I just viewed another video/animation distributed by ACX that describes how to do this process manually so I must believe that others might find this useful as well. I don’t know enough about plugin programming at this time to create such a function but I may very well learn. Does anyone know of anyone that has expressed similar ideas or done any programming toward such a function? Thanks
in the Los Angeles area is one of the most challenging … parts of voice over work.
That’s because Everybody Knows you can record audiobooks from your kitchen table. There was a tiny studio across the street from a breakfast place and I don’t think it’s there any more (the studio, not the breakfast place). I don’t think it was the cash cow they were expecting.
There are studios closer than you think. I once got a very good quality voice track from one of the producers and I asked how he did it. He said, “My Toyota.”
pre-selected segment of audio
Which is the one thing you can’t do. The inserted audio has to be tailored to the space left by the cutting function. The cutting function to be automatic and “know” what your voice is. If you don’t do that, the theatrical timing goes off. Recognizing human voices is usually something Humans are good at.
I just viewed another video/animation distributed by ACX that describes how to do this process manually
Link?
This question comes up a lot and I have no good ideas other than manually editing. Remember this is a forum. Users helping each other.
Koz
I thought there was a way for Audacity to paste and leave the hole size. More research…
Koz
Perhaps you’re thinking of the “Punch Copy / Paste” plug-ins: Punch Copy/Paste
The problem is, that while it is easy for you to see if a sound is your voice, or a page turn, or a chair squeak, it is extremely difficult for a computer to know if the sound is part of the performance, or “just noise”.
Perhaps you’re thinking of the “Punch Copy / Paste”
No, I thought one of the existing tools would do it, but that works.
Koz
Yes, the Punch Copy/Paste plugin, that I mentioned in my post of August 10, could do the “paste-some-preselected-room-sound” part of the function I envision. I am currently looking at the “Sound Finder” and “Silence Finder” plugins to do the “search” part of the function. The issue I am finding with these “finder” functions is that “Silence Finder” only marks the end of a silence and not the beginning. Has anyone done any work on improving the “Silence Finder” plugin to actually create labels that mark the beginning AND end of a silence?
That does seem to be a mistake doesn’t it? Sound Finder does exactly the opposite of what you want, and Silence Finder is missing options. I think Silence Finder was designed to point the operator to the silence rather than key locator tools to change it into something else.
The forum has a section for improving, changing or expanding Audacity.
https://forum.audacityteam.org/viewforum.php?f=20
Note both of those tools are “Sound” tools, not “Voice” tools. Audacity doesn’t know what a voice is.
Koz
Strangely enough “Sound Finder” contains a “based on…” credit to “Silence Finder” as a comment in the code. The two are similar but not the same. I have posted a bit more detail in the Nyquist programming forum in hopes of understanding the programming side of things. Thanks
I was thinking about this in the shower (some of my best work…). If only there was a way to train the system what your voice is…
We do have a tool you can train. Noise Reduction.
And Noise Reduction has a “Residue” setting. So in theory, if you select enough of your voice sample as Profile and Residue it, all that’s left in the clip would be the cat, gasping, mouth clicks and cars going by…
Koz
There’s a long history behind this.
The original effect was written by Alex S. Brown and was called “Silence Marker”, and it was designed to mark track breaks when copying audio tape to digital. (For those too young to remember: Audio Tape). The effect would place a label just before the end of the detected silence, so that the recording could be split with a little bit of lead-in before the start of the song. This was before Audacity had “region labels”.
Some time later (though we are still talking ancient history), region labels were added to Audacity, plus the ability to use “Export Multiple” with labelled regions.
One of the limitations of “Silence Marker” was that if you have a tape with long gaps between songs, when the tape was recorded in Audacity and split into separate song files, there would be long silences at the end of each song. Jeremy R. Brown wrote “Sound Finder” (based on “Silence Marker”) to address this limitation by using region labels that start just before the start of the sound, and end just after the end of the sound.
Since then, there have been many discussions about merging Sound Finder and Silence Finder into one effect. I personally spent a lot of time writing improved versions, but the senior QA person at that time blocked every suggestion. I’ve probably still got all those versions somewhere. I’ll see if I can find one of the better ones.
Here’s one that can mark sounds or silences with either point labels or region labels (I’ve also updated it to a “version 4” plug-in):
mark-sound-silence.ny (3.21 KB)
Thanks Steve. I thought that somebody may have worked on those very useful Sound/Silence Finder functions. I look forward to trying out your plugin but I have to mention the misspelling of “slience” in the file name.
Doh!
Fixed.