LibriVox Forum

annise wrote: ↑July 17th, 2021, 12:55 am The reason I wrote what I did was for Phil's thoughts - I just felt it was important for him to mention clipping if he made a video for this effect because we have had recordings from people who have fixed the volume by just lowering it after it was clipped - and it sounds terrible
And we are talking about newbies - there is no control over who watches the videos. We all learnt how not to clip quite quickly. That is why I prefer to tell newbies to look at their waveform first

Anne

Yes, absolutely, looking at the waveform tells us a lot, and it can soon be learnt by new readers that sudden peaks are to be avoided. Then we can remedy the peaks by use of voice and/or editing. A properly set up compressor can lower the peaks without reducing the normal level - but these compressors can be hard to set up correctly, and are probably too much for most people, especially new readers. I find them hard to set up as well, as there are so many parameters to consider. I think Checker requires 86dB to 91dB which gives a certain latitude and you do not have to aim for 89db as 86dB or 87dB is perfectly acceptable.

For a long time before a PLers husband came up with checker, we advised people to check that the waves on Audacity averaged about the .5 mark and that worked pretty well . There is a certain amount of levelling that can be done in the cataloguing stage as long as it doesn't cause clipping. Readers can be quite successful without understanding anything much about audio-engineering or having a degree in English Literature or voice production, they just tell a story well .

Anne

annise wrote: ↑July 17th, 2021, 5:02 am For a long time before a PLers husband came up with checker, we advised people to check that the waves on Audacity averaged about the .5 mark and that worked pretty well . There is a certain amount of levelling that can be done in the cataloguing stage as long as it doesn't cause clipping. Readers can be quite successful without understanding anything much about audio-engineering or having a degree in English Literature or voice production, they just tell a story well .

Anne

Yes Anne. I still recommend just looking at the wave form to get a good idea of it's height. I still recommend my video on this as the simplest way of getting an acceptable volume. Noting technical involved which is hard for a lot of people. The reason I am kind of excited about this Edit tool is the perennial problem of volume. New people are told their recording is too low at 83 bB or too high at 93 dB but they haven't a clue as to what those strange things are and get frustrated looking for a 'dB meter' inside of Audacity. Of course there is nothing inside Audacity like that. Then they must install Replay Gain and/or the Checker and learn how to use that. Tedious for all of us.

This Effect just simplifies things; I can see that my track looks ok, then apply it and voila, it's very consistently near 89 dB. Spikes have never been much of a problem in my experience with recorders, low volume has been the never ending problem.

Is this tool in the new versions of Audacity?
Which version are you using, Phil?

I'm still on an old version (2.1.2) and don't have it, so I can't play with it to check it's limitations/drawbacks. But I'm covered by InTheDesert's answer

Most new readers will have the newer audacity versions so I think it'll work.
It does sound excitingly good and simple!

I've got 2.4.2 and it's there. (You may have to enable it.)

I am using 2.4.2 and it has showed up there. The latest version has it too but I still use 2.4.2 'cuze I am lazy.

I did check the Add/Remove plug-ins but didn't see it there. Maybe I can download it.

I have two questions related to using the compressor in Audacity:

Does it matter whether you noise-clean or use the compressor first?

Is it okay to use the compressor twice? For example, if I compress a recording first, then run it through Checker and find that it's not quite loud enough, then amplify it, then see that it has some high peaks after amplification, would compressing it again cause any problems? Nothing sounded amiss to me when I tried this, but I don't have the most sophisticated headphones and speakers.

Thanks!

I don't think it would matter if your did compressor or noise cleaning first. Compressor only really reduces volume of the loudest things - hopefully your noise levels are not so loud that they would be affected at all by compressor.

Yes you can run compressor multiple times. Better to do that at low settings than to try compress too much at once. I often have to do light multiple compression for spiky parts in plays where I can't get enough volume to match other parts without knocking down the tallest spikes.

Amplifying between compressions should not have an effect. That is what I do when I do the multiple compressions - compress, see if I can amplify enough to be at proper loudness - without clipping, compress again, check again. The key is not clipping when you amplify - clipping is bad and can't be undone by deamplifying or compressing - you have to use control-Z to back up and undo the amplification. (Audacity should prevent you from amplifying enough to clip unless you check a box that you want to allow it - and you should not do that. Also - do not let compression automatically amplify - cuz that can clip without warning.)

Thanks, Todd

One thing I've noticed with Audacity's compressor is that it tends to "fade-in" the beginning of the recording. It compresses the first little bit, regardless of how quiet it is. So what I've been doing is going the beginning of my file and generating 1 second of tone (doesn't matter the frequency or size/volume/amplitude), doing the compression (and you can see the shrinking on the generated tone's wave form), then deleting the 1 second tone. (I often do noise removal on the tone, too, because it really doesn't matter since I'm going to delete it anyway.)

TriciaG wrote: ↑August 5th, 2021, 6:05 am One thing I've noticed with Audacity's compressor is that it tends to "fade-in" the beginning of the recording. It compresses the first little bit, regardless of how quiet it is. So what I've been doing is going the beginning of my file and generating 1 second of tone (doesn't matter the frequency or size/volume/amplitude), doing the compression (and you can see the shrinking on the generated tone's wave form), then deleting the 1 second tone. (I often do noise removal on the tone, too, because it really doesn't matter since I'm going to delete it anyway.)

I've noticed this too, but not just at the very beginning of the recording. I've also known it to happen at the beginnings of paragraphs or anything after a significant pause. I don't know why this happens, or what settings I could change top prevent it, but I am careful with the compressor these days.

(And, of course, it may not be that it would happen to everyone. Maybe just something about my speech patterns, who knows?)

Thanks, everyone! I hadn't noticed a fade-in happening with quieter parts of my recordings after a silence, but it's definitely good to know to be on the lookout for that.

Hi Phil, I have seen your tutorials for using the "checker" but I can never seem to find the correct forum. Once I leave that page I get lost! Any suggestions on how to find the correct page for the tutorials. I am really new here and need direction. Thank you

Roselbex wrote: ↑August 21st, 2021, 3:37 pm Hi Phil, I have seen your tutorials for using the "checker" but I can never seem to find the correct forum. Once I leave that page I get lost! Any suggestions on how to find the correct page for the tutorials. I am really new here and need direction. Thank you

Sure. Here is the link that I use for my tutorials https://wiki.librivox.org/index.php/Instructional_Videos

This should be the one on using the Checker program. https://www.youtube.com/watch?v=QKVOYsaDoRo&feature=youtu.be

Keep asking if this does not answer your questions, it is easy to get lost here.

Scarbo wrote: ↑August 4th, 2021, 5:52 pm Does it matter whether you noise-clean or use the compressor first?

I also don't think it makes a difference.

Scarbo wrote: ↑August 4th, 2021, 5:52 pm Is it okay to use the compressor twice? For example, if I compress a recording first, then run it through Checker and find that it's not quite loud enough, then amplify it, then see that it has some high peaks after amplification, would compressing it again cause any problems? Nothing sounded amiss to me when I tried this, but I don't have the most sophisticated headphones and speakers.

I tend to save my edited chapter (in the Audacity native format) just before I do compression and the final level setting. Then I run a gentle filter to reduce anything below 100 Hz (because: Why Not?), run the compressor (generally at 2:1 because my voice generally doesn't need more), and make a guess at my final amplification. When I look at the volume with Checker it's usually a bit off, so I can go back to my saved file and try again with settings that are closer to ideal.

But that's just the way I do things. The order in which you do things and your settings will differ from anyone else. Just pick some way to do things that makes sense to you and stick with that unless you find some reason to change.