How Do I Limit Spikes and Raise Volume of a Podcast

Hi there.

I am trying to figure out a better way to adjust MP3 files of a podcast that I listen to. These are five year old files that he is not going to remaster, I’m afraid. The podcaster has some very loud effects at the beginning and end of his podcast and occasionally during the episode. Unfortunately, I can’t just use MP3Gain on the files because the spikes compared to the narration are too high. MP3Gain just makes the loudest part louder and doesn’t raise the narration enough to be heard over road noise. I have to have the volume up to max and I still have trouble making out what the guy is saying and I’m afraid I’m going to blow my speakers when the sound effects play.

So what I have been doing is using “limiter” effect in Audacity to cut the peaks to -9db and then exporting the MP3. I then use MP3Gain to raise the volume to 95 db which is a level that allows me to use medium to medium high volume in my vehicle.

I know that there has to be a better way to do this all in Audacity, but I can’t seem to find any information about how to do this. I added Chris’s Dynamic Compressor plugin to my plugins, but I have no idea what any of the adjustments do. I can’t find any written guides as to how to use it. I played with the sliders and listened to the sample but I couldn’t tell if there was any difference at all. I don’t even know if it is the right tool for what I want to do. I don’t really do audio. I’ve used Audacity to make ringtones and radio PSAs, but not for any real adjustments.

Can anyone point me in the direction of a tutorial on how to do what I’m doing better?

I appreciate any advice or direction anyone has for me. Thanks in advance.

I have no idea what any of the adjustments do.

It launches from Effects > Compress dynamics…

You only need the first one, Compression ratio. Increase that one from the default 0.5 to 0.7 and Chris’s compressor will turn any unruly podcast into broadcast safe. Chris designed it so he could listen to opera in his noisy car. I use it so I can listen to unruly podcasts while hiking.

It’s amazing how that setting produces well behaved shows. I used to grab a radio show off-air and listen to that. They starting publishing a podcast version without the brodcast noise, but of course, it didn’t have the broadcast transmitter processing, either. It was unlistenable. One performer mumbled in his beer and other had a thermonuclear laugh. Sound familiar? One pass through Chris produces a show all but indistinguishable from the broadcast show.

It does have a bug. Leave or put something at the end of the show you don’t care about. Chris is a look-ahead processor and doesn’t like sailing off the end of the show. It can produce odd effects damage. In my podcast case, there is coming attractions and other production and station stuff at the end that I just cut off later, after Chris gets done with it.

The bug is unlikely to be fixed. Chris reached End of Life.

Koz

You should know that Audacity doesn’t edit MP3. It converts the content into its own super high quality internal format and edits that. Then it has to make a new MP3 when it’s done, thus increasing the compression distortion. So don’t try to replicate the original file sizes. Go for a higher MP3 quality—larger file sizes so you’re not likely to notice the change.

Koz

Another plugin worth trying for speech is “LevelSpeech2.NY”, it has a fast-acting limiter which should cap the loud noises.

That is the guidance I needed. I took the episode I am currently listening to and grabbed a copy off Twamley’s site. I ran it through Audacity using Chris’s Compressor using the default setting except the compression ratio changed to .7. I hopped in the truck and ran to the store. The volume is comparable to what I was getting with my hatchet method. Here is what the cut and gain looks like above Chris’s effect.
Hachet v Chris.PNG

It does have a bug. Leave or put something at the end of the show you don’t care about. Chris is a look-ahead processor and doesn’t like sailing off the end of the show. It can produce odd effects damage. In my podcast case, there is coming attractions and other production and station stuff at the end that I just cut off later, after Chris gets done with it.

I didn’t get any weirdness at the end of my test podcast. What should I look for?

Chris reached End of Life.

I saw that. The site I found it on said it was hosting the effect out of respect. He lives on, though. I never knew him and here I am feeling sad that he is gone.

You should know that Audacity doesn’t edit MP3. It converts the content into its own super high quality internal format and edits that. Then it has to make a new MP3 when it’s done, thus increasing the compression distortion. So don’t try to replicate the original file sizes. Go for a higher MP3 quality—larger file sizes so you’re not likely to notice the change.

I’m not sure what you mean here. I realized that Audacity does its work in another format. Which is why we have to export into other formats instead of save the files directly. But I have always just clicked export as MP3 and never even gave settings any thought. Ringtones and spoken content don’t exactly beg for high fidelity tweaking.

Should I do something about the saving options or is there another dialog I should be looking at?
MP3 Export dialog.PNG
I really do appreciate the time you are spending here to help me.

Thank you for the suggestion. I’m going to try that effect as well on what has become my reference podcast episode. I’ll do some side by side comparisons and see what I conclude. Of course, when you go listening critically to audio you suddenly notice things that you didn’t before. :laughing:

You realize, I will probably be back with questions… It’s like feeding a cat.

One note that Chris was designed for a music and speech mix of unknown initial volume. Levelspeech was designed to process spoken word podcasts, etc. It’s unknown how it would react to music interstitials and interludes.

Post back.

Another very common shortcoming is a tool taking some desirable action from a defined volume. Defined Volume is exactly what a live production doesn’t have. I started recommending Amplify or Normalize at the default settings just so any follow-on tools would have a common starting place and everybody would get the same result. This can be less successful on a phono transfer with cat-hair pops or someone dropping the microphone in the middle of the show. Amplify would think that was the volume of the whole show.

There is a little-known setting in some compressors of knowing when to give up. That sound can’t be part of the show, so don’t even try.

I think some of the BBC processing did that. They didn’t just brute force everything to a standard volume. They picked the most noticeable features of the sound. That was the surprise in researching the BBC PPM instruments.


I didn’t get any weirdness at the end of my test podcast. What should I look for?

I’ve never experienced the damage, but on a guess, I’d say if there was any quiet noise at the end of the show, Chris might try to meet and match the noise volume by badly changing the last volume of the actual performance.

“…and so that brings US TO THE END OF OUR SHOW /p/ SHSHshshshshsh…s…s…”

I don’t know what Chris is thinking about, but I know he was really good at this and my guess is based on how the rest of it works.

Koz

The last setting, maximum amplitude is pretty simple. The default is for the volume peaks to stop increasing at 99% of the total volume. That’s the adjustment at a different setting that allows some people to use Chris for Audiobook Mastering.

Koz

There is a “talk” episode coming up where Zack and an associate discuss the subject. Usually, Zack is loud and his friend is not. So I am going to point Levelspeech at that and give it a hearing when I get through the First World War (Just a few more hours to go). There will be some music to start and end the podcast if he stuck to form. It should be a good test.

I’ve never experienced the damage, but on a guess, I’d say if there was any quiet noise at the end of the show, Chris might try to meet and match the noise volume by badly changing the last volume of the actual performance.

“…and so that brings US TO THE END > OF OUR SHOW > /p/ SHSHshshshshsh…s…s…”

These podcasts begin with music and end with music, originally very loud music compared to the rest of the podcast. That is why I ended up trying to adjust them.

I am going to go back an try Chris’ on one of those early episodes. I think it will be an interesting test.

I will report my findings. Thanks again.

But I have always just clicked export as MP3

Even if you deal with shows that aren’t museum quality, you might give a thought to your MP3 settings. MP3 always creates sound distortion. You don’t get those tiny file sizes for free. It deletes sounds that are too quiet to hear anyway and in more extreme settings, converts what you gave it to approximate sounds simply because it’s easier to compress.

MP3’s chief talent is that most times you can’t tell it’s doing that—if you do it once. If you MP3 compress an MP3, it does it twice and if you really offend the sound deities, your new show is unsuitable for publication.

If you do produce paid work, for example an AudioBook, it is recommended you export the work as WAV (Microsoft) 16-bit. You can make those into anything else you want. That’s a perfect quality archive.

It’s easy to miss the damage. Open one of your stereo shows and export it as MP3 32-bit constant. That will give you a bad compression and an idea what it sounds like when you do it wrong. Bad cellphone, talking into a wine glass.

If that’s a total disaster, open one of your MP3s and export it at 64-bit constant. 64 is the bare minimum for a stereo show assuming you started with a perfect WAV.

MP3 can be a time bomb. Produce a work for someone and export it as MP3. Open it later and do some production editing, filtering and effects. Export it as MP3 and send it to the client. If the client makes a new MP3 as a product, that MP3 will very likely be trash.

A poster used to make sound files for broadcast by downloading MP3 internet music and then listening and commenting. He would make an MP3 for the station who would broadcast it that way.

They couldn’t make the station’s MP3 podcast from the broadcast show because the sound turned to garbage.

He posted wanting to know how to stop that.

Stop using MP3 for everything.

Koz