No Free Pass For Podcasts

[Donovan]

Donovan and everyone else reading this thread, pay very close attention to Paul's advice. While I realize his analysis is over a lot of heads, it's well worth spending the time to understand what he is saying. Donovan, you are free to follow his advice about maximum levels. Reducing the peak level on your files by 1.5dB will not alter the tonality, but it may save you some clipping. In fact, it would be helpful if Paul would analyze one of your released MP3 files.

Paul, what software are you using to do the analysis? It's useful to separate the men from the boys. I will confess that I am guilty of normalizing and hard limiting wave files to levels higher than -1.5 dB. That procedure changes today. The additional 1.5 dB does not really add to the loudness of the audio, and it does contribute to the clipping and other undesirable results from MP3 conversions.

Guys, this stuff is good. Embrace it. Once you do, you will be leagues above the "on fire" and "answerman" guys when it comes to professionalism in audio quality.

If you just can't get the hang of what Paul is saying, just normalize or hard limit your audio files to -1.5 dB instead of 0dB. It will LOOK different in your software, but you will get used to it.

I made the adjustment last night for episode 26 of my show. Though I can't hear a difference, I do feel better watching the levels never go above -1.5db, which is what I wanted in the first place.

The other thing I'm going to test is my process. I apply the multi-band compressor first and then apply the dynamics processor that has my own settings. I'm going to switch it up and see how it sounds if I apply the dynamics processor first and then the multi-band compressor. I think I've done that in the past, but I need to try it and hear it again to be sure. As I said, I'm always willing to learn more!

[Donovan]

And if he wants to analyze one of my podcasts, might I suggest last night's episode 26 which is at http://www.donovanadkisson.com/tds26

I'd love to see what he finds.

[ptfigg]

And if he wants to analyze one of my podcasts, might I suggest last night's episode 26 which is at http://www.donovanadkisson.com/tds26

I'd love to see what he finds.

I already did , although I think it was the previous episode. The Peak Amplitude was 0 dBFS and there were Intersample Peaks. With respect if you submitted that to me it would not comply.

Your Program Loudness was close to the recommended spec, so that's good. OTOH that's entirely different issue that I'm sure we will get into.

I'll check the most recent episode ...

-paul.

[Dan Ortego]

Well, if this doesn't limit the playing field of thread participants I don't know what will. I was almost with you until it drifted into command line tools, syntax, and compiled binaries, at which point it made Adobe Audition seem like child’s play. Yep, its over my head but I admire those willing to work at this level.

[ptfigg]

And if he wants to analyze one of my podcasts, might I suggest last night's episode 26 which is at http://www.donovanadkisson.com/tds26

I'd love to see what he finds.

I measured the file:

Peak Amplitude: +0.27 dBFS (Clipped)

True Peak: +1.02 dBTP

Total Intersample Peaks: 144

Program Loudness: -17.4 LUFS (Mono)

* * *

I'm not sure what tools you have access to. Fill me in and I'll make some suggestions. As far as the Program Loudness, it's slightly hotter than what's slowly becoming the recommended standard. But that's ok. As long as you produce consistent results on a show to show basis, we'll save that discussion. There are variables, mainly the differences in how to handle processing Mono files vs. Stereo. And there are various aspects of "Loudness Normalization" that we can discuss. Note that it's not the same as Peak Normalization or RMS Normalization.

Like I said we can get into this after we solve the clipped audio issue ...

Keep me posted.

In any case I know this stuff may be a bit complex. However - if you (not you personally Donovan ) want to learn how to get it done right, there's no way around the complexities. Otherwise the space will continue to be occupied by people giving bad advice and "students" not realizing that they are being misinformed.

-paul.

[Donovan]

I measured the file:

Peak Amplitude: +0.27 dBFS (Clipped)

True Peak: +1.02 dBTP

Total Intersample Peaks: 144

Program Loudness: -17.4 LUFS (Mono)

* * *

I'm not sure what tools you have access to. Fill me in and I'll make some suggestions. As far as the Program Loudness, it's slightly hotter than what's slowly becoming the recommended standard. But that's ok. As long as you produce consistent results on a show to show basis, we'll save that discussion. There are variables, mainly the differences in how to handle processing Mono files vs. Stereo. And there are various aspects of "Loudness Normalization" that we can discuss. Note that it's not the same as Peak Normalization or RMS Normalization.

Like I said we can get into this after we solve the clipped audio issue ...

Keep me posted.

In any case I know this stuff may be a bit complex. However - if you (not you personally Donovan ) want to learn how to get it done right, there's no way around the complexities. Otherwise the space will continue to be occupied by people giving bad advice and "students" not realizing that they are being misinformed.

-paul.

Thanks, Paul! I have Adobe Audition, ffmeg, etc. If I don't have a program I need, I can possibly get my hands on it.

Can you share the command line parameters you're using to measure these stats?

Thanks.

[Donovan]

I scanned my processed .wav file from last night's episode in Audition. This is what it told me -


https://s3.amazonaws.com/amn_shows/m...Statistics.PNG

[ptfigg]

I scanned my processed .wav file from last night's episode in Audition. This is what it told me -


https://s3.amazonaws.com/amn_shows/m...Statistics.PNG

It looks like you're running an older version of Audition that lacks True Peak analysis and Loudness stats ...

Here's what I get in Audition CC:

analysis.jpg

-paul.

[Donovan]

It looks like you're running an older version of Audition that lacks True Peak analysis and Loudness stats ...

Here's what I get in Audition CC:

analysis.jpg

-paul.

Yeah, I'm still using 5.5. Adobe gave me a year of everything when an article was written on me and my process using their software, but after the year was up, I couldn't justify the $50 per month. I realize that Audition is just $19.95 per month, but I haven't felt the need to upgrade. Maybe it's time. I like statistics.

[ptfigg]

Thanks, Paul! I have Adobe Audition, ffmeg, etc. If I don't have a program I need, I can possibly get my hands on it.

Can you share the command line parameters you're using to measure these stats?

Thanks.

You're very welcome and I'm happy to help.

For the comprehensive SoX analysis:

At the prompt:

sox --show-progress Your_Input_File.wav -n stats

For Loudness Specification Analysis using FFmpeg:

At the prompt:

ffmpeg -nostats -i Your_Input_File.wav -filter_complex ebur128=peak=true -f null –

You should be good to go in Audition by accessing Amplitude Statistics and running "Scan."

If you are interested in learning how to use a Loudness Meter, we can open a discussion. Keep in mind if you are running an up to date version of Audition, a scaled down version of TC Electronic Loudness Meter is bundled:

Effects/Special/Loudness Radar Meter

There are many other meters. I own most of them and I'm happy to offer anyone assistance.

"afclip" is part of the "Mastered for iTunes" toolkit, and only available for Mac.

Again, keep us posted ...

-paul.