How does a newbie reduce bass?

Special Interest Groups Audio Processing
1

spectrum.txt (5.91 KB)
spectrum.txt (5.91 KB)
Hello Everyone,

While I’ve been a audiophile for about 35 yrs, this is the first time I’ll be trying to edit some recorded music to make it sound better.

I went to a great concert (Jason Bonham’s Led Zeppelin Experience). At the exit, a company named Ovation had set up some kiosk’s where you could download the concert to your MP3 player on the spot.

After listening to it, I found the bass frequencies to be extrememley overly loud. Probably on the order of 15-20dB below 150Hz, and by up to 30dB below 80Hz.

I’d like use Audacity (for the first time) to make it listenable.

Some questions:

  1. Can you expand the Spectrum pref to show freq below the indicated 86Hz cutoff?
  2. Can you have it display the music in a Spectrum analyser form on a continuous basis like many home audio eqiupment can do?
  3. Should I use FFT or the Equalizer to reduce the bass frequencies?
  4. How to I set it up to apply the curve I generate to all the songs automatically?

I’ve uploaded a clip so you can see what I’m referring to.

Thanks again for helping this newbie!

David

2

If the venue was getting the show like the audience is getting it, then it’s no wonder the sound is off-balance.

Many bands let people record them from the audience because they know the chances of getting a clean show are almost zero.

First thing you need to know is that if the show started out life in MP3, it needs to not be edited and re-exported (the way Audacity does it) at all; or edited and upgraded to uncompressed WAV or AIFF. You can’t ever go back to the same size MP3 file after editing. MP3 music damage is cumulative and permanent.

You can grab a corner of the frequency display and pull. It will get as big and detailed as your screen will allow.

It doesn’t do a lot of good to expand very much below about 100Hz because you can’t tell the different between 100Hz and 95Hz, and filters don’t work down that far anyway. Nobody in Hollywood would dream of doing a live shoot including anything below 100Hz. It’s a magic place.

Load a sample show segment so you don’t have to wait for the whole thing each pass at the equalizer. Effect > Equalizer. The bar in the middle is a rubber band. You can click and pull up and down at each frequency until you get what you need.

http://kozco.com/tech/audacity/voiceEq.jpg

Once you get one that works, save the curve under an appropriate name and you can use it as many times as you wish.

If equalizer is listed in the Batch (Chains in Audacity) then you win. If it’s not. You lose. Audacity Chains components are hard-wired and unique. If it’s not there, that’s the end until one of the developers writes a new one.

Koz

3

The equalizer tool will also get bigger and more detailed as you pull it bigger and bigger.

There are ways to edit MP3 without recompressing and adding damage, but those won’t do equalization.

You need to be in Audacity 1.3 for all this. Audacity 1.2 has problems.

Koz

4

Thanks, I’ll try 1.3.

It appears that the mp3 didn’t upload so I’ll try it again, then you can see how boosted the extreme low bass is and why it drives me crazy.

5

Ok, it appears on the spectrum graph I can view down to 20Hz or below by changing it to view it as log freq (which is what we audiophiles are used to), and changing the size to 2048.

Doing this it appears the bass below 100 Hz or so is 20-30dB too strong.

I can use the Equalization Effect or the Apple IUGraphicEQ to get the desired response. According to the Spectrum ananlyser, there is signal down to 5Hz!

Is there any way to get a preview of what the Spectrum will look like, (not the usual waveform)? It doesn’t appear to change.

Lastly, I think you said that you can’t do any of these changes while still in MP3 format, correct? So, is there a best format to let Audacity convert it to to do the processing?

Thanks again.

6

Is there any way to get a preview of what the Spectrum will look like, (not the usual waveform)? It doesn’t appear to change.

I don’t know… but, once you’ve approximately identified the problem frequencies it’s best to listen with your ears rather than with your eyes! :wink: Load-up a known-good “reference recording”, but don’t try to “match” the reference. Just use the reference as a guide to help determine if there’s too much bass or too little bass. Your ears are the best tools, but it’s a good idea to “calibrate your ears” by listening to a reference as you are working. (Pros do the same thing*****.)

Lastly, I think you said that you can’t do any of these changes while still in MP3 format, correct? So, is there a best format to let Audacity convert it to to do the processing?

First, MP3 is lossy compression (data is thrown-away during compression) but it can be quite good at high bitrates. With lossy compression, high-bitrates = bigger files = less loss = better quality. The guys at HydrogenAudio.org have done lots of blind listening tests, and it turns-out that most music can be encoded transparently (it can sound identical to the uncompressed original).

The “damage” is done during encoding, not during decoding. Audacity (and all “regular” audio editors) have to decompress the MP3 to PCM (like WAV) before processing. And of course, it also has to be decompressed when you play it. Like most audio editors, Audacity will decode to uncompressed floating-point PCM. The additional damage comes when/if you re-compress to MP3 or other lossy format.

So the question is, what format do you want or need? If you make an audio CD, there will be no further damage (assuming you save-as WAV after adjusting the EQ.) If you were to make a CD (or other lossless format) It would sound identical to the original MP3 (if you didn’t use any EQ or other intentional processing). If you can use FLAC, ALAC, or WAV, these are also lossless.

If you need MP3 or AAC for your portable player, you are going to go through a 2nd lossy compression step and the only thing you can do is use a high bitrate (high quality setting). You can always create a CD or FLAC archive in additon to the MP3.

*** P.S.**
Another trick pros use is to listen on everything they can get their hands on. This is even more important for us amateurs who don’t have studio monitors… Listen on your living-room system, make a CD and listen in your car, your iPod/earbuds, your headphones… It doesn’t have to sound perfect on all these systems, but it shouldn’t sound terrible and should compare well with your reference.

7

One of the shortcomings of Audacity: no “real-time” preview of equalization (or other effects), it’s on the wish list …

Proposed Feature

Real Time Adjustment means that when adjusting the parameters of an effect, you can do so AS you are listening - imagine playing with a graphic equalizer as you are listening to a sound - today you have to make a change, preview it, and then make another change.

http://wiki.audacityteam.org/wiki/RealTimeAdjustment

“real-time” adjustment is possible with other DAWs.

8

But there is a “Preview” button.
The default preview duration is only 3 seconds, which I find to be a bit short so I have it increased to 5 seconds. This is set in “Edit menu > Preferences > Playback”

9

Yes, I saw the “Preview” button, but I couldn’t tell what it did! When you say the duration is short, do you mean the preview lets me hear the difference? Or see it on the main waveform display? Or see it on the Spectrum display I created?

10

By default the Preview button will play the first 3 seconds of the selected audio with the effect applied to it. If the first three seconds of your selection is silence then you won’t hear anything. It takes a few moments for Audacity to process the audio so it is best not to set the preview length too long. I find that 5 seconds of preview is a good setting.

Before applying the Equalization effect to a song, I will select (click and drag with the mouse to highlight) about 10 to 15 seconds from a “typical” part of the song. Then I will open the Equalization effect, have a guess at the settings and click the Preview button. When I think I’ve got the settings about right I click the OK button and the effect is applied to the selected region. I can then listen to that region in context and decide if I like it. If I do, then I will press Ctrl+Z to Undo the effect, then select the entire track (double click on the track or click on a space in the track control box) and press Ctrl+R to reapply the effect to the entire track.

I usually use the “Draw Curves” view in the Equalization effect as I find it less fiddly than all of those sliders.
Something like this should produce a noticeable reduction in the amount of bass:
bass-cut.png

That’s pretty extreme, but you could counter that with settings something like this:
bass-cut2.png
Note that the vertical sliders on the left allow you to adjust the vertical scale of the graph.
After reducing the bass you may want to run the Amplify effect to bring the overall level back up.