Audacity Forum

Hi,

I have converted a song from a UMX mod format into MP3. It was a bit muffled/muddy sounding, so I cleaned it up a bit with Audacity's Equalizer. However, there are a few spoken words that include "T" "Ch" and "S" sounds, which now peak pretty high in the upper frequencies, and tend to hurt the ears during playback. So, my thought was to select only those peaks (on the waveform) where those sounds are, and equalize them again, taking down the 4-10K frequencies.

Problem is, doing this results in a big gap at the beginning of the selection: the volume drops off and then comes back up. Isn't Audacity supposed to be able to apply Eq to a SELECTION of a song without any kind of problem?

Secondary question...

Barring a solution for the Eq-dropoff problem above... perhaps someone has a perfect solution for cleaning up a somewhat muffled/muddied (and bassy) song? The original MP3 I created from the mod file lacks dynamic range and clarity when played alongside songs in my playlist that come straight from a CD. Is there a sequence of tools and/or settings that are recommended for restoring clarity to an MP3?

Thanks for any expertise!

I'm going with no. Once you MP3 compress a song to death, there's literally nothing left to boost back up to normal. MPEG1, Layer III (MP3) is a lossy compressor. It does it's job by very selectively and gently throwing away quality. The designers realized that there might be times when you would be willing to sacrifice significant quality to get a small file size, so they made the tool open-ended. They will let you produce a scary-small music file, but it's not going to sound pretty and you won't be able to bring it back. This is how a lot of home answering machines work.

The harsh "S" sounds might be helped by a "De-Esser" tool. There may be one buried in here, somewhere...

http://audacityteam.org/download/nyquistplugins

Koz

kozikowski wrote:I'm going with no.

Koz, which question were you answering -- No, there's no way to equalize a section of song without the annoying volume glitch? Or, no, there's no decent combination of tools to use to maximize what sound quality remains in the MP3?

I will definitely look for the de-esser. I might also look back at my conversion utility and see if I can get a WAV file from it instead of just MP3. That would be better?

I think Koz was answering "No" to the question "Is there a sequence of tools and/or settings that are recommended for restoring clarity to an MP3?"

On the whole I would agree with him.
Converting an uncompressed sound to MP3 essentially makes the file size smaller by throwing away some of the audio information. This is called "lossy compression". You can not "rebuild" the quality, because you do not have the bits that have been thrown away.

Doofnac wrote:I might also look back at my conversion utility and see if I can get a WAV file from it instead of just MP3. That would be better?

That is likely to be better, or at least it should be as good as the original, however if the sound quality of the original MOD file is not very good, then there is little you can do about that.

I've not worked with MOD files for years, but if memory serves me correctly, check out "BassMod" and "Modplug".

I'm using MPTrack to convert from the UMX mod... and YES I can export to WAV. And I gotta say that sounds MUCH better. Maybe not quite like the original probably sounds, but at least the upper frequencies aren't muddied, despite the bass. Of course, the file is 12 times as big...

I'm a neophyte at this stuff; I didn't realize just how compressed an MP3 is. Anyway, I think with a bit of of a high-pass filter, this WAV's good enough to go. Thanks guys!

...

Now... about the first part of my original post: Why is there a volume glitch when I try to equalize just a section of a song?

Doofnac wrote:Now... about the first part of my original post: Why is there a volume glitch when I try to equalize just a section of a song?

I'm not really clear what the problem is. Could you try and describe it in a bit more detail - possibly giving a step-by-step method for recreating the problem, and maybe some screen shots?

We are reminded that audio production in a highly compressed or restricted format is crawling with difficulties. In the case of MP3, each time you export and import, the compressor subtracts some quality all over again. If you do enough production management, editing, and cutting, there will be nothing left but audio garbage at the end. The damage is cumulative and in some cases multiplies.

This is why people capture live performances at the very highest digital uncompressed format they can lay their hands on and then keep that same sparkling, broad, high quality format through the production and editing cycle. Uncompressed audio formats don't degenerate at each step. And no, this isn't going to fit on a 10G laptop hard drive. Uncompressed is big. Very big.

In general, the compressors (MP3, AAC, M4A, etc.) are delivery formats. That's what you do right before you push the performance off to your iPod or post to your web site. If you're delivering to a client, you do it in the original huge uncompressed format and let them figure out what to do with it later.

This is why you're getting enormous swings in your show quality. Because you're editing in compressed formats, each production step comes with a knob that adjusts G00D---Medium---BAD. If you're not paying attention to that, you can cause a lot of damage.

Koz