PG Music Home Forums General Discussion Beginners Forum Why does mp3 version sound dull?

Hello all -

First post for this newbie. I liken working with computers and software to learning to drive an 18-wheel rig: it's cumbersome, it's slow, I'm lumbering, there's a lot of gears to navigate, a lot of mistakes, etc. etc.

So I'm muddling thru Windows BIAB 2016 w/ 2017 upgraded RTs.

I created a song on my laptop (Windows 10), converted it to .mp3 and e-mailed it to my PC (Windows 7) to play on iTunes, but it sounds faint and dull on the PC.

The mp3 version sounds great on the laptop, tho the file is identified as .WAV and not mp3 as I had specified at file conversion. What gives?

Thanks.

MP3s compress the data. In other words you will lose data when you use MP3. The more compression you use the more data you lose. This is called lossy compression. If you chose 128 kbps then you will lose a lot of data. Lower than 128 kbps MP3s are really terrible. To get wav type quality you should use 320 kbps MP3s, although 192 kbps and up also sound OK with some genres of music.

Wavs do not compress data so that is why they sound so good. There are other music file compression programs that do not lose data. Programs like Flac and Ape are two such programs. They are called lossless programs. If you have a player that can use either of these that would be the optimum way to go IMO. Otherwise use MP3s that are 192 kbps or higher.

Note that sometimes when sending others 128 kbps MP3s I have had to have a song set for wav and another version of the song mixed for MP3. However if you use the 192 kbps or higher you should not have to do this. YMMV

PS - I use the free Audacity program for converting wavs to MP3s.

{edit} - I forgot to say welcome to the most friendliest site on the web!

If you render your MP3s at 256 or 320, I find almost imperceptible differences in the sound. When I rip my music, I generally rip it twice. Once at 128 for my iPod Classic that gets used in the car (where there is lots of noise anyway, and that allows for a lot of songs) and again at 320 for playback on my home audio system (a much better listening environment).

And just to clarify, Mario, is talking about lossy data compression (ripping to different audio formats, such as MP3 or WMA), not audio compression (such as running a VSTi compressor). Two completely different things.

Thanks for the quick response, guys, but back up the truck a second...

Where am I specifying a kbps?

I hit the WAV icon, giving me "Export Song as Audio," then I click "Export as MP3," then Save (to my desktop). I attach it to an email & send to the PC to open in iTunes: dullness ensues.

If you haven't installed updated codecs on your computer, you may very well be getting 56kbps encoding by default.

I generally don't do it directly from BIAB; I usually just export to WAV and from there convert to MP3 using a dedicated program, where I can choose my codec (usually LAME) and bit rate settings.

Thanx John for clarifying. I wasn't even thinking about audio compression when I wrote that message! My bad! I should have specified the difference but I'm glad that you did.

I use Audacity for all of my conversions. You can download
Audacity here:

http://www.audacityteam.org/download/windows/

Then scroll down and download LAME MP3 encoder from the same page.

If you look at Audacity you will see that you can do a lot more than conversions with it.

When converting a wav to a MP3 you can choose from many data compressions. Click on File/Export Audio and scroll down to MP3 files in Save as type. Then click on Options and pick your kbps in the scroll down arrow in the Quality window: see attached image.

FWIW, as far as I know, mp3 format has a 16kHz top end cutoff regardless of the bitrate, so you lose something in the top end no matter what. Regardless of codec or software used.

I had not heard that but it makes sense. A search came up with an article that states MP3 top end frequency cut off is related to the sampling conversion. The chart below is copied from ++ THIS ++ link.

bit rate > cut-off frequency > compression
1411kbps > 20kHz > 1:1
320kbps > 19.5kHz > 1:4.4
192kbps > 18kHz > 1:7.3
160kbps > 17kHz > 1:8.8
128kbps > 16kHz > 1:11
96kbps > 15kHz > 1:14.7
64kbps > 11kHz > 1:22
32kbps > 5kHz > 1:44

Check out Nyquist rate for a further explanation.

Yeah Matt, I'm very familiar with Nyquist as it applies to a wave file but never thought that it applied to lossy MP3 formats. Don't know if I didn't think it through or just had blinders on.

Of course sometimes even I wonder about how much I don't think things through. For instance, do you also know that the only data recorded in a sample is waveform gain? I was dumbfounded when I found that out.

Unless I'm mistaken, the Nyquist rate will tell you the highest frequency that can be produced by a specific sample rate, so that would explain why the highs disappear in successively lower MP3 encoding rates.

BTW, the trick to getting the free LAME mp3 codec into Audacity's access is you have to go to Edit | Preference and then Library. From there it will install it sort of automagically for you. It will take you to a link at the Audacity website. In that link you will choose LAME or something like that.
Then when you want to convert a .wav to mp3.
In Audacity...
File
Open | navigate to your .wav file. It will import it for you.
File
Export Audio
At the bottom of the Export dialog you can change the compression rate.

So... what every said above is correct... every time you covert a WAVE file to MP3 you are losing 80% or more of the original data depending on the bit rate of the MP3. I always convert to 320kbs MP3 files unless there is a really good reason not to use 320kbs.

But... lets back up a bit. When you finished the song and exported the wave, did you do anything else to that wave at that point or did you proceed directly to the MP3 converter?

The reason I ask is because many folks at that point will process the wave file with what is called "mastering software" to add some sparkle and thump to the music. I prefer to call it polishing or sweetening the song because "mastering" is something totally different but I am OK calling it "small "m" mastering" as well. It's simply the process of adding compression, reverb, and EQ tweeks to the finished song to make it sound a bit better and more cohesive.

Essentially, when you listen to the Wave or the MP3, there should really be very little sonic difference between the 2 files. Most people would not be able to tell you which one was which. Especially at the higher MP3 bit rates.

You also should understand that the different music players software will affect the sound quality more than the Wave to MP3 conversion. I would check that MP3 on several different players to see if that is the problem. Barring that, go back to the original Wave and do some of the polishing and sweetening I mentioned.

One other thing to check on. Be sure that the exported WAVE is normalized to just under 100%. I like to shoot for 96% on the highest peak. Run some compression first to bring the entire wave into a relatively consistent looking form. You want the finished wave to look like this.



In this photo.... the compression rate is consistent but not as high as it could be because there's quite a few spikes and "airspace" above the main body of the energy. The song is on my website and sounds good. By cranking up the compression a bit more, the song would have gained more "apparent volume" but lost some of it's dynamic range... the difference between the loudest and softest parts of the music. Once you get the compression right, and it might take a few tries if you're not worked with this before, you can then use the normalize function. Normalize brings the highest peak to a designated level that you set relative to 0dB. You don't want to go above 0dB.

As a footnote to the compression thing.... in the picture above, if I had cranked the compression up to max, that section toward the end that is obviously lower in volume would have been indistinguishable from the rest. To much compression is not a good thing.

All of that is important because you mentioned a "faint and dull sounding MP3"..... again, that should not happen if your levels are all correct before you convert to MP3, and you have done some proper sweetening.

I hope this helps you find the reasons for the weak and dull sound you ended up with in the MP3.

Herb, do you use WavePad for your 'mastering software'?

Even simpler; use PGRTA plugin (or any other frequency analyzer) and look the frequency range of the mp3. See how much it differs from the same wav file.
The better the analyzer the more you'll notice.

I'm sure quite a bit of real-life results depend on the actual codec/rate being used for the conversion.

I think it always will. I'm no expert, but my understanding of history is that Apple wanted to load as many songs on to the 32Mb memory of the iPod as it could. Bear in mind, this was the dark ages of memory. Mine has 4Gb now.
OK, so this guy comes along, takes an MP2 (stereo) and adds a third track.
That third track went along and eliminated everything it could. If there was a pause or a silence, the third track would simply tell the audio file when to cut and when to resume. We're talking high tech nano second stuff, here. Basically, a midi on steroids.
The MP3 are made to be played on portable players. What you are hearing is compression quality. Your machine is probably just fine.

Yes.

It's a good program and I use it to listen to and work on my exported waves.

In Wavepad I trim the start and the end, and I also run the normalize function on the wave.

After I have done those chores, I convert it to mp3 in Wavepad.

Wavepad seems to have a player that sounds extremely close to my DAW, which is always a good thing. In other words, it doesn't sound to me like it's adding any color to the sound. AND.... I can't tell any sonic difference in Wavepad between the final Wave and the mp3 that came from it.

I do the same as Herb, but I use Diamond Cut DC8 to do the same thing. I use it to cut off the front and tail. I use it to normalize the wave, as well as apply any mastering effects. DC8 has a lot of filters and effects, mainly designed for cleaning up records and tapes that have been digitized, but works just as well on audio files I create. And DC8 lets me rip to MP3 (or use a host of other CODECS). We're going to start beta testing DC10 pretty soon, so I'm looking forward to that (I help beta test). They also have a very friendly forum (like here) where you get to interact directly with Craig and Rick (the founders of Diamond Cut).

My method is similar to John's. I use Adobe Audition 1.5, which is the old Cool Edit Pro converted to Adobe and patched once before Adobe mucked it up. [I also have Adobe Audition 3 so I know].