Converting MP3 to Wave

I have a few downloaded MP3s and they have different noise levels.

Some of the song are loud and some are less when played at the same volume.

How can I change their noise levels so that all the songs have the same loudness when played at the same volume.

I want to burn these MP3s to CD. It is quite irritating because some songs are loud and some are not. I have to change the volume level for each song.

http://www.geocities.com/mp3gain/

I think for MP3’s from different sources you just want to do track analysis and track gain. leave the setting at 89dB.

in case you got the new nero 6, there’s the feature “normalizing” in the nero smart-start menu, pretty useful, cuz don’t have da start another program…
:bow:

Originally posted by chrisi
in case you got the new nero 6, there’s the feature “normalizing” in the nero smart-start menu, pretty useful, cuz don’t have da start another program…
:bow:

Actually, this feature has been in Nero since version… euh I dunno… it already was there in Nero 4.x, perhaps even earlier.

Anyhow, just select all files in your Nero compilation, click right, choose normalize from the filters tab and off you go!

So whats better, to get the MP3Gain or use it in Nero? I am a Nero user for some time and don’t know if I need the extra software. I like to minimize as much as possible.

Nero is easier, but MP3Gain doesn’t affect quality. If you have any questions PM Pio2001, he probably knows a lot more about it than me.

I have downloaded MP3Gain and use it to normalize my MP3s to 89dB. The MP3s still sound a bit different but much better.

I know the Nero have some enhancement features but doesn’t know whether to use it or not. Which features should I use ? Noise,hiss reduction, stereo widening ?

hehe there’s a news post about mp3gain today!

Noise level is not the correct expression for this consideration. What you are “referring to” is the volume level of the Mp3’s that you downloaded. Say you have three songs: SongA.mp3, SongB.mp3, and SongC.mp3.

It is only natural that these have different volume levels SongA was originally on some CD-Music001 and some Audio-Engineer had mastered (adjusted volumes) of all the songs on that particular CD so that the CD would play with a nice even volume level. Perhaps, the CD-Music001 was mostly slow, romantic songs and SongA was a fast, rock-roll song. Then the Audio-Engineer most likely would have decreased the volume of SongA to blend it in with the romantic songs on the CD. The objective of the Audio-Engineer is to create a smooth sounding CD, not necessarily a true CD without clipping, so he may even allow clipping or make adjustments in specific octaves of certain songs

Same goes with SongB and SongC. They would have different volumes based on the CDs they were originally on

Nero has a feature to Normalize volumes when burning a regular music CD. But one must open Nero Express to activate it. Thereafter, one can switch Nero back to the old familar view. Burning time will just about double then, as Nero will first go through all the songs to record their volume levels and select a Normailization factor to use on the songs. Nero does not change the volume levels of the original Mp3’s on your hard drive.

Normalizing all those Mp3’s on your hard-drive is a foolish venture. You have two ways to Normalize the volume level of a song, by adjusting the Maximum (peak voltage) level of the song, or by adjusting the RMS level of the song. Most techniques basically maximize the “peak” volume level to 1.0 Volts Trouble is, the “loudness” heard by the Ear is mostly the RMS level of the song. Adjusting songs to all have the same RMS level gets very tricky, as if not very carefully done, one introduces clipping - - and clipping means adding unwanted harmonic garbage to your song. Neither Normalizing method will work unless the music is similar, as a song with a lot of base drum and very deep-baritone singer will sound louder than a smooth slow-dance song and often even louder than a regular rock-&-roll song

Goldwave is a beautiful software to use. One can Normalize volume levels by Maximum (peak) level or by RMS level. One can also edit soundwaves: trim some seconds off, delete clicks-&-pops, fade in-or-out. It can nicely handle blending of songs into one single song, say for a Jive session. And if you do karaoke, it is very handy to expand the waveform to help in synchronizing the lyrics to the music

One normally use noise level to refer to the background noise (noise floor) of the music. Most often, one refers to the signal-to-noise (S-N) ratio. This S-N is about 80-85 db for CDs, 30-35 dB for cassetes, 40-45 dB for industrial $500 one-inch tapes running at 30 IPS on $250K Honeywell tape recorders. Each 6-dB means doubling or halfing. On a CD, with a one-volt sinewave at 1000 hertz, the noise floor is about 30 microvolts.

Consider the below songs. The problem here is that these songs have their acoustical energy in different parts of the frequency spectrum and the Ear has different senstivity at different octaves. One would be hard-pressed to make all these songs have the same volume “as heard by the Ear”.

Die Flippers - Bye bye Belinda.mp3
Buddy Holly - Love me.mp3
Brooks & Dunn - Deny, Deny, Deny.mp3
Calimeros - 10 Jahr sind vergangen.mp3
Connie Francis - Diddley Medley: Fool’s Paradise, Love Made A fool Of You, Not Fade Away, Bo Diddley.mp3
Celine Dion - Thats The Way It Is.mp3

Normalizing these songs so all have a Maximum peak of 1.0 is worthless. These songs have to be Normalized by RMS-level, but to the lowest RMS-level less one introduce clipping. Clipping means adding harmonic garbage to the music. Even if one Normalizes to the (lowest) RMS-level, some songs are still going to sound louder than the others due to the acoustical energy being in different octaves, where the Ear may be more or less sensitive to sound

It would take an experienced Audio-Engineer some time to create a CD of these songs that sounded good and smooth. He would know what octaves are important on each song, how the Ear responds at various octaves, and use this knowledge to blend the songs on a CD. Most likely he would vary the order of the songs to make the inevitable volume level “changes” flow smoothly.

For the record, many stereo units have a Loudness adjustment to try to compensate for the way the Ear hears as one changes volume level, meaning quiet to loud or vice versa. This does nothing for adjusting volume-levels of different songs to blend the songs together. Also, for the record, I never adjust the volumes of the songs stored on my hard drive. This has no benefits and is a total waste of time. Do your volume adjustment when you form the music CD

Originally posted by John375
Each 6-dB means doubling or halfing.

3 db? //unsure

Volume (amplitude) goes by 6 dB steps

+6 dB means twice as loud …
+12 dB mean four times as loud (2x2) …

  • 18 dB means eight times louder (2x2x2)

-6 dB means half as loud …
-12 dB means one-fourth as loud

The standard argument is that smaller dB steps (say 2, 3, 4, or 5 dB changes ) are rarely noticeable to the average person. In fact it is often argued that one can not even measure the Amplitudes to better than 6 dB accuracy, on a day to day basis. So one normally considers only 6 db steps for volume. (Those with “golden ears” can do better than the 6 dB steps)

Power (being proportional to the square of the amplitude) goes by 3 dB steps

+3 dB means a doubling of power
+6 dB means the power increased by a factor of four

-3 dB means the power went down by half

The ear hears Amplitude, that is volume, or more specifically the pressure changes in the air. The Ear does not really hear power, as power means a time integration, that is work-done divided by the time it took to do it

Relative to - - I know the Nero have some enhancement features but doesn’t know whether to use it or not. Which features should I use? Noise,hiss reduction, stereo widening ?


Avoid monkeying with these features, as you can easily do more damage than good.

If you have Noise (usually meaning Hiss, Clicks, and-or Pops) than the song came off some tape or some vinyl LP record. Such a song is already of limited bandwidth, thus low quality, as it probably does not have anything above 8,000 to 10,000 Hertz. It can take hours (say, 4-8) to properly clean just one song, and when you are done, the song is still a dud as it lacks full frequency bandwith. (When I have to do such cleaning, though I really hate the wasted time & tedious work, I prefer to use Goldwave as it gives me better control. But one can use Steinberg Clean, the Nero waveform features, LP-ripper, or several other cleaning softwares.)

For the record: LP’s are good to maybe 8,000 Hertz and awful lucky to ever reach 10,000 Hertz, but have major low frequency limitations. New “good quality” tape cassettes are only good to 10,000 to 12,000 hertz, though one may get a little more using the Dolby compression schemes, while old cassttes are lucky to even reach 8,000 Hertz . CDs claim a 20,000 Hertz bandwith, but one has to take this with a grain of salt.

Years ago, some Guy on the Internet would clean LP’s and convert them to CDs for those who wanted to save their LPs to CDs. He charged $30 per LP. (LP’s only have about five songs on them per side). He also had some $3000.00 worth of equipment that he used to do the job

Rather than clean a song, it is far easier to go and search the Internet for a better, clean copy than correct what you have. Say use the p2p software WinMX. If you find a clean copy, most likely it came from a CD re-mastered from the original recording and has far better bandwidth than your bad copy.

The major problem with cleaning a song is that you are going to lose some music sound, and all too often the resulting music sounds “flat and lacking” after it has been “cleaned”. This is going to happen no matter what you do. The trick is to make the loss not “readily” detectable by minisculing appling filters for short durations here and there, and this involves a tedious trial-&-error procedure that takes hours. One tries a filter, then listens to the cleaned music, and re-do’s with filter refinements. One has to repeat this at every bad spot in the song. Plan on a half-day to a day per song to do this properly

There are two ways to clean a song, work in the Time Domain or work in the Frequency Domain, and often one does both. (Goldwave does both). Each domain has advantages and disadvantages. Removing tape Hiss can be done simply by applying the standard tape Hiss filter, starting with a mild beginning application. (Goldwave tells you what to use and how to start). But this Hiss filter makes the music “sound dead” in quiet (low volume areas), so one only uses it discretely throughout the song. A much better way to remove tape Hiss is to use a quiet area of the song (say song beginning or end) to generate a tape Hiss noise spectrum in the frequency domain and then subtract this spectrum from the running spectrum of the music. Clicks and Pops are usually miili-second or less duration. One first tries a standard Click-Pop filter, very mildly applied, to remove these annoyances, and usually one does this in small sections of the music less the the resulting music “sound flat”. If that does not work, add more strength to the filter and re-try. All too often, one has to go into the spectrum (frequency domain) and try to cut out the noise spikes. Sometimes a small area of the song is so bad that it just sounds better with that small section cut-out (removed) (Goldwave does this easily)

Stereo widening - - forget it. One never makes CDs using “altered” songs. The songs on the CDs are suppose to be “flat”, meaning no enhancements (other than noise control). One does not add “base” gain to the songs before placing them on CD. One does not add “treble” gain to the song before placing a song on the CD. One does not add “stereo widening” to the song before placing the song on the CD. These enhancements can be done when you play-back the CD. Nearly all home stereo euipment allows you to make these play-back effects. The reason for not embedding these enhancements into the songs placed on the CD is that your taste may change over time or others may listen to the CD and not like your choice of enhancements

BTW, I don’t have Goldwave stock

Relative to - - - Converting MP3 to Wave

I never did understand what you wanted here. Use CDex to do this. It is a very small beautiful piece of software, with all the CDDB capability one would want. CDex can match or beat the quality of any CD-Ripper or Wav-Mp3 converter on the market. CDex is free.

An advantage of CDex is that it is easy to change the Mp3 Encoder. CDex comes with that encoder commonly considered the best as of today. Nowadays, this is the Lame Encoder, a file named “lame_enc.dll” within the CDex folder. To update, just change this DLL. Or add other MP3 encoders as you want to the CDex folder, as change the option in the CDex dialog box

Originally posted by John375
Power (being proportional to the square of the amplitude) goes by 3 dB steps

thats why the unsure :wink: - couldnt remember which was which, and i was too tired to pop open a textbook to find out

I prefer Steinberg Clean to reduce hiss and noise and to adjust volume. It kicks @ss.

Relative to - - I prefer Steinberg Clean to reduce hiss and noise and to adjust volume

Steinberg is Okay, I have used it. But it is for amateurs. It has all the knobs & whistles built-in and set for you beforehand. One just presses the de-click button, de-hiss button, etc.

Trouble is this broad-scope application of the de-noise and usually it will change the sound. Few persons knowledgeable in de-noising signal use this approach

Goldwave can do exactly what Steinberg does with the same simple tools. (The de-noise filters are International standards) But I like the waveform expansion available in Goldwave and its ability to minisculy apply filters

I never tried Goldwave but I know steinberg is used on a professional level, and for that reason I started using it. It has done wonders on old tapes I transformed and made digital. I just stick to what works and it works amazingly well. I dunno if you have ever tried the audio mastering tools in WinOnCD, but I also like them a lot an duse them too. If someone out there has the time, I guess for advanced mastering AcidPro is the way to go, but this thread started out simple to convert mp3towav, so discussion of goldwave, steinberg clean, and all those are useless to this topis in regards to what was originally posed as a qauestion.

Ins hort, for a simple program, todo the job, BadReligionPR gav ethe best link I think thus far.

I agree. It is time to end this tread, - - - or move all my replies to a separate thread

I was an aircraft structural dynamics engineer and spend decades studing time signals for evidence of what caused structural failures, be they from shock loads, vibration loads, or acoustics loads. I have used every technique immaginable to extract intelligence out garbage data signals, and can compute & check all calculations made by the analyzer.

Goldwave, though simple, is closest to the analyzers I previously used and I can easily make it do what I need done. The other softwares mentioned have little interest to me. If I need more, then I rather just use a full-blown Spectrum Analyzer software where I can readily switch from the time domain to the frequency domain, use Fourier or Wavelet Transforms as needed

Time to close this - - bye

I have a program called Nero Wave Editor that came with Nero 6. One can load and edit a number of different formats of sound-file with it but it displays all formats in wave format on screen. Because there is a visual representation of the wave file onscreen, it is easy to see whether the level is too low and increase the volume accordingly.using the tools. If increasing the volume universally would cayuse clipping, it is possible to apply a curve which ‘normalizes’ the volume (i.e. makes it so that the range is narrower.) You can then increase the volume overall without clipping. If clipping does occur, you can see it on the image because it will extend beyond the window. You therefore have the choice of saving it or not, so no harm is done. I have been using it successfully for years, but it certainly doesn’t take half a day to do one file; half an hour for a tricky file at most. There are several other tools for fine-tuning. I have even successfully removed noise from files made from recordings from the 1920s. This required a little longer but was quite doable. I have zero training as a sound engineer.