Wav to Mp3 without loss of quality

[freezwalm]

Hey guys, Freezwalm here.

I'm having a little problem, I make my tunes in Reason, but Reason is only able to export music in Wav-format.
No problem, I searched the web for a wav to mp3 converter and it worked fine. Untill today, when I converted my last new tune from wav to mp3, the mp3 file sounded really bad. The bass was clipping and crackling, while the wav file sounds perfect. So I messed around with the mp3 converting settings of the program, but after trying all possible combinations, I still haven't been able to create an mp3 file with the same audio quality as the original wav file.

Can anyone help me in any way? Which programs are best for converting audio? How do you guys do it? Anyway, I've uploaded it to soundcloud in wav format cuz soundcloud supports a lot of audio file types, but I really need to convert it to a perfect mp3 for being able to upload it here on newgrounds.

Thanks in advance,

-Freezwalm

(by the way, the audio converter I use is called Switch Sound File Converter)

[Buoy]

Would help if you told us what software you're using for the conversion. I always recommend Audacity because it works every time, it's common, open source, and you can do convenient edits like fading and cutting as well when preparing the final mp3.

Use a quality setting of 320kbps (constant) or 220-260kbps (variable) for the best quality. There is always some loss of quality but high quality mp3 files work fine for full songs.

[Buoy]

Also i didn't see the bottom line sorry. I used Switch as well but I don't recommend it because I've also had issues with it and Audacity is does everything Switch does but better

[freezwalm]

Thanx man! I 've allways had audacity and used it for sampling but never used it for converting audio files before, worked perfectly!
-Feezwalm

[Biggyzoom]

In my experience, Audacity should do the job just fine and it's free. If you wanted to go posh and hi-tech then you could invest in something like Toast which has all kinds of media conversion features as well as mastering and CD making applications and the like. Don't want to insult you at all, but do make sure that it does in fact convert it to a good quality. something like 192kbps stereo should do just fine, but you could possibly get it a bit lower and not hear the compression.

[MetalRenard]

Fun Fact: Most people (average listener, not specialist) can't tell the difference between a .wav and a 192kbs mp3 file.

[SourJovis]

It also helps not to export it as the best sound quality in the first place. When you export it, set it to no more than cd quality (44.1 KHz and 16-bit) and mark the box dither (masks the lower sample rate). This is way you've compressed the sound as much as you can before converting it, so the wav to mp3 conversion program will have to do less work (so to speak), and fewer things can go wrong.

And yes, as mentioned before, converting wav to mp3 will always decrease the quality.

Btw, I don't understand why files as wav and avi are called "lossless". I've tried to import and render those files a few times over, and the quality decreases quite rapidly. I'm told that's not supposed to happen, so maybe my software is bad (I used Sony Acid for the wav and Sony Vegas for the avi), but IâEUTMm not so sure. Maybe few people actually tried this.

[jpbear]

odds are your sub bass is way too fucking loud,

for some odd reason wav can handle it but mp3 cant

[seel]

At 11/28/12 07:20 PM, SourJovis wrote: Btw, I don't understand why files as wav and avi are called "lossless". I've tried to import and render those files a few times over, and the quality decreases quite rapidly. I'm told that's not supposed to happen, so maybe my software is bad (I used Sony Acid for the wav and Sony Vegas for the avi), but IâEUTMm not so sure. Maybe few people actually tried this.

Could be due to bit depth conversion and lack of dithering. Every time you convert down (32fp to 24bit, 24 to 16, 16 to 8 etc., you don't have to dither when converting up though) you'll introduce some artifacts unless you dither, convert up and down a lot and you'll be left with one hell of a mess.

@ OP, you should always limit your audio to -0.2dBfs or lower before converting to a lossy format such as mp3 to avoid intersample clipping.

[MetalRenard]

At 11/29/12 09:21 AM, seel wrote: @ OP, you should always limit your audio to -0.2dBfs or lower before converting to a lossy format such as mp3 to avoid intersample clipping.

^
Solid advice. +1

[Buoy]

At 11/29/12 11:18 AM, dem0lecule wrote:

still, get a flac if avalible if you're into djing and remixing stuff because using an mp3 and then saving it as an mp3 really shits on the quality

Since human ears cannot distinguish such range of quality; what's the point of using FLAC? Are we now being audio-file-whores? A simple OGG or even MP3 that has small size can serve good purposes for good-enough quality, just exactly what the ears desire to enjoy.

He never said anything like "you should only listen to FLAC", he said if there is a lossless version available you should use that when you're using it in a remix, a mixtape, or whatever. This is because the mp3 artifacts accumulate, so if you export a 192kbps remix of a 192kbps audio file it will not sound as good as a 192kbps file. If someone then uses your remix in a mixtape, and exports their mixtape at 192kbps, it will sound even worse. Hence why you should use the lossless version of your source material when it's available. Ya dig?

This is like saying a grand piano sounds better than an upright piano but fact is when played in blind spot, people don't know which is which.

Not a very good comparison. They sound different (for being pianos, of course), it doesn't matter if people can't name them in a blind test. One might fit better for a song than another. It wouldn't matter if it did anyway, because which sound you prefer is a subjective thing, not a quantifiable thing like audio fidelity.

Btw, have you ever done 'DJ' before?

Please discuss without being a condescending dick.

[jpbear]

At 11/29/12 11:18 AM, dem0lecule wrote:
This is like saying a grand piano sounds better than an upright piano but fact is when played in blind spot, people don't know which is which.

i lold

[SourJovis]

At 11/29/12 09:21 AM, seel wrote: Could be due to bit depth conversion and lack of dithering. Every time you convert down (32fp to 24bit, 24 to 16, 16 to 8 etc., you don't have to dither when converting up though) you'll introduce some artifacts unless you dither, convert up and down a lot and you'll be left with one hell of a mess.

@ OP, you should always limit your audio to -0.2dBfs or lower before converting to a lossy format such as mp3 to avoid intersample clipping.

I kept the quality constant, and there was no dithering to select or un-select. Maybe it's things like that that make the difference, but unfortunately I don't have any control over it, so no way to prevent quality loss.

Your setting the volume limit to lower than the max is a solid tip. I do this as well for videos (mp4), because I've had clipping problems there. Not for songs though, because I never had a problem with mp3 myself. Not maxing out the volume while you can seems a little counterproductive. Though -0.2dB isn't too much, and better be safe than sorry.

At 11/29/12 02:46 AM, jpbear wrote: odds are your sub bass is way too fucking loud,

for some odd reason wav can handle it but mp3 cant

Yes. good pint. I don't know if that's the cause for clipping, but I think it's wise to filter out any frequencies below or above the human threshold. No use keeping frequencies no one can hear anyway. Even though the frequencies can't be heard, they can still get in the way of other sounds or cause clipping. So get rid of them.

[KiddGimmickMusic]

http://media.io/ Just sayin'.

[SuperGoodSound]

At 11/29/12 09:21 AM, seel wrote: Could be due to bit depth conversion and lack of dithering. Every time you convert down (32fp to 24bit, 24 to 16, 16 to 8 etc., you don't have to dither when converting up though) you'll introduce some artifacts unless you dither, convert up and down a lot and you'll be left with one hell of a mess.

Dithering really does nothing. Do you know anyone that can detect quantization noise at -90dbfs? While listening at a normal reference volume? Not if the normal reference volume isn't going to destroy your hearing and the hearing of the people in the next room.

@ OP, you should always limit your audio to -0.2dBfs or lower before converting to a lossy format such as mp3 to avoid intersample clipping.

This has nothing to do with intersample clipping. It has to do with the algorithm the psychoacoustic masking process uses. On lower bitrates, the process causes aggressive variation in the output decibel level (up to around 1.5db overs for heavily limited material).

Also, ask someone who thinks they understand intersample clipping to describe the significance of why it causes audible signal degradation and it becomes really really funny to watch them try to answer. True story. Go ahead. Try it.

[ErikMcClure]

I should point out that with modern technology, it is usually trivial to utilize FLAC for maximum possible quality without much of a hassle. If you are a DJ, it seems obvious that you should be using lossless quality, because why the hell not? We have USB flash drives with 128 gigs of space on them. For streaming over the internet, sure you want ~192 kbps, but for anything else, it's just as easy to use lossless quality as it is to use 192 kbps MP3. You might as well use the highest quality possible - sometimes it makes things even easier if you don't have to convert to MP3.

[seel]

At 12/2/12 07:20 PM, joshhunsaker wrote: Dithering really does nothing. Do you know anyone that can detect quantization noise at -90dbfs? While listening at a normal reference volume? Not if the normal reference volume isn't going to destroy your hearing and the hearing of the people in the next room.

Good job not addressing the point I was making.

This has nothing to do with intersample clipping. It has to do with the algorithm the psychoacoustic masking process uses. On lower bitrates, the process causes aggressive variation in the output decibel level (up to around 1.5db overs for heavily limited material).

How does it have nothing to do with intersample clipping when that's what happens?

[Back-From-Purgatory]

At 12/4/12 02:58 AM, dem0lecule wrote:

At 12/4/12 02:55 AM, Lasse wrote:
Idiot speakin'

Nah, I'm just warming up.

No, you're not.

This stops here.