improving quality of a taken sample

[afasio]

Hey all, ive got a problem.

Me and my friend usually sample fragments of beats, like leads from old anime series or things like that. but when i get the song from youtube, the quality is really horrible, probably because it has been compressed x number of times. what can i do to make the sample sound better / less horrible? i could really use some tips!

looking forward for your answers!

[CPhoenix]

You can't really "improve the quality" as easily as you want to. Once you get anything mixed down, obviously you can't revert to it's prior state.

All you can really do is use your tools to try to correct problems. You're essentially working w/ what you got lol. The main thing is to identity what the hell is wrong with it in the first place.... b/c that's what you focus on.

I got a really bad vocal track earlier this week. I mean... really bad. They sent it in a low quality MP3 for starters. There's a high pitched noise throughout it (i later found out they recorded next to a freakin' refridgerator), and the vocals are distorting from too much shouting & not enough backing up off the mic. They also included tons of reverb and delay, like that would make it sound better lol. They did pretty much everything you're NOT supposed to do. I asked them to re-record, but the vocalist just had surgery n couldn't perform. So I asked them to resend it as a WAV and without the effects. Well... that file came out somehow remarkably worse than the MP3. It sounded like they rendered it in 8-bit lol. Luckily they weren't looking for a miracle song. My job was just to mesh it with a well recorded vocal from another poet and put background ambience to it (they're both poets).

So here's what I did to "improve" the quality of the recording (I use the term improve very loosely). I decided to use their original MP3. For the 2nd poet's well recorded verse, I added some reverb and delay to it, to try to at least put it in the same room as the poor recorded mp3. I also took some of the bass out of their voice w/ my EQ. On the mp3, I used Waves x-noise to try to get rid of as much hiss as i would, before you started to notice the voice sounding muffled. Once the voice was muffled, I eased up on it and allowed some of that hiss to come thru. Then I played w/ my EQ and decided to accentuate the lower-mid tones of the poet's voice instead of the highs. I wanted those to be stronger. I upped the low-mids, I lowered the mid-highs, and I kept the highs the same. It started to sound better. After that... it was all about the ambience behind it. They're Muslim, and it's against their religion to use anything except a daf drum. So, I found a good daf drum loop and sampled it. I carve the low end out (so they're voice would be more present in the lows), and I allowed the high's from the daf to stay there.

Overall... the mix came out audible. It didn't have a chance to be outstanding b/c of the poor recording quality of the vocal... but i made it so that at least you can listen to it and still enjoy what they're saying.

I said all of that just to give an example of how you can't really improve the quality... just work w/ what you got. I'm sure there's a few miracle workers in here that could've done even more w/ what I had...... especially with a better monitoring system than me.

In your case.... if your track is overly compressed, think about what that means. That means the dynamics are squeezed too much and the quiet parts are too loud. So... theoretically... I would try to automate or cut/slice/adjust the levels on those parts that should be more quiet and lower em to try to add dynamics. Depending on the sample, it may work... or it may sound like even more shit. If you decide to use EQ, subtraction usually will get you much farther than addition. Cut out the bad frequencies so you can focus on the good ones. Showcase the parts of the music that sound good.

[headshotsvol7]

Or an extremley easy way of doing this is finding a MIDI File for the sample you want to improve upon and use your own instrumentation and effects on it. =}

[shady11]

You can't really "improve the quality" as easily as you want to. Once you get anything mixed down, obviously you can't revert to it's prior state.

All you can really do is use your tools to try to correct problems. You're essentially working w/ what you got lol. The main thing is to identity what the hell is wrong with it in the first place.... b/c that's what you focus on.

I got a really bad vocal track earlier this week. I mean... really bad. They sent it in a low quality MP3 for starters. There's a high pitched noise throughout it (i later found out they recorded next to a freakin' refridgerator), and the vocals are distorting from too much shouting & not enough backing up off the mic. They also included tons of reverb and delay, like that would make it sound better lol. They did pretty much everything you're NOT supposed to do. I asked them to re-record, but the vocalist just had surgery n couldn't perform. So I asked them to resend it as a WAV and without the effects. Well... that file came out somehow remarkably worse than the MP3. It sounded like they rendered it in 8-bit lol. Luckily they weren't looking for a miracle song. My job was just to mesh it with a well recorded vocal from another poet and put background ambience to it (they're both poets).

So here's what I did to "improve" the quality of the recording (I use the term improve very loosely). I decided to use their original MP3. For the 2nd poet's well recorded verse, I added some reverb and delay to it, to try to at least put it in the same room as the poor recorded mp3. I also took some of the bass out of their voice w/ my EQ. On the mp3, I used Waves x-noise to try to get rid of as much hiss as i would, before you started to notice the voice sounding muffled. Once the voice was muffled, I eased up on it and allowed some of that hiss to come thru. Then I played w/ my EQ and decided to accentuate the lower-mid tones of the poet's voice instead of the highs. I wanted those to be stronger. I upped the low-mids, I lowered the mid-highs, and I kept the highs the same. It started to sound better. After that... it was all about the ambience behind it. They're Muslim, and it's against their religion to use anything except a daf drum. So, I found a good daf drum loop and sampled it. I carve the low end out (so they're voice would be more present in the lows), and I allowed the high's from the daf to stay there.

Overall... the mix came out audible. It didn't have a chance to be outstanding b/c of the poor recording quality of the vocal... but i made it so that at least you can listen to it and still enjoy what they're saying.

I said all of that just to give an example of how you can't really improve the quality... just work w/ what you got. I'm sure there's a few miracle workers in here that could've done even more w/ what I had...... especially with a better monitoring system than me.

In your case.... if your track is overly compressed, think about what that means. That means the dynamics are squeezed too much and the quiet parts are too loud. So... theoretically... I would try to automate or cut/slice/adjust the levels on those parts that should be more quiet and lower em to try to add dynamics. Depending on the sample, it may work... or it may sound like even more shit. If you decide to use EQ, subtraction usually will get you much farther than addition. Cut out the bad frequencies so you can focus on the good ones. Showcase the parts of the music that sound good.

This was a great post.

[TheEstroMan]

c phoenix do u know how to master a mixtape? jw im getting close to being done with my mixtape and was looking for someone who knew how, check out my page on here i posted acouple teasers to it its the top 2 videos so idk get at me when u can

[Dre Major]

As already said you're pretty much stuck. Best you can do is use it as is making the sound of it, as crappy as it may be, seem to be done on purpose. Or try to use some heavy EQ to get it close to workable again.

[afasio]

thanks for the helpfull answers, i do wanna tell you that i sample fragments of beats / so i have a fragment of like 12 seconds with somthimes 2-3 instruments. so i cant really set the eq on each instrument but i have to eq all 3 at the same time which is pretty difficult but thx a lot!