How To Guide|Midi/Sheet Music to Audio

[pcgumshoe]

PCGumshoe’s way of converting midi to descent audio tracks:
Comments and Recommendations Encouraged

Purpose: These instructions are for people who want to do their own thing, not for people who want to save $20 to sing just one song.

Hardware Needed:
A PC with a lots of RAM and a good Sound Card (Stereo is all, not 6.1 inside your head sound)
A good set of headphones ($50 doesn’t buy a good set, but will do)
Software Needed:
Finale 2002, Sibelius or Print Music (or a pre-made midi file) Cost $80-500
Sonar 2(2) or Sonar 2(2) XL or similar product $250
Reason 2.5 $250
KPro Check this site
Microstudio Check this site
Notepad Free


The only reason why you would do this is your song does not exist in karaoke and you really want to sing it, or you want to make songs for people. I won’t address the legal issues here, but for private, non-commercial use I would relate this to playing a piece of sheet music on your piano. You should be able to find at least the Piano/Vocal Score of what you want to sing. If that is good enough for you, you might not need as much of the aforementioned software. Piano/Conductor’s Score will have all the movements you want or need, but you will have to place them yourself. You should have a basic understanding of music, notation and a good ear. Remember, you are trying to imitate an original here.

If this is more of a hobby, hope that someone knows the song that you want to sing better than you and has already sequenced it. You can do a search and check out midi rings to look for the song you want. If you find one that sounds like all the notes are there, it might work. Don’t listen to the quality, you CAN change that.

Now, with your notation software (Finale) key in all the music to multiple staves and then have each instrument's music represented. Don’t worry about the sound quality; you can’t get some instruments to sound descent from your sound cards audio files. Once you have the music in your notation software, then save it as a midi.

I have not found any recognition software for music to be useful. Also, I don’t play piano so well, so I use speedy notation. If you play piano, you can play and record all the tracks independently.

Now open the midi file in Sonar. In Sonar, create copies of some of the instruments have many tracks of the same instrument. This will be important because some instruments play differently than others at certain parts of the song. For example, you may have a trumpet line and at a certain point, the trumpets are muted. Two midi lines allow you to assign two different sounding instruments to play the same notes. Cut and paste is a nice thing. Copy a full line and then delete the parts you don’t want the new sounding instrument to play.

In Sonar, you will use REWIRE to open up Reason. Reason has a lot of nice features and looks like a rack of audio equipment check Reason out HERE

I don’t spend a lot of time on my drum lines except in the case of dramatic rolls of a Timpani or Snare. If using a pre-sequenced midi file, use Edirols sounds in Sonar. That way you don’t have to reassign the drum notes.

Take time to find the right sound for an instrument. There are many adjustments in Reason. You may have to invest in new instrument refills for Reason. In no way should you think that by buying Reason you will get the instrument that you want, but you can buy it and add it to Reason. Reason accepts many formats for recorded sound fonts. If you don’t like the piano, you can always buy the Steinway Sound Fonts (around $200 but cheaper than a Steinway or a musician).

Play and replay certain sections in whole or in part or just solo instruments to make sure that everything blends smoothly and sounds great. Also, add some extra space before the first note and after the last. Use one measure before the first note. This give you time to have an introduction. At the end of the song, add two or three measures so that any echo or reverb isn’t immediately cut off in the final audio file.

You may want to have two versions of the song. One with the vocal part played and one with it muted. If you, or your singer doesn’t want to hear the vocal part and just sing with the instruments, having a copy of a file WITH the vocal part played may help you to tap in the lyrics in KPro later.

Once you have a descent sounding audio file, mix it down through Sonar. This mix down will take and play the song at an increased speed (inaudible) and produce a wav file for you. This wav can be used now to create your karaoke song.

It doesn’t hurt to play the wav on a CD in a real sound environment (where you do your karaoke shows, for example, or your home theater system).

You can find the lyrics anywhere, even in the sheet music! But to save time, I usually search for them on-line. Cut and paste them into Word. I don’t use word for my final lyrics file, because I sometimes need to make quick adjustments to something and Word takes to long to load. In word, I will make all sung words in Caps. Select all, and choose font and select “ALL CAPS” once I have the lyrics in this format, I will copy them to a txt file. You are under no obligation to use this style, but it works well for me. Be consistent. Sound Choice has the same style all the time, only rarely do they change it.

It is nice to add things to your song that help you and your singer keep on track. If I am making a song that I don’t know too well, these cues also help me synchronize the lyrics better. Anything that is not sung, like spoken words I will us in lower case. Using words like INTRODUCTION, INTERLUDE or OUTRO will help let your singer know when they are to sing or pause. For short pauses I will use the a couple equals signs ====. Another very helpful thing is to insert “*” in between syllables. That way you tap in Kpro according to the syllables and don’t get ahead of yourself. A word like “ANYBODY” written that way in the txt file will need only one tap, typed like this, “AN*Y*BO*DY” will require 4 taps. Also, two or more words can be done in one tap by holding down the “ALT” key and typing “0160” on the number pad and then releasing the “ALT” key. This creates an “M” space that connects two items as one. I will use these for many things “(TAKE A BOW)” could be used as a closer to a song as an instruction to a singer, using the “M” space would make it sweep as one, with out them in place of the spaces, it would take three taps.

If you get everything in correctly, don’t forget to change your wav file if you were using two versions. Then export to CDG and use Microstudio to burn to

If you have any other questions or comments for me personally on this, email me.
Otherwise, I would love to hear any other recommendations.

This is just a beginning of my how to, but I hope it helps

[gduns - with the Lord]

Have you ever tried a midi file from your midi out (on the sound card) to a synthesizer. I have a keyboard, and a synthesizer and was wondering if this would sound better than running a midi file on say..... soundblaster, or is it about the same.

[pcgumshoe]

Reason effectively emulates a synthesizer INSIDE your computer. No need to run it out. If, however, you have a synthesizer that you are found of, yes, you can use that. You would have to record the sound, however.

So you would go out digitally to your midi and then back into your compuer analog and convert that to digital and then have a wave. Assuming your computer could record that input flawlessly, you would be fine.

If you have a newer sysnthisizer, it should have the ability to import sound fonts. Reason can use the same sound fonts as the synthesizer.