Here is my problem:

I have to transcribe (make a score) of a recording
the recording (mp3) is a piano playing rythmical paterns + an at the time electronically generated duplication starting about 0.5 seconds after the piano

my question: would it be possible to deconstruct this soundfile to the piano only (so it'd become much easier for me to transcribe it)

Thanks and by the way, i am a complete layman as far as soundmanipulation is concerned, so be gentle.

greets,

R

Welcome!

I'm not sure I fully understand the question, but...

...if you are asking "is it possible to separate one instrument from an audio recording where it is mixed with other sounds/instruments playing at the same time?" the simple answer is "no". Many liken this to trying to make a cow back out of a box of beefburgers (or maybe a horse nowadays...)

...if you are asking "is it possible to edit sections of audio where the single instrument I want to hear is playing on its own?" the answer is "yes". All you need is a simple audio editing package and you can chop the file up into as many sections as you like.

If you can get a MIDI file of the piece you are working on then it's job done - pop this into a typical DAW (Cubase does it, if you know anyone with a copy) and it will display the score.

Hope this helps - or please correct my understanding of your question!

--------------------
An Eagle for an Emperor, A Kestrel for a Knave.

You understood correctly (answer 1)

though i do think the meat simile is a bit out of proportion
In a way all i'm trying to do, or rather undo is a delay effect (0.5 sec) on a mp3 file.
Mmm, I feel technology is letting me down on this one
("why is software not smart enough to decode raw soundfiles?")

But thanks anyway!

So basically you want to kill an echo because it makes transcribing harder? Don't think that can be done.

Some things could make the job easier:

(1) If you know the person who made the recording, ask them for a version without the echo.

(2) Play around with EQ, this may help bring the main notes out more (assuming the echo doesn't sound exactly the same as the main notes).

(3) Put the file into an audio editing program like Cubase, Reaper, Logic, Wavelab, etc. and slow it down. That generally makes transcribing easier. It may also help you distinguish the main notes from the echo (unless the clash of note and the echo from an earlier note happen very closely together).

Quote Dragonetti:

---

In a way all i'm trying to do, or rather undo is a delay effect (0.5 sec) on a mp3 file.
Mmm, I feel technology is letting me down on this one
("why is software not smart enough to decode raw soundfiles?")

---

Because they aren't 'raw' soundfiles. An MP3 file doesn't contain separate pieces of data for each instrument or part. It simply holds a representation of a single waveform of the entire mix of audio.

Software has no real understanding of the audio content, or of musical content - and has no concept of which parts you might want to hear and which parts you don't.

--------------------
An Eagle for an Emperor, A Kestrel for a Knave.

Quote Dragonetti:

---

why is software not smart enough to decode raw soundfiles?

---

For the same reason you can't easily separate warm water into hot and cold. I like the horse idea, myself.

--------------------
Ohm's Law states, "Your PA isn't as powerful as you think it is".

If all the file is is a delayed piano part, Melodyne Editor (with DNA access) will be able to decode the two (piano and delay), with some initial setting up. You can then just delete or mute the delayed notes and even export it as midi. The results and accuracy however depend on the cleanness of the original recording and the resolution of the mp3 conversion.

Cheers,
Sam

--------------------
http://www.soundcloud.com/sambrox
seedy.dk