Hello,
I am new but have tried to read everything I can. I select the entire 3.5 minutes of an mp3 and then export to “other type” then SDS, and all seems to go well, but despite multiple attempts and reinstalls, the file is always only about 30 seconds long. I have a ton of space, so that is not the issue. The computer is a data cruncher, so it is not challenging its processors. Any ideas on what would truncate all attempts at conversion?
Thanks,
“SDS” is not a MIDI file. It’s a “MIDI Sample Dump”. That’s a format for putting audio data into a sampler. MIDI is not audio and it is not possible to “convert” audio to MIDI.
MIDI is like a digital form of musical score. It does not contain sound, but rather a list of commands that tell a MIDI device what to do - typically a MIDI file contains “note data” to tell a synth which notes to play. Creating a MIDI file from audio is like a digital version of writing out a tune in musical notation.
There are some apps that can automate the creation of MIDI files from audio, but mostly they work only with monophonic sounds (simple tunes that have just one note at a time - no harmonies or chords). There’s a free on-line app of this kind here: https://basicpitch.spotify.com/
A number of sites claimed this software could convert MP3 to MIDI. I know the differences and was pretty impressed by the claim. Maybe they stretched the programs abilities just a tad . I’ll make sure no one else adds more but suspect you are correct.
Yes. Many of them are scam sites, or click bait advert sites, while others are (like the “basicpitch.spotify.com” site) genuine but limited. The problem is the word “convert”.
The word “convert” is usually meant in the sense of modifying a thing into a different form, where the “before” and “after” are both essentially the same thing, or at least the same sort of thing. Example: convert ice into water by heating.
but what if you listen to a poem and write it down with pencil and paper. Have you “converted” speech to text? The “before” and “after” are completely different things - different kinds of things - one is sound and the other is text, where the text is an abstract representation. “Converting” audio to MIDI is like this.
I was very pleased to find the “basicpitch.spotify.com” example as it is a safe example (currently does not even have adverts) of what kind of “audio to MIDI conversion” is actually possible. There are also some expensive apps that can work to some extent with polyphonic sound (multiple notes at the same time), though the more complex the sound, the less accurate they become. “Melodyne Assistant” is one of the best (€249), but even that struggles with complex music and some degree of manual correction is usually required.
Thanks, and I agree. Interestingly, a discerning human ear can do this, albeit slowly. However, computer software lags a bit even though music is simply a mathematics construct converted to sound energy. One would think this would be “easy” for a computer to discern the frequencies and timing and deconstruct this information back to basic numbers. This could then be reproduced into sheet music or a MIDI file. Just shows the actual potential power of the human mind, especially if trained and experienced in a task.
One would think this would be “easy” for a computer
Yeah… When it comes to audio there are lots of things a computer can’t do very well… For example, if you record an interview in a noisy restaurant noise reduction software can’t remove the background noise without damaging the audio you want to keep. When the noise is bad, the cure can be worse than the disease… Noise reduction woks best when you have a constant low-level background noise… When you don’t really need it. Sometimes in movies or on TV they will pull a conversation out of party noise. But that’s science fiction… The human brain is often the best filter…
There are other examples too… Most audio defects can’t be fixed! And noise is the biggest “remaining problem”. You can make a pretty good recording at home with relatively inexpensive equipment (compared to the analog days) but building a soundproof studio is still expensive and it’s the biggest difference between a “home studio” recording and a pro studio recording.
And all real world sounds contain harmonics & overtones. That’s why no two singers sound the same, even when singing the same notes and it’s why a trumpet sounds different from a guitar.
Usually the fundamental note-frequency is the strongest, but sometimes a harmonic is stronger. That’s not a huge problem with single-note detection since the harmonics are the same note at higher octaves.
But with chords and/or multiple instruments there can be thousands of simultaneous frequencies! Audacity’s [u]Plot Spectrum[/u] can give you an idea of the complexity.
But, computers are getting more powerful everyday! The speech recognition and content analysis recognition with Alexa and Siri are amazing, but they run on powerful servers on the Internet or in ‘the cloud’.
So true. I am still trying to block my computer hum every time I record educational videos. I could encase it in foam but then it would melt.
Randy