View attachment BOTW Screen Shot.doc
Hi all,
I just found this Forum by searching "Spectrum Analysis" and this is my first post. This Forum is for very advanced people, I think, but I hope I will be steered in the right direction to find my simple need.
1. I listen to 2-ch music and at the moment I am listening to CDs with very low frequency content, including Church Organ music.
2. I would like to be able to identify the frequency of the note at certain portions of the track.
3. I have Audacity and a spectram of the sound wave. There are, from left to right, peaks that show 80hz, 39hz, 26hz ---reducing to 5hz on the right which is almost flat.
So which is the one I can be confidently say is the frequency of the note?
4. What other software can I use so I could have the figure stamped on the screen as a record?
Your help will be most welcome.
Regards
William

 

One of your problems is that few instruments produce pure tones - http://en.wikipedia.org/wiki/Overtone In your organ music, you may hear the pedal notes, plus what is being played on the manuals.

Almost forgot, we have a related thread - http://forum.allaboutcircuits.com/showthread.php?t=54969

 

First, welcome, and all that. I can already tell you are not the most beginnerish person on this site, but we also have people here all the way up to "rocket scientist". Very handy!

The biggest and lowest frequency peak is the "fundamental" frequency of some note. As beenthere said, most notes are not pure. That is why multiples of the fundamental note show up on the graph. The multiples are always higher than the fundamental, usually multiples of 2, unless perhaps a door is rattling and thus transmitting all of its resonant frequencies.

'scuse me. Family interruption.

 

So...it is good to know the frequencies you are looking for, like "E" at multiples of 40 Hz or "A 440", but most "standard" notes are very strange frequencies. Mathematically, the notes are figured at (the twelveth root of two) times the fundamental. 40 Hz (bottom string on a bass guitar) times 1.05946 = 42.3784 Hz for "F". Multiply that answer for F# and you get 44.89822 Hz, and on up until "E" @ 80Hz. You are not going to find perfect tuning in the real world. Don't expect it. Remember "the well tempered clavier"? Mathematical precision doesn't actually sound the best, but a few hours of practice and you'll be naming notes almost immediately.

It would help a lot if you could set that graphing program to display the frequency instead of the time. ps, frequency is 1/time, but you already knew that. I don't know how to work with your graphing program. This is only my 2 cents worth. Other people will be watching and some of them will be helpful. Ask another question if you want to.

 

If you are after a functional result you can get right now, you could download any of the free wave editors (like Audacity) and click the waveform to set loop points, so if a organ note of 1 second length is playing just loop it and it will play as a continuous note.

Then play a note on your keyboard or guitar etc and you will very quickly find out what the note is.

I have done similar things on home studio equipment trying to transcribe notes in fast guitar solos. Looping it to play a single note (or chord) continuous or a few notes in a tight loop, you can soon work out the notes.

 

It would help a lot if you could set that graphing program to display the frequency instead of the time. ps, frequency is 1/time, but you already knew that. I don't know how to work with your graphing program. This is only my 2 cents worth. Other people will be watching and some of them will be helpful. Ask another question if you want to.

Thanks #12,
Please find the sample screen shot of the wav file (which I am not able to attach) attached. I have used Free Audacity software.
can you take me further with this?
Regards
William

 

I didn't look at your screen shot but what I described is quite straightforward.

Click START and END points on the waveform, then press PLAY. Audacity will play that little chunk in a repeating loop.

Then adjust the start and end points to just loop and play one note, you may need to zoom in on the wave to place the start and end points.

Then once it is continually playing one note in a loop, play some notes on your piano/keyboard/guitar until it sounds like the same note.

 

I see you have accomplished frequency labeling and found a note at about 78Hz. That is closer to E than D#. Out of tune or inaccurate software? I don't know.

I also don't know what your next question is. "take me further" does not make sense to me. Please try again.

 

I see you have accomplished frequency labeling and found a note at about 78Hz. That is closer to E than D#. Out of tune or inaccurate software? I don't know.

I also don't know what your next question is. "take me further" does not make sense to me. Please try again.

Thanks again.
You picked the 78Hz, is it because that is the loudest (-30dB)? There is another peak of 70Hz on the left at -35dB and to the right of 140Hz at
-39dB and others as well. So, is the note with the highest SPL is the fundamental note we are looking for? May be I am asking a very simplistic question of a complex waveform.
Just so that you may be able to enlighten me further, the track was from Simon and Garfunkel's Bridge Over Troubled Water at about 3.10 to 3.15 seconds into the track. If you were willing to spend the time and do the analysis in a professional way, perhaps you could comment some more with graphs?
William

 

Referring to post #3, the biggest and lowest frequency bump is the fundamental frequency of some note.

Telling me that this a graph of some music tells me why there are so many bumps. This is not an impure note, it is a whole orchestra. The fundamental frequency of the root of the key they are playing in might be a lower frequency than the biggest bump, but it got covered up by so many other instruments or bandpass limitations in the processing stream from the recording studio to the graph you presented.

I am not a professional audio analyst. I am an old, precision analog engineer that worked with a band for several years. That is probably why I become more puzzled as you continue asking. I don't know what you want.

You can now identify the frequency of a note at certain portions of the track, as requested in your first post, but not because I did anything helpful.

 

Thanks to all who responded to my post. I know a bit more now then when I started. I will need to learn some more about Spectrum Analysis.
William