Audio editing:

Discussion in 'other software & services' started by TECHWG, Oct 2, 2006.

Thread Status:
Not open for further replies.
  1. TECHWG

    TECHWG Guest

    How can i take a recording and just transpose its frequency ? By this i dont mean use the rediculous pitch strech feature, i mean almost mathmatically take the sound and up its frequency to another start point, so for example if the base frequency was 1000 Hz, up sample it so it started at 5000 hz etc

    Does anyone know o_O?
     
  2. TECHWG

    TECHWG Guest

    any ideas ?
     
  3. Mrkvonic

    Mrkvonic Linux Systems Expert

    Joined:
    May 9, 2005
    Posts:
    8,698
    Hello,
    Do you mean how you can do it in theory or some software?
    Just shooting in the dark, but you need Fourier transforms and compression algorithms. Could you clarify what you intend?
    Mrk

    P.S. Excellent programs for editing audio are opensource projects:

    Audacity

    VirtualDub

    You can do practically anything with those, if that's what you mean.
     
  4. TECHWG

    TECHWG Guest

    No, i have and own Adobe audition, but i am looking for a kind of a "agh thats easy click here drag there type something like this in and Boom you have it" or if it can not be done with audition what will i need and how it can be done

    Thanks
     
  5. Mrkvonic

    Mrkvonic Linux Systems Expert

    Joined:
    May 9, 2005
    Posts:
    8,698
    Hello,
    Like I said, try those two programs. Adobe might not be able to solve your problem. On the other hand, I believe you will succeed using VirtualDub and Audacity.
    Mrk
     
  6. TECHWG

    TECHWG Guest

    i tried audacity, but it does not seem to have the ability to do what i want. With adobe i am kind of able but with a lot of fudging with the audio . . all i want is clean upping or downing of an audio signal to another pitch .. Almost mathmatically so there is no loss kind of like let a = 99 let b = 100 etc
     
  7. TECHWG

    TECHWG Guest

    you seem to know what you are on about, maybe you could help me find something to do this with perhaps ? adobe seems to go weird when making signels higher and lower they sound like a angry chipmunk with larangytis or a darlek that fell down some steps and is hurt
     
  8. Mrkvonic

    Mrkvonic Linux Systems Expert

    Joined:
    May 9, 2005
    Posts:
    8,698
    Hello,
    Patience is important. Why don't you try VirtualDub. Try resampling. Try compressing the sampling. Stretching it. That might work with frequencies. And I really have no idea what Adobe does to invalid chipmunks.
    But if you present your problem in detail, I might be able to help.
    Mrk
     
  9. TECHWG

    TECHWG Guest

    i have described my issue in as much detail as i can . . think of a slider knob on a recording studio desk . . and that rectangle knob is the chunk of audio . . and its sat at the 0 mark . . if you slide the knob up now its at 12 (the audio - (knob) - was slid up to a higher frequency than original)
     
  10. dog

    dog Guest

    Audacity should do what you want to do (I think --- unless this was what you were referring to in post #1 -- "By this i dont mean use the rediculous pitch strech feature, i mean almost mathmatically take the sound and up its frequency to another start point, so for example if the base frequency was 1000 Hz, up sample it so it started at 5000 hz etc") ... see example in screenshot.
     

    Attached Files:

  11. TECHWG

    TECHWG Guest

    i have seen this but it freezes up i am going to do it again and see if ti unfreeses and if it does what i need . . thanks i am taking a voice sample and putting it from normal voice frequency to 20khz above my music i made so i can brand the copy with the persons name so i know if its given out . . .
     
  12. TECHWG

    TECHWG Guest

    ok the program works, but its not doing what i want unfortunatly . . it spreds the frequencies all over the range . . i just need to push them up and possibly compress the frequencies ? so that the audio gets pushed up in the frezuency range where it will all fit from the 16khz - 19 khz
     
  13. dog

    dog Guest

    Edit: Sorry I didn't see your latest post before I posted. ;)

    You're creating a seperate/new audio track for the voice --- then changing the freq of it --- and then combining the two tracks, right? - as opposed to recording the voice track directly to the main audio track.
     
  14. dog

    dog Guest

    Just a thought but ... What about using the High Pass and Low Pass Filters after applying the freq/pitch change to restrict the range?
     
  15. TECHWG

    TECHWG Guest

    well see i can fix the excess with the spectral view . . select what i dont need adn delete. but the thing is i record my voice on a sterio track, then alter the frequency using my own method that unfortunatly is not perfect, then place it on the track so it looks like this. i tried to attach the pic to this post . . i hope it workes. But my thing is i CAN do what i want, but the quality of the voice is not what i wanted since its loosing frequencys.

    All i need is a way to just move or nudge the freequencies from where they are straight through the roof where ever i want . my method is lossy


    ALSO compress them to a smaller range as you see above the track: My main talking is the big bulk at the bottom , then another of me talking was compressed the best i can and upped on the freequencies . . when i reverse the process i can tell my words but the quality is not good
     

    Attached Files:

    Last edited by a moderator: Oct 3, 2006
  16. TECHWG

    TECHWG Guest

    i edited my last post to let you know . . i need to do this what is on the picture but i need t do it with quality, by doing it properly . . compressing the audio into a smaller space of frequencies adn then send it up the freequencies to 16khz + etc
     
  17. StevieO

    StevieO Registered Member

    Joined:
    Feb 2, 2006
    Posts:
    1,067
    I understand what you're trying to do, but unfortunately physics comes into play here.

    The human voice covers aproximately 6 octaves, give or take. So for this example let's say it ranges from 100Hz to 6.4kHz. Therefore you will be trying to upconvert to between 16kHz and 1024MHz, which is not possible due to the limitations of the digital equipment and processing !

    For a start the filters that are required for A/D/A conversion operate around 22kHz, so most of your signal would be filtered out anyway, even if it were possible with the software. Also the software is not designed to work at frequencies outside it's normal working range, 20Hz - 22Khz approx. So whichever way you look at it, it can't be done as you wish.

    So these are the reasons the converted voice sounds the way it does. All you are getting is a very limited range of frequencies being converted. The rest is missing as it's not even getting converted due to the inbuilt filtering which has to be there to make digital audio work. Otherwise you would be hearing 100% distortion all the time.

    An alternative suggestion, would be to forget about the upconversion voice idea, and instead record a series of tones, maybe spaced in time, and of different frequencies too, to identify each individual track. For example, you could have an 18Khz single tone running all the way through track one, an 18.1Khz single tone running all the way through track 2 etc. Right up around approx 22KHz. Then you could get creative by introducing the ideas i mentioned above, which would give you many more combinations.

    I would also experiment with the levels of the tone/s, as i don't expect you would need them raised very high. And it definately wouldn't be wise to do so, as you would burn out yours and other peoples tweeters in their speakers, and the amplifiers would be working overtime as well !

    Try it out and let me know how it works out.


    StevieO
     
  18. Chris12923

    Chris12923 Registered Member

    Joined:
    May 31, 2004
    Posts:
    1,097
    I hope to have you an answer soon. I emailed someone that can probably have you a good answer.

    Thanks,

    Chris
     
  19. TECHWG

    TECHWG Guest

    thankyou very much for all your help and i hope also that i can get an answer . . the reason i wanted to do this originally was because i saw this is what they do for silent subliminals and thought what a cool idea it would be that you can have a whole monologue with thoughts/ideas/identification that is on the track but nobody can know its there . . what i have done works, and works well, but what i need to do rea,lly is make the whole thing less lossy, by properly compressing the frequencies it uses before i go altering the pitch. but altering pitch seems to be the most lossy thing . . upsampling to 17Khz and then to test down sampling back to normal leaves lots of artifacts and interferance, although i can hear my voice and discern the words , its not clear as i hoped. This is my issue. that picture above is of a sample that works . . but as i say the downsample makes it sounds crapdoody

    I look forward to hopefully getting a resolution to this and thankyou foryour continued help
    WG
     
  20. TECHWG

    TECHWG Guest

    also i forgot to mention, that becausre this technology can be used to improve the self , i also want to make my own few subliminal mp3's like the test ones i have done, (when i get the quality / method fixed ) with your guys help to figure out how to do it more correctly !

    Cheers alot guys

    WG
     
  21. TECHWG

    TECHWG Guest

    Does anyone else have any ideas while we wait for this person who asked someoneelse ?
     
  22. TECHWG

    TECHWG Guest

    ? i am really stuck
     
Loading...
Thread Status:
Not open for further replies.