Connect with us

Resample MHz signal

Discussion in 'Electronic Design' started by James Barlow, Aug 19, 2007.

Scroll to continue with content
  1. James Barlow

    James Barlow Guest

    Please excuse me for starting another thread on this topic. My
    apologies for not being specific enough the first time. But I did
    learn something.

    What I am really looking for is external hardware that will allow me
    to upload an audio frequency WAV file, such a voice or music, from a
    PC and then play it back at 100 times the original sample rate.

    The program I am using is CoolEdit, which has a built in tone
    generator with modulation and eveloping. It can produce 8, 16 or 32
    bit resolution at 12 discrete sample rates between 6000 and 192,000.

    The idea is to take advantage of this program's functionality and
    user-friendly GUI to produce complex signals above the audio range.

    PC software (alone) is not a solution as the resulting bandwidth would
    be 2KH-2MHz, and therefore beyond the capablities of a standard
    soundcard. DSP-based boxes for musicians are equally unsuitable for
    the same reason.

    I seem to need a lab-type device, or something I can modify or build
    from scratch. Do I have any viable options here?

    I have Googled big time and so far found nothing. Signal transposers
    would be a great project for an electronics magazine. Think of all the
    applications.

    I hope my aim is clear enough to attract an expert response.

    James Barlow.
     
  2. MooseFET

    MooseFET Guest

    Basically what you are after is a DAC that can go up to several MHz.
    If the record is fairly short an "arbitrary function generator"
    may be good.

    Signatec makes some PCI DACs IIRC

    Dyneng makes some PC104 ones, I think but they aren't likely fast
    enough.

    If your computer has an ISA slot doing a home made DAC wouldn't be
    very hard. You would want to provide a little bit of buffering FIFO
    and self clocking. The timing of writes on a PC is quite jittery.

    The easy way to build this would be to make the PC into a "box that
    does this function." Basically what I am thinking here is that you do
    this part in DOS or Linux and basically take over the whole machine
    while you do the output.

    On DOS machines that aren't laptops, the 18.2Hz interrupt can be taken
    over and can be used to transfer bursts of data at a much faster
    rate. Laptops and Windows machines are less likely to be able to do
    this. For some reason the BIOSes commonly used in laptops leave the
    interrupts off for longish times.

    Linus may be able to do it. Windows is extremely unlikely. In both
    cases you need to be able to get to the code that sends the burst at
    regular times.
     
  3. Phil Allison

    Phil Allison Guest

    "James Barlow"


    ** Which will NOT be the case with a 100 times increase.

    eg 50Hz x 100 = 5000 Hz = very audible.

    You are obviously a clueless jerk on a fool's errand.

    God knows what crackpot idea you have in mind but will not tell us.

    Frequency shifting the audio band up into the supersonic region is easily
    possible with analogue circuitry, but will not compress time duration of
    course.




    ........ Phil
     
  4. colin

    colin Guest

    a video card has a dac that will go upto 200mhz+ wich you can use.

    I dont know how to do it myself exactly but ive read about it being done and
    considered doing it myself a few times.

    for a repetative pattern its probably quite easy.

    you have to set the video timings so you dont get blanking etc,
    and put your output into the video display memory.

    ofc you wont be able to use your display at the same time,
    unless you have a seperate vga adaptor.


    COlin =^.^=
     
  5. Bob

    Bob Guest

    I'm guessing that you want to send the resulting signal through the
    air. All the ultrasonic transducers I have read about have a resonant
    point
    and the frequency response drops either side of it. There are some
    with -6dB points at around 50% and 150% of the resonant frequency
    but I'm not aware of any with multi-octave response. Anyone?

    Perhaps a combination of frequency shfiting and frequency
    multiplication
    might be more practical eg shift 300Hz-10000Hz to 40300-50000
    then play back 44 times faster to get a signal about 400KHz wide
    centred on 1.98MHz.

    Perhaps do the frequency shift mathematically into a 192KHz PCM .wav
    file then a DAC running at just over 8Msample/sec.

    Getting data out of a PC that fast is not that easy. I'd be thinking
    of
    an ethernet device with some buffering but that's mainly because I
    view low level drivers as a horrific thing to take on.
    Many audio applications can loose the bass without anyone caring.
    The tiny loudspeakers in my laptop do nothing below 300Hz.

    For voice you can get away with 300-3300hz though up to 8000hz
    helps intelligabilty.
    The simplist approch, mixing with 40Khz produces a voice signal
    at 40300-43300 and an inverted voice signal at 39700-36400

    An analog signal that avoids that seems to involve messing
    with a multi-octave 90degree phas shifter.

    Bob
     
  6. Phil Allison

    Phil Allison Guest

    "Bob"
    "Phil Allison"
    ** A very wild guess indeed.


    ** The OP has stated he wants good waveform accuracy.


    ** For which low cost, easy to build, published designs exist.

    Including one from me in the August '97 issue of Electronics Australia
    magazine.

    ( 22Hz to 20 kHz , +/- 2 degrees ) as part of a + 5Hz shifter for
    acoustic feedback suppression.




    ........ Phil
     
  7. Nobody

    Nobody Guest

    The DAC is only 8-bit, which may not be enough for the OP.

    If it is, then a video card gives you 3 x 200MHz+ DACs for a fraction of
    the cost of the same thing sold as "test equipment".
    Not all hardware supports this. There's invariably a fixed upper limit on
    the number of pixels per line (determined by the number of bits in the
    corresponding register), and you can't always reduce the blanking to zero
    (the horizontal blank may be used to refresh DRAM).

    Also, if you want to program the video hardware yourself, getting
    documentation for modern video chips is far from straightfoward. Even
    with the documentation, programming is far from straightfoward. If you
    can get something from before the era of hardware 3D, it will be a lot
    simpler.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-