sugestions for 6 channel 16 bit 44.1Khz

Discussion in 'Embedded Systems and Microcontrollers' started by David Walker, May 7, 2015.

  1. David Walker

    Thread Starter New Member

    May 7, 2015
    5
    0
    Hi Guys,

    I'm new to the forum, I am looking for suggestions on where to start with a project idea I have and the best way forward. I apologise if something similar has been asked but I can't find anything that directly relates to what I'm looking for and thought you guys would like to discuss the project.
    Also this question is mainly about the selection of a microelectronic and parts so I assumed this was the correct place to post this question.
    I am looking for a programmable controller that will allow me to take 6 audio inputs and convert them into a series of signal-derived information messages rather than actual digital audio (i.e. data packets; if USB(frequency and intensity), if Ethernet(target IP, socket, info packet(frequency and intensity) ) ) and broadcast them, ideally, via Ethernet or USB (not to fussed about Ethernet or USB at the moment as this is just a prototype to prove proof of concept), to be used by a desktop application.
    I would appreciate any advice on what kind of hardware you guys would choose for this type of project. Also, please don't worry about the development language as I already work with c, c++, java and others and I am not at all worried about learning a new language if it suits the project and hardware.
    Thanks
     
  2. Papabravo

    Expert

    Feb 24, 2006
    10,136
    1,786
    The big problem with your proposal is that it ignores several crucial points.
    1. Nobody cares about 16/44.1 anymore, 24/196 and up is where it is at.
    2. If your scheme has ANY possibility of introducing latency to the listener, it will fall with a great big THUD.
    3. USB audio formats are already different than USB data formats for a reason.
     
  3. David Walker

    Thread Starter New Member

    May 7, 2015
    5
    0
    Thanks Papabravo, I appreciate you taking the time to comment.
    To give further detail;
    1. Nobody cares about 16/44.1 anymore, 24/196 and up is where it is at.
    Of course, as an audio file, I would love higher sample and bit rate and if cost of hardware is similar and latency of ADC and Music/Audio information retrieval is not effected
    then I would definitely consider it. However, as a specification for minimum requirements 16bit 44.1Khz is more than ample.

    2. If your scheme has ANY possibility of introducing latency to the listener, it will fall with a great big THUD

    The idea is of a visually representative system played back in semi-real time, intermittent packet loss (similar to expected loss when using UDP) and delay of up to
    30millieseconds would be acceptable.
    It's a system for semi-real time real time and comparison than playback.

    3. USB audio formats are already different than USB data formats for a reason.

    All that is required is to send data packets with frequency and intensity data, in reality the rate of this would be at a fraction of real digital audio and would guess that data
    packet transmission would be effective with message rates between 100 and 200 per second.

    The idea is that the target computer receives audio information (as opposed to actual audio) which can process it further.
     
  4. Papabravo

    Expert

    Feb 24, 2006
    10,136
    1,786
    The positive benefits of your idea escape me at the moment. Most people who listen seriously are deeply suspicious of any compression or data efficiency scheme. Of course if the audience could care less about fidelity then any such considerations are superfluous.
     
  5. bertus

    Administrator

    Apr 5, 2008
    15,645
    2,344
  6. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,772
    931
    if your idea is to capture data for things such as lighting and effects timing sequencing then the low data capture rates and high milli-second latency acceptance are understandable. If so, or if not, please indicate.
     
  7. David Walker

    Thread Starter New Member

    May 7, 2015
    5
    0
    Again, thank for your comment, I do appreciate you taking the time to help.
    This is not going to be a system for playing back audio, a new compression method or standard for audio data transmission. what I want to do is take 6 audio inputs and convert them from raw analogue audio into quantified information such as frequency and intensity and send it to another computer for further processing.
    The computer receiving the data can then use it as control data for things such as switches, etc (i.e. at frequency X with intensity Y from channel 1 then run function XY1).
    I understand that is novel and the benefits are not immediately apparent and would like to go into further detail on its final application but as it is very technical and would require a much longer description I am hesitant to describe it further here.
    Essential I envision a box that will take 6 audio inputs, convert them to data packets containing frequency, intensity and channel, then send them off these data packets to another computer for further processing. I could do this with a small form pc such as am Intel NUC, or Raspberry PI but would much rather use a programmable controller.
    How would you approach it?
     
  8. Papabravo

    Expert

    Feb 24, 2006
    10,136
    1,786
    Then it really doesn't matter what you do or how you do it because there will be no obvious relationship between the system inputs and the systems outputs. Why bother with the transmission of data, why not just make it up out of whole cloth on the receiving end. How could anyone tell the difference between Pearl Jam and Beethoven's 5th?
     
  9. David Walker

    Thread Starter New Member

    May 7, 2015
    5
    0
    Hi Guys,

    Thanks so much for the input so far.
    I've mocked up a diagram to clear up any confusion;

    upload_2015-5-8_12-1-8.png
    I've changed the design slightly in order to utilize the 8 input channels from the Sigma Delta Pi 18 (https://www.abelectronics.co.uk/products/3/Raspberry-Pi/14/Delta-Sigma-Pi). I would like as many inputs as passable but 6 is the minimum number of separate channels required to show proof of concept.
    so far I'm thinking of using a Raspberry Pi 2 with the above 8 channel ADC, what do you guys think?
     
  10. David Walker

    Thread Starter New Member

    May 7, 2015
    5
    0
    Probably a major point, but all input singals are assumed to be monophonic. Maybe should have stated that in the original post.
     
  11. Papabravo

    Expert

    Feb 24, 2006
    10,136
    1,786
    Your block diagram clarifies a great deal. Certainly the pieces that you want to assemble exist in the libraries of the FPGA vendors. You just have to select one that has enough resources to do everything you need. I would suggest you start with a single channel and work your way up from there. What does your budget look like and how familiar are you with various methods of developing an FPGA?
     
  12. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    2,907
    2,163
    At the sample rate of that ADC (The data rate for analogue to digital conversions is 3.75 (18 bit), 15 (16 bit), 60 (14 bit) or 240 (12 bit) samples per second.) a RPi2 should be able to handle the task easily.
     
Loading...