Shop OBEX P1 Docs P2 Docs Learn Events
Digital recording chip suggestions — Parallax Forums

Digital recording chip suggestions

Hi All,

I have a project where I would like to record some audio. Essentially it is a old school answering machine for ham radio. The user will hit a couple of DTMF tones and be able to record a short message.

So, used any chips that you liked? Minimum external parts and easy uC control being first, record time and multiple messages second.

Thanks!

Jonathan

Comments

  • What Microcontroller are you planning to use? What language do you use?

    The Propeller chip and a ADC chip should be able to do it. It you want to go that route I can post additional info.
  • I've got a board all setup for this, the microphone is coupled into the Prop using CTRA in POS mode with feedback to digitize the audio and save this to a wave file on SD card. I even feel like just writing a few lines of Tachyon code for this so along with a few other things I might try it out in the next day or so.
  • What are the bit-depth/sample-rate requirements?
  • Hi All,

    Mick, just enough for voice mail. I'm not sure about the exact requirements but I was planning on using the lowest sample rate that the chip has to offer and get the maximum record time out of it.

    To be honest, I was thinking of using a PIC for this job. But, if I can use a Prop and skip the voice recording chip, I just might reconsider. It certainly would up the features possibilities. If I need an external ADC to use a Prop that could be a problem as I will be making a dozen or so of these devices, so cost does come into play.

    Thanks folks!

    Jonathan
  • No ADC necessary as mentioned, just a feedback resistor from one Prop I/O going back into the other I/O with small caps to Vdd and Vss and then your microphone is coupled into this with a cap. If you have an SD card connected it doesn't matter if you sample at 11kHz but I would go for 12 or more bits for dynamic range. Either way it is very simple to save each sample as a signed 16-bit mono WAV sample in a file. So that's 22k bytes/sec minimum will still record up to 54 hours even on the smallest 4GB card.
    Prop + RC ADC + SD = voice recorder
    Put an RC on another pin and you can playback audio.
  • Peter,

    Hmm. That sounds pretty cool. However, it would drive the price up. What is the minimum bit rate that would yield comprehensible audio? If I could get it down, could I use EEPROM to store it? I do like the idea of the SD card for myself and I can see a number of other possibilities that sound fun. But for the run of 12 or so that I need to make I'm not sure if it is a good fit.

    Thanks!

    Jonathan
  • I don't think you are being very realistic then if you are thinking EEPROM as those things are very very limited in terms of memory, they are too slow and they are "expensive". An SD card can be had for a few bucks and although connectors are really only less than 20 cents you probably won't pick up any in small volumes but it is perfectly fine to solder directly to the card contacts, it is a PCB after all.

    So you must have a budget in mind, do your sums, count your costs but be realistic and balanced. Sometimes trying to save a buck each on a run of 12 is applaudable if your weekly income is only $100 but is that so? Otherwise if you spend $12 on fast food then what's a few more bucks?
  • Peter,

    Well, the Prop and an SD card adds quite a bit to the cost of a unit. So I was trying to see if I can shoe-horn the Prop in there because it would add the possibility for more features. However, it seems that most of the folks just want the basic functions.

    I guess everyone here agrees, as there have been no recommendations for recorder chips.

    Thanks! I will play with the recording method you describe with a Prop for myself.

    Jonathan
  • @Jonathan

    There are various of recorder chips around if you search. I use Jameco and they have a few listed. For example:

    http://www.jameco.com/1/1/25771-isd1110p-single-chip-voice-record-playback-device.html
  • kwinnkwinn Posts: 8,697
    You could google "audio recorder chips", which will bring up a list of chips and modules to consider. Not that I am really recommending recorder chips over the propeller and SD card. When all is said and done the cost difference is small and more than made up for by the ease and flexibility the propeller and sd card provides.
  • You may be able to expand the recording tine on Raymon's post:

    http://forums.parallax.com/discussion/105292/parrot-auto-voice-record-and-playback

  • I'm always at a loss in understanding the desire/need/reason for saving peanut money when making design decisions over a handful of units. The Prop is a nice clean solution and cheap even with SD card but if I were needing to make thousands or more I might not use the Prop, and then again I might. It all depends upon the reasons for saving peanut money. You could use SPI Flash for a couple of dollars or then again people normally have old SD cards lying around too. But as I said, I'm, at a loss.
  • I'm always at a loss in understanding the desire/need/reason for saving peanut money when making design decisions over a handful of units. The Prop is a nice clean solution and cheap even with SD card but if I were needing to make thousands or more I might not use the Prop, and then again I might. It all depends upon the reasons for saving peanut money. You could use SPI Flash for a couple of dollars or then again people normally have old SD cards lying around too. But as I said, I'm, at a loss.

    Yup, totally agree. When I think about what I could buy for the cost of taking the lil lady out for dinner/drinks :-D
  • Seems the SD card offers the added flexibility of putting it in to a PC to allow easily saving greeting .wave files if you want that feature.
  • Hi All,

    I did look over the chip offerings, I was just wondering if anyone has used one that they like. I'll just order a couple and play with them. Bob, the one you linked from Jamco is one of the ones I have looked at, thanks!

    These units will be kits assembled by newbies, so no surface mount. A DIP prop is large, making a larger (and more expensive) board. It also needs an external crystal and an EEPROM, more board space and cost. I'm a Prop fan since day one, went to the first release meeting in Rocklin and use them for most of my projects. But they aren't the right fit for everything. And it may be peanut money to you, but some folks have a different definition of that. And it will be quite a bit more expensive than a PIC and a recorder chip by a good margin. On top of that, most of the users of this kit don't want and won't use the added features that would be possible with a Prop. So for them it is just money wasted. For myself, as I indicated, I will be making a Prop powered version.

    I am also going to look at using a surface mount Prop and see what I can price it out at. That would keep the board size down. Then I would have to solder the Prop on there for them. Kinda takes a bit away from the kit aspect and then harder to replace if they mess something up.

    Thanks!

    Jonathan

  • Erco,

    Hmm. That just might be an idea, because some folks don;t even want the recorder option.

    Maybe I should explain the project a little more. Irma, an 80 year old ham, wants to be able to turn off the radio at night but still wants to be there to help if a fellow ham needs help. So, I told her I would build a "long tone zero" detector that would un-squelch her radio if it detects a long tone zero. A long tone zero is the signal of a ham in distress, transmit a DTMF tone zero for ten seconds or so.

    Anyway, I built one and it occurred to me that I could add a simple message capability. Hit the right DTMF codes and then leave a message for your pal. Talking about it to local hams, many wanted a long tone zero detector but only a few want the recording capabilities. The kit idea with a class to build them was met with enthusiasm.

    For myself, I'm going to make a cool prop version that can handle multiple messages and so on. But most don't want the features. I bet a few will when they get a load of mine though... :-)

    Anyway, the external recorder might be a good fit, allows adding it if you want it and ignore it if you don't. Think I'll order one of those to play with too.

    Thanks!

    Jonathan
  • ercoerco Posts: 20,256
    Those modules have small cheap speakers. Volume & sound quality increase if you swap to a bigger/better speaker, which is simpler than adding an amplifier.
  • pjvpjv Posts: 1,903
    Johnathan,

    I'm suspecting that a Propeller all by itself could do this very nicely. I would use the Prop's sigma delta capabilities as an A/D converter, and a Goertzel transform to detect the DTMF tones, and then use some CVSD (continuously variable slope Delta) modulator in software to do compact audio sampling and storage.

    With some tricks, I expect you could get passable audio with as little as 4 kbits per second.... and perhaps if you could add some adaptive algorithm (I have some ideas on that) with only 2 kbits per second.

    Such audio files however could only be "played" by a matching reverse algorithm unit, as they would not be standard waveform encodes. And that may not be suitable for you. But if all parties had the same thing, thin that might work.... all in a single Propeller, no extra chips.

    Would be a real fun project.

    Cheers,

    Peter (pjv)
  • Jonathon - is this the wilderness protocol discussed here?

    http://www.mcminnarc.com/wilderness.html

    I listen to a talknet every now and then and never heard anyone mention this that I can remember.

    I have heard of plenty of hams doing other experiments with Raspberry Pis, so that might be an alternate platform. You can get a USB audio interface relatively inexpensively.

    If a listener were internet connected, then it could arrange to send an email or text message upon detecting this long tone zero.
  • @erco - Good to know. With any luck the volume will be ok, but if not changing the speaker would be easier as you said.

    @peter - It does sound like fun. I tried to get DTMF decode with a Prop going a while back. If I did get it to work that does save on a decoder chip. I have no experience with audio at all really. I'll look and see what objects are available. I imagine it would be buffered in RAM and you might be able to extend that a bit by writing it to EEPROM as fast as you can. And as long as the Prop can play it back, that is all that is required. However, I am by no means sure my programming skills are up to the task and it would take me a while. The place to start would be recording some audio.

    @Kieth - Yeah, that's it. A lot of hams don't seem to know this, not sure why. It's a good way to get attention. Text or email, now there is another idea. I can send email and texts by radio already, so ya never know...

    Thanks!

    Jonathan

  • @Jonathan

    re:. I tried to get DTMF decode with a Prop going a while back. If I did get it to work that does save on a decoder chip.

    This object does tone detection.

    Go down to Phil Pilgrim (PhiPi) comment:

    " Attached is a DTMF input demo I did awhile back that uses the Goertzel algorithm to detect the tones. It includes an Excel worksheet for computing the bandwidth of the filters."

    http://forums.parallax.com/discussion/119568/propeller-dtmf
  • Bob,

    Thanks, I'll take a look at that. For myself, I'm looking at making a unit with a Prop and this radio unit:

    http://www.hobbypcb.com/index.php?option=com_virtuemart&view=productdetails&virtuemart_product_id=101&Itemid=425

    Still figuring out what to do for the kit. If I get DTMF working, then I'll try my hand a sound recording with a Prop. If THAT works out, maybe the kits can be a prop only solution.

    Thanks!
    Jonathan

  • ercoerco Posts: 20,256
    FYI, my 30-sec module arrived today but it records a full 60 seconds. Not bad for $4.51: http://www.ebay.com/itm/181337044626
  • If you want to have some fun, go get a SanDisk Sansa Clip Zip. Internal flash will be plenty, no matter what, and RockBox is an open-source firmware for it. It has plenty of CPU power to do DTMF detection, and it has an electret mic onboard. Desolder the mic, capacitively couple your audio input to the pads, and you're done with the hardware. Then, you can either write a plugin for RockBox (and have your function buried in a menu somewhere) or you can make a few changes to RockBox itself to boot directly into the function you're trying to make.
Sign In or Register to comment.