multiple sound recorder circuit or ic?

[rambinator]

hi guys!

I have to build a toy that can play sounds for a birthday present. I allready worked with one of those ready-to-use sound ics where you records something through the speaker and play it back.

what i now want is something that can record and playback like 3 or so different sounds (maybe in sequence whenever you push the button or in random order, doesnt matter). Does anyone know an ic that can do it? or a circuit i could build for that purpose? long time ago ive seen something with a uC and eeprom but i cant remember where. it should be quite small also.

thanks....

rambinator

[Jaicen]

I've designed a looper that uses an ISD 2560 sound recorder chip, and there are a couple of different designs using the same family. It's possible to record and play back different loops by addressing the chip, which can be done with logic IC's as well as simple switches. If you check out the Payback looper on Andrew coleman's site, you'll see how this is posssible. It might be overkill for your needs, but works really well.

http://geocities.com/thetonegod/

[stryd_one]

You're the tone god?

Yerh I saw that IC on smallbear and was curious but I couldn't find schems to go with it... At a glance it seems a midibox looper would be possible.... hmmmm ;)

[The Tone God]

what i now want is something that can record and playback like 3 or so different sounds (maybe in sequence whenever you push the button or in random order, doesnt matter).

If look at the Payback at the site Jaicen linked to there is an explanation of how the addressing feature works in the ISD25xx ICs. You could control the address lines with some simple logic to select the different addresses. You can even use a shift register to spin through the addresses with a fast clock to give a pseudo random result.

Jaicen's design might be better depending on your control needs. If you only need one playback per button push then you use the internal "pushbutton" mode of the ISD25xx. It makes the circuit very simple.

I've designed a looper that uses an ISD 2560 sound recorder chip, and there are a couple of different designs using the same family.

Jaicen was referring to his design which came before the Payback. His design might be better for the OP's purposes though as the Payback was designed with the limitation of only two switches for playback and record controls.

If you check out the Payback looper on Andrew coleman's site...

My last name is not Coleman. I don't want Mr. Coleman to get confused with me, well unless we are in a police line up together. ::)

You're the tone god?

No, I am. Just dropping by for this one thread.

Yerh I saw that IC on smallbear and was curious but I couldn't find schems to go with it... At a glance it seems a midibox looper would be possible.... hmmmm ;)

That IC, the ISD1020, is for the looper offered at General Guitar Gadgets. That looper was designed a long time ago when the ISD1020 was around. It is no longer produced so Smallbear keeps a stock of them for those who want to build the looper at GGG. That looper is probably the first looper that was created in the effect DIY community.

Andrew

[Jaicen]

Cheers for the clarification, good to see you over here o' god of tone!  ;D

I didn't want to imply that my design was any better (far from it) than any other, just a comment. I really designed it to replace the Mobius Trip looper designed by Dean Hazelwanter (I'm fairly sure that's the right name ;) ) simply because I didn't want to order the 1020 all the way from America when my local rat-shack carries the 2560! I use the same analogue circuit as the Mobius with a few extra features thrown in.

Using the internal pushswitch mode is very very easy with the ISD2560 especially if it's for a record once application so i'd look at that first.

[stryd_one]

Thanks for dropping by Andrew! You're one of those people who's names I already recognised when I first started reading about DIY FX, not sure how you got famous, but you are ;)

The subject of voice recorder IC's has come up around here before and lord knows how we've managed to miss the ISD IC's time and time again, especially when the synth and FX DIY community have already done it.

For those who missed those threads, in a nutshell we had two hurdles that made the designs impractical - frequency response and latency. As Andrew has already demonstrated in the PayBack looper the ISD can be kept powered up and this gives a trigger time of 300usec so latency is clearly not an issue ;)

It's a darn shame that it isn't possible to adapt (externally augment or bypass and replace) the antialiasing filters so that we could clock the device faster to get a better samplerate and therefore better frequency response... I'd love to quadruple-clock the thing and get 15 seconds of 32khz samples.

But ignoring the low frequency response, this IC could be really useful... Midibox control of the control and address lines could introduce a lot of cool features like tempo synced loops, setting sample lengths before recording by time or by beats at a given tempo, chaining a few of the IC's for more samples, etc etc. Hey with crafty use of a noisegate (to gate the input when readdressing the IC inbetween recording each slice) you could make an automated beatslicer with the thing....hmmmm

Anyway this is getting wayyyy off topic now, but I thought I'd share those ideas seeing as this IC has come to light around here. Could make a top addition to the MBFX projects :)

[rambinator]

i just ordered an isd2560 and i guess it would do with the delete eom deactivated and consecutive adressing. like that there will be the next message everytime you push the button. that will be ok. i dont need random. does anyone knows what happens when you reach the end of all messages in this mode?

i found this sample circuit for the 1400

http://www.winbond-usa.com/products/isd_products/chipcorder/applicationbriefs/apbr11.pdf

and i guess it would work with the 2560 too

[Jaicen]

You should look up the datasheet for the 25xx series, as the external logic is different.

I don't know exactly how you're planning to address it either, it's not very clear. Check out Andrew's design to see how it's possible to address two loops with one switch.

With regards to running to the end of the capacity, unless the chip reads an EOM marker, it'll keep going to the end and stop, it won't loop unless it gets an EOM.

Finally, the low sample rate is more to do with the technology itself than any restrictions imposed by the manufacturer. It takes time to store the sampled audio and this is the limiting factor, which is all in the datasheets on winbond.com.

[The Tone God]

Cheers for the clarification, good to see you over here o' god of tone!  ;D

Thanks for the warm welcome and no problem. :)

I didn't want to imply that my design was any better (far from it) than any other, just a comment. I really designed it to replace the Mobius Trip looper designed by Dean Hazelwanter (I'm fairly sure that's the right name ;) ) simply because I didn't want to order the 1020 all the way from America when my local rat-shack carries the 2560! I use the same analogue circuit as the Mobius with a few extra features thrown in.

Using the internal pushswitch mode is very very easy with the ISD2560 especially if it's for a record once application so i'd look at that first.

I know you were not trying imply anything about each of our designs so its cool. I was pointing out that the rambinator did not give any control specifications for their project so your design might actually be easier to tweak to their needs then trying to hack the Payback but they can still use things like the addressing feature from the Payback with your control design.

From what I remember Dean designed his looper a long while back when you could get the 1020 at Radio Hack. When they stopped carrying the 1020 people got concerned so Smallbear got some so people could still build the project. I think the upgrade to the 25xx series was smart move all around.

I might still do a 1420 version though as it is real simple to do the control circuitry with the same design restrictions given.

Thanks for dropping by Andrew! You're one of those people who's names I already recognised when I first started reading about DIY FX, not sure how you got famous, but you are ;)

How the hell did that happen ? I mean I understand the fame I have with the local cops but not this. To be honest I haven't been an active member in the community for too long so I am kind of baffled at that. In any case thanks for the kind words. :)

The subject of voice recorder IC's has come up around here before and lord knows how we've managed to miss the ISD IC's time and time again, especially when the synth and FX DIY community have already done it.

The ISD series can for some be hard to get. Also most of the simpler ICs are intended for driving speakers so the output interfacing can be difficult. Even more so since there is no schematic of the output section or how the power down function operates it. :::shakes fist at Winbond:::

But ignoring the low frequency response, this IC could be really useful... Midibox control of the control and address lines could introduce a lot of cool features like tempo synced loops, setting sample lengths before recording by time or by beats at a given tempo, chaining a few of the IC's for more samples, etc etc. Hey with crafty use of a noisegate (to gate the input when readdressing the IC inbetween recording each slice) you could make an automated beatslicer with the thing....hmmmm

If you were to go that far you would be better off move up to one of the other ISD series that are better intended for microcontrollers. Then alot of the features you mention would be pretty easy to implement.

Finally, the low sample rate is more to do with the technology itself than any restrictions imposed by the manufacturer. It takes time to store the sampled audio and this is the limiting factor, which is all in the datasheets on winbond.com.

Just to expand alittle more on that thought. The technology does have limitations, what they are is unknown, but the sample rate is connected to the internal clock set by the manufacture hence why the rate and cutoff is affected in the various length ICs. There is an external clock input that is intended for applications that need very accurate playback but you can only adjust it so far as the internal filters and other parts still rely on the internal clock.

Enjoy everyone.

Andrew