Colleagues, please tell me!
I want to select a not very complex chip, which is a USB interface with an I2S output bus and parameters no worse than 32bit 384khz.
I found one or two, but they are in very uncomfortable packages and with built-in microprocessors! (The processor, then why?)
Are there simpler and more convenient solutions?
Thankful in advance,
Ogogon.
You can get devices without a built in microcontroller (cp2124). What package did you want? USB is a high frequency, impedance controlled bus, so parts are probably going to be surface mount, small packages as appropriate for the frequency. If that's a problem, you could buy an evaluation board and use that.
32-bits* and 384kHz are both "exotic" and rare. It's only for "audiophiles". (And, higher sample rates are useful for non-audio applications.) You're not going to find much of a selection.
Pro studios generally record at 24-bits/96kHz which is pretty-much overkill, although 24-bits does allow you to leave extra headroom while recording. (Of course, that headroom is removed in the final production.)
The rumor is that most 24-bit audio DACs & ADCs are only accurate to about 20-bits and I'd assume that's about the real-world limit.
24-bits gives you 144dB of dynamic range. That's more than you can practically use... On they loud-side you're above the threshold of pain and hearing damage (if your system can go that loud) and on the quiet-side you're below the environmental background noise (or below the threshold of hearing).
32-bits gives you 192dB of dynamic range which is of no practical use. There's no issue on the "digital side" but you'll never get that kind of range on the analog (or acoustic) side.
...And. there's more to making a good audio interface than simply "wiring-up" the chip. The chip specifications is just one of the things that affect audio quality/performance.
* There are practical reasons for processing in 32-bit (or 64-bit) floating-point (Audacity uses 32-bit floating point "internally).
P.S.
I'm not saying you shouldn't experiment with this stuff but you should understand why it's hard to find.
The LTC2500C-32 (https://www.digikey.com/product-detail/en/linear-technology-analog-devices/LTC2500CDKD-32-PBF/LTC2500CDKD-32-PBF-ND/6670442) almost fits the bill, except that it uses an SPI interface. It's 32 bit with 104dB SNR at 1 megasamples per second. Oh, and it costs $56 a chip when bought individually. :))
Now, if you're willing to throw away some bits, and don't mind dealing with a more complex LVDS inteface, the ADC12D2000RF (https://www.digikey.com/product-detail/en/texas-instruments/ADC12D2000RFIUT-NOPB/296-43535-5-ND/4251096) will get you 2 gigasamples per second on two channels with 12 bit resolution. Only $4400 a chip, minimum order 40. :D
Maybe a very stupid answer, but I'll try anyway.
I have a Gustard U12, USB to SPDIF interface. It also has a I2s output via HDMI and supports 32 bit 384KHz.
Isn't that exactly what you are looking for? These units are not that expensive and if needed you can even take out the circuit board (or parts) you really need.
Perfectly suitable for vinyl captures.
Aloha from Kaua'i, bitches!
Unless your doing outside recording like nature sounds. There's no reason to record media using that level.
Ah okay, you want to listen to bats, then I understand the need for equipment which can record way above 100 kHz.
24-bits gives you 144dB of dynamic range. That's more than you can practically use... On they loud-side you're above the threshold of pain and hearing damage (if your system can go that loud) and on the quiet-side you're below the environmental background noise (or below the threshold of hearing).
32-bits gives you 192dB of dynamic range which is of no practical use. There's no issue on the "digital side" but you'll never get that kind of range on the analog (or acoustic) side.
Doug, I believe using dB units while talking about dynamic range can be very confusing as most people associate dB with loudness. In reality, it's just a unit that measures a ratio between two values.
Let me explain this in more details...
24 bits gives you 2^24=16,777,216 values to represent your analog waveform (that has values from Vmin to Vmax in Volts).
The dynamic range is the ratio between the largest and the smallest digital values and in this case is equal to 16,777,215/1 or you can represent it in dB units using the following formula: 20*log(16,777,215/1)=144 dB. That is just a ratio and it has nothing to do with the sound that has the loudness of 144 dB!
Whether you have 16 bits or 24 bits or 32 bits, you still represent the same analog waveform that goes from Vmin to Vmax in Volts. So using different bit depth won't make your sound louder - it just means you have more resolution for the amplitude of the same standard volume range.
I hope that helps!
24-bits gives you 144dB of dynamic range. That's more than you can practically use... On they loud-side you're above the threshold of pain and hearing damage (if your system can go that loud) and on the quiet-side you're below the environmental background noise (or below the threshold of hearing).
32-bits gives you 192dB of dynamic range which is of no practical use. There's no issue on the "digital side" but you'll never get that kind of range on the analog (or acoustic) side.
Doug, I believe using dB units while talking about dynamic range can be very confusing as most people associate dB with loudness.
I think if someone is confused by units like dB, they're probably not going to be reading this old thread you bumped about DACs.