Question about 384khz sound devices

GoldWave general discussions and community help
Post Reply
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Question about 384khz sound devices

Post by Videogamer555 »

Is there any sound device that has 384khz sample rate for both input and output, and that has both the input and the output being analog (not digital) audio?

And now an explanation for why I'm hoping that such a device exists. 384khz sample rate on both input and output, and with both the input and output being analog, means I could have sinewaves with frequencies anywhere between 0 and 192khz, either on the input or output. This would be very useful for radio frequency experimentation (both transmission and reception) in the ELF (0 to 3khz), VLF (3khz to 30khz) and a limited portion of LF (30khz to 300khz, or in this case limited to a maximum of 192khz). ELF means Extremely Low Frequency, VLF means Very Low Frequency, and LF means Low Frequency. This is well below the lowest frequency used for ham radio, and in fact is well below the lowest frequency used for conventional AM radio broadcasting (or any other type of normal radio communications). However exotic and/or experimental radio transmitters are often used in this range (for example underwater radio communications), and possibly radio signals generated from natural phenomenon. So being able to transmit and receive in this range would be VERY interesting. To turn any sound card into a radio transmitter (with the maximum usable frequency being limited to half the sample rate), just connect a straight wire to the left channel of the line-out or speaker-out port. To turn any sound card into a radio receiver, just connect a straight wire to the left channel of the line-in or microphone-in port. Because maximum usable frequency is limited to one-half of the sample rate, the higher the sample rate, the better. And because radio waves are an analog phenomenon, the input must be analog if you want to receive radio signals with it, and the output must be analog if you want to transmit signals with it.
DougDbug
Posts: 2172
Joined: Wed Feb 16, 2005 3:33 pm
Location: Silicon Valley

Re: Question about 384khz sound devices

Post by DougDbug »

Google turned-up MSB Technology but it's "audiophile" stuff and an ADC & DAC is going to be very expensive!
384khz sample rate on both input and output, and with both the input and output being analog, means I could have sinewaves with frequencies anywhere between 0 and 192khz,
You're right about the Nyquist theory, but there's no guarantee that an audio DAC or ADC is going to go that far above the audio range. i.e. It's easy to built an audio amplifier that goes DC to to 1MHz or higher, but most don't... Most amplifiers filter out higher frequencies that can cause interference, and they often filter out DC to protect the speakers from DC signals.
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Re: Question about 384khz sound devices

Post by Videogamer555 »

DougDbug wrote:Google turned-up MSB Technology but it's "audiophile" stuff and an ADC & DAC is going to be very expensive!
384khz sample rate on both input and output, and with both the input and output being analog, means I could have sinewaves with frequencies anywhere between 0 and 192khz,
You're right about the Nyquist theory, but there's no guarantee that an audio DAC or ADC is going to go that far above the audio range. i.e. It's easy to built an audio amplifier that goes DC to to 1MHz or higher, but most don't... Most amplifiers filter out higher frequencies that can cause interference, and they often filter out DC to protect the speakers from DC signals.
I am not interested in DC actually, just anything that's oscillating, as ELF radio basically includes all frequencies below 3kHz except DC, VLF is all frequencies from 3kHz to 30kHz, and LF is all frequencies from 30kHz to 300kHz (though with a sample rate of 384kHz the maximum frequency possible is going to be 192kHz).

And I'm guessing that most devices that actually have a sample rate as high as 384kHz, don't do any filtering, as the whole point of having such a high sample rate is to be able to record signals that are well above the highest possible audible frequency. If you want to only capture signals that are within the audible range, having a sample rate of exactly 40kHz is all you need, as the highest audible frequency for humans is 20kHz.

Could you please point me to a device that has both a 384kHz DAC and a 384kHz ADC? The one you linked me to has only a 384kHz ADC, but not a DAC.
DougDbug
Posts: 2172
Joined: Wed Feb 16, 2005 3:33 pm
Location: Silicon Valley

Re: Question about 384khz sound devices

Post by DougDbug »

Could you please point me to a device that has both a 384kHz DAC and a 384kHz ADC? The one you linked me to has only a 384kHz ADC, but not a DAC.
Poke around the MSB website a little more, they make both. Or just Google.
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Re: Question about 384khz sound devices

Post by Videogamer555 »

DougDbug wrote:
Could you please point me to a device that has both a 384kHz DAC and a 384kHz ADC? The one you linked me to has only a 384kHz ADC, but not a DAC.
Poke around the MSB website a little more, they make both. Or just Google.
I meant I was hoping they sold one unit that did both (so I don't have to buy 2 separate units), and that both the input and output were analog, and that both the input and output had a 384kHz sample rate.
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Re: Question about 384khz sound devices

Post by Videogamer555 »

After looking at both the DACs and ADCs at the site you linked to, I found them to be absurdly expensive. I was expecting devices 384kHz DACs and ADCs to be in the several hundred dollar range. But those things cost TENS OF THOUSANDS of dollars for the most part. Their "cheap" one still cost over $6000. Can it be that hard to manufacture a device with a 384kHz clock oscillator crystal?

Well, after doing some more Google searching, I finally found this device that has 384kHz input and output analog connections, and it only costs a bit over $400, which is a much more reasonable price. http://www.amazon.com/dp/B00CYPTOM4

And here's the official webpage for this product, on the official company website from the company that makes the device http://www.brik-audio.com/style/frame/t ... &id=355692
DougDbug
Posts: 2172
Joined: Wed Feb 16, 2005 3:33 pm
Location: Silicon Valley

Re: Question about 384khz sound devices

Post by DougDbug »

But those things cost TENS OF THOUSANDS of dollars for the most part. Their "cheap" one still cost over $6000. Can it be that hard to manufacture a device with a 384kHz clock oscillator crystal?
I know... I assumed you didn't have that kind of budget.

It's not just the clock... You can't simply "overclock" your average DAC/ADC chip. But, the chips aren't expensive either. I once researched the cost of the chip used in a in a high-end DAC, and it was $5 USD. And it was one of the best audio DAC chips available...

You've got to understand the audiophile marketplace. The more expensive it is, the more desirable it is!!!! And when you're only selling a few per year, the price has to cover all of your engineering, marketing, and distribution overhead. But, don't confuse selling price with manufacturing & distribution cost. The only relationship between price and cost is that you have to cover your costs to say in business. Otherwise, price is determined by competition and what the market will bear.
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Re: Question about 384khz sound devices

Post by Videogamer555 »

DougDbug wrote:
But those things cost TENS OF THOUSANDS of dollars for the most part. Their "cheap" one still cost over $6000. Can it be that hard to manufacture a device with a 384kHz clock oscillator crystal?
I know... I assumed you didn't have that kind of budget.

It's not just the clock... You can't simply "overclock" your average DAC/ADC chip. But, the chips aren't expensive either. I once researched the cost of the chip used in a in a high-end DAC, and it was $5 USD. And it was one of the best audio DAC chips available...

You've got to understand the audiophile marketplace. The more expensive it is, the more desirable it is!!!! And when you're only selling a few per year, the price has to cover all of your engineering, marketing, and distribution overhead. But, don't confuse selling price with manufacturing & distribution cost. The only relationship between price and cost is that you have to cover your costs to say in business. Otherwise, price is determined by competition and what the market will bear.
Usually (from my understanding) an ADC chip isn't clocked. It's free running. Whatever analog signal goes in, is instantly represented in digital form on the output pins (16 pins for 16 bit ADC, or 24pins for 24bit ADC). Where the sampling clock oscillator comes in is in the binary register chip, that holds the digital data (which was output from the ADC chip) for a certain duration (the faster the clock oscillator, the shorter the duration), which must be long enough for the bits to be read off into a serial format (one wire for all the data) via a binary shift register chip. And that is where the second clock oscillator comes in, the bit clock. The bit clock must be at least 16 times the speed of the sampling clock for 16 bit data, and must be at least 24 times the speed of the sampling clock for 24 bit data. So there should actually only need to be 2 crystal oscillator chips (the sample clock and the bit clock), and 4 signal handling ICs. First signal handling IC is the ADC. Next one is the Register (also known as a "sample and hold" chip). Next one is the Shift Register. And the last one is the USB interface IC, which allows the device to connect to a PC via the USB port, and formats the sampled data into a format suitable for the USB protocol, and responds to requests for data from the PC by sending that data to the PC.

Basically the reverse of this process is done on the DAC side (the output side) as well. Again, it needs 2 crystal oscillators, and 4 signal handling ICs.

And there may also be a 3rd crystal oscillator (unless it's built into the USB interface chip itself), and that would be needed to clock the USB interface chip, which is basically a microcontroller, and therefore is basically the device's own CPU, and it must obviously be clocked to process commands and data. Depending on how many pins that microcontroller has, the device may be able to skip the parallel to serial conversion step completely, and just handle the parallel data signal directly (again, bringing the number of external oscillators down to 2, or even only 1 if it has its own internal oscillator built into the USB interface chip)
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Re: Question about 384khz sound devices

Post by Videogamer555 »

Anybody here have a chance to try out this yet? http://www.amazon.com/dp/B00CYPTOM4
What is its actual frequency response? Does it actually have the full nyquist-limited frequency range available on bout output and input?
DewDude420
Posts: 1171
Joined: Fri Mar 11, 2005 11:15 pm
Location: Washington DC Metro Area
Contact:

Re: Question about 384khz sound devices

Post by DewDude420 »

Sorry to kick up an old topic; I don't get around much anymore.
Usually (from my understanding) an ADC chip isn't clocked. It's free running. Whatever analog signal goes in, is instantly represented in digital form on the output pins (16 pins for 16 bit ADC, or 24pins for 24bit ADC).
No. There is still a limit to how fast the chip can convert the analog data to digital without suffering from error. You cannot take a chip, apply a faster clock, and get a faster sampling rate...not when it comes to audio. There is *much* more going on in the audio as far as digitizing audio.

The other aspect is that's A LOT of data to process; so you get to a point where you need an FPGA to handle it.
I am not interested in DC actually, just anything that's oscillating, as ELF radio basically includes all frequencies below 3kHz except DC, VLF is all frequencies from 3kHz to 30kHz, and LF is all frequencies from 30kHz to 300kHz (though with a sample rate of 384kHz the maximum frequency possible is going to be 192kHz).
You would be *much* better suited using an ADC designed for SDR than you would a sound card.
To turn any sound card into a radio transmitter (with the maximum usable frequency being limited to half the sample rate), just connect a straight wire to the left channel of the line-out or speaker-out port. To turn any sound card into a radio receiver, just connect a straight wire to the left channel of the line-in or microphone-in port. Because maximum usable frequency is limited to one-half of the sample rate, the higher the sample rate, the better. And because radio waves are an analog phenomenon, the input must be analog if you want to receive radio signals with it, and the output must be analog if you want to transmit signals with it.
Ok...stop. For starters, "turning any sound card in to a radio transmitter" would not be legal; it would violate several regulations for unintentional radiators. They are not type-accepted as transmitters. The rest of what you said is close...but unnecessary.

Let me give you the skinny on 384khz. IT is *not* any "standard" that I've seen/heard of. The only time this resolution is really used is in the DSD/SACD production phase; where you use a 352khz sample rate of 32-bit audio for editing/mastering/production in the PCM world; then convert that back to 1-bit DSD. There is currently nothing on the consumer market taking advantage of this high of a rate. This kind of stuff is expensive because it's 99% geared toward studios. No one captures in 384khz; if they're going to go that high they *generally* capture to a 1-bit (or 8-bit) DSD format first. In fact, the one 384khz device I got a chance to play with was internally a 6-bit delta-sigma that converted "down" to 384khz PCM...the same way any dac over 16-bit uses the "oversampling" technique of converting from PCM to 1-bit Delta-Sigma...and uses the 1-bit signal to physically convert to analog.

If you want to make a VLF receiver; doing it with a 384khz sound card is going to be super expensive and not have the results you expect. You would be better off taking a RTL-SDR dongle and modifying it for direct-sampling...hooking your antenna directly up to it's 28msps ADC (which runs at 8-bit) and using that. The problem with that is you lose a lot of efficiency when doing a direct-sampling method if you don't have the right stuff in front.

You're looking for a cheap/easy way out for what reason? Have you ever been down in the VLF range? I'll put it to you this way; all the signals you can do anything with are going to be below 100khz. Everything down there is usually highly encrypted military comms that won't do anything but sound funky. Sure, below 20khz you do get some "natural radio" hat's neat to listen to; but you don't need 384khz for that. I got a radio here that'll tune down to 30khz; I got a sound card that'll get me up to 48khz. There's nothing down there. Zip. Nada. It's fun to do; but I would not waste money on it. Take a look at the allocations below 100khz on the FCC chart. There's time signals at 20khz and 60khz; (the US only runs 60khz, but we protect 20khz for other countries)...the rest of it? It's seriously all just encrypted military comms and beacons. I've been down there. The only thing of slight interest would be the experimental amateur radio allocation at 135.7 - 135.8khz; but all you're going to hear down there are extremely low-baud digital modes and, more likely...QRSSCW (which is basically extremely slow morse code where the dit and dahs are upwards of minutes in length). The military comms? You're not going to decrypt those...they will just be shrill sounding chirping. WWVB isn't even that interesting...it's just beeps..it's actually not even beeps...it's a continous carrier that rises and drops in power output.

What you really should do is head over to a radio forum; like maybe an amateur radio forum...and ask those guys. This is an audio forum...and how an audio card performs for SDR usage is likely entirely foreign territory to most people.
Videogamer555
Posts: 43
Joined: Wed May 25, 2011 2:28 am

Re: Question about 384khz sound devices

Post by Videogamer555 »

DewDude420 wrote:Sorry to kick up an old topic; I don't get around much anymore.
Usually (from my understanding) an ADC chip isn't clocked. It's free running. Whatever analog signal goes in, is instantly represented in digital form on the output pins (16 pins for 16 bit ADC, or 24pins for 24bit ADC).
No. There is still a limit to how fast the chip can convert the analog data to digital without suffering from error. You cannot take a chip, apply a faster clock, and get a faster sampling rate...not when it comes to audio. There is *much* more going on in the audio as far as digitizing audio.

The other aspect is that's A LOT of data to process; so you get to a point where you need an FPGA to handle it.
I am not interested in DC actually, just anything that's oscillating, as ELF radio basically includes all frequencies below 3kHz except DC, VLF is all frequencies from 3kHz to 30kHz, and LF is all frequencies from 30kHz to 300kHz (though with a sample rate of 384kHz the maximum frequency possible is going to be 192kHz).
You would be *much* better suited using an ADC designed for SDR than you would a sound card.
To turn any sound card into a radio transmitter (with the maximum usable frequency being limited to half the sample rate), just connect a straight wire to the left channel of the line-out or speaker-out port. To turn any sound card into a radio receiver, just connect a straight wire to the left channel of the line-in or microphone-in port. Because maximum usable frequency is limited to one-half of the sample rate, the higher the sample rate, the better. And because radio waves are an analog phenomenon, the input must be analog if you want to receive radio signals with it, and the output must be analog if you want to transmit signals with it.
Ok...stop. For starters, "turning any sound card in to a radio transmitter" would not be legal; it would violate several regulations for unintentional radiators. They are not type-accepted as transmitters. The rest of what you said is close...but unnecessary.

Let me give you the skinny on 384khz. IT is *not* any "standard" that I've seen/heard of. The only time this resolution is really used is in the DSD/SACD production phase; where you use a 352khz sample rate of 32-bit audio for editing/mastering/production in the PCM world; then convert that back to 1-bit DSD. There is currently nothing on the consumer market taking advantage of this high of a rate. This kind of stuff is expensive because it's 99% geared toward studios. No one captures in 384khz; if they're going to go that high they *generally* capture to a 1-bit (or 8-bit) DSD format first. In fact, the one 384khz device I got a chance to play with was internally a 6-bit delta-sigma that converted "down" to 384khz PCM...the same way any dac over 16-bit uses the "oversampling" technique of converting from PCM to 1-bit Delta-Sigma...and uses the 1-bit signal to physically convert to analog.

If you want to make a VLF receiver; doing it with a 384khz sound card is going to be super expensive and not have the results you expect. You would be better off taking a RTL-SDR dongle and modifying it for direct-sampling...hooking your antenna directly up to it's 28msps ADC (which runs at 8-bit) and using that. The problem with that is you lose a lot of efficiency when doing a direct-sampling method if you don't have the right stuff in front.

You're looking for a cheap/easy way out for what reason? Have you ever been down in the VLF range? I'll put it to you this way; all the signals you can do anything with are going to be below 100khz. Everything down there is usually highly encrypted military comms that won't do anything but sound funky. Sure, below 20khz you do get some "natural radio" hat's neat to listen to; but you don't need 384khz for that. I got a radio here that'll tune down to 30khz; I got a sound card that'll get me up to 48khz. There's nothing down there. Zip. Nada. It's fun to do; but I would not waste money on it. Take a look at the allocations below 100khz on the FCC chart. There's time signals at 20khz and 60khz; (the US only runs 60khz, but we protect 20khz for other countries)...the rest of it? It's seriously all just encrypted military comms and beacons. I've been down there. The only thing of slight interest would be the experimental amateur radio allocation at 135.7 - 135.8khz; but all you're going to hear down there are extremely low-baud digital modes and, more likely...QRSSCW (which is basically extremely slow morse code where the dit and dahs are upwards of minutes in length). The military comms? You're not going to decrypt those...they will just be shrill sounding chirping. WWVB isn't even that interesting...it's just beeps..it's actually not even beeps...it's a continous carrier that rises and drops in power output.

What you really should do is head over to a radio forum; like maybe an amateur radio forum...and ask those guys. This is an audio forum...and how an audio card performs for SDR usage is likely entirely foreign territory to most people.

My setup is pretty simple for transmitting. A long wire (about 20 feet long) comes out of my soundcard's line-out or headphone-out port, and software plays a constant "tone" within the frequency range for the soundcard, at the desired frequency, for a carrier wave. The actual RF power coming out of this setup is going to be very very weak, as a 20foot antenna (though it sounds long) is only a tiny fraction of the wavelength of the radio waves emitted at this frequency. There's no way it will cause any interference with any other device, and therefore won't violate FCC rules.
DewDude420
Posts: 1171
Joined: Fri Mar 11, 2005 11:15 pm
Location: Washington DC Metro Area
Contact:

Re: Question about 384khz sound devices

Post by DewDude420 »

There's no way it will cause any interference with any other device, and therefore won't violate FCC rules.
That is a VERY simplistic interpretation of FCC rules. If you were to transmit say on 144.100mhz, with 2mw of output...if you got found out you'd face very stiff fines.

The rules cover intereference; but they also cover what you can and can't do regardless of interference.

The area below 8.3khz is not allocated; but you wouldn't get away with it (should you get caught) since they still require special authorization to run in that band.

8.3 - 9 khz are meterological aids
9 - 11.3 khz are meterological aids with radio navigation
11.3 - 11.4 are just radio navigation
11.4 - 19.95 are maritime mobile
19.95 - 20.05 are offlimits becuase standardized time signals
20.05 - 70khz are maritime mobile.

These allocations do not have a FCC Rule associated with them...which means they're part of the government allocation. It's all DoD stuff down there. If you want to take the risk...fine. But even with little power...those signals can travel. Interfere with one submarine communication and they will at some point find you.
Post Reply