You're pretty much stuck to 16 or 24bit at 44,100, 48,000, 96kHz or 192kHz samples per second at n channels in the sound card realm. Some motherboard audio I/O chipsets can do this. Depends on your intended SNR, mostly.
Unless you are doing ultrasonics, why do you wish so high a sampling rate? Can you hear 120dB quantisation noise? You can get cards from measurement computing or national instruments.
That sounds like a load of Tosh. Similar to the crowd that have a Golden Ear that can hear up to 30kHz and need gold plated electrodes.
Thank you for the URL, its performance is listed at 13 bit, 1Ms/s, which is a 'bit' anemic compared to the 20+ bits, 100MS/s I'd like to get.
Very interesting information, but describes phase shift modifications to the signal, not the creation of distortion.
hasn't seemed to be recently. nor have many of the contributors to sci.electronics.design recently.
On a sunny day (Mon, 31 Mar 2014 06:41:58 -0700) it happened RobertMacy wrote in :
What do you mean by that? Any position depending phase shift will create a change of the waveform. You moving your head a bit while chewing chewing gum will completely change the waveform at your eardrums.
There is no absolute way to place your head relative to the sound source, not even with in-ear phones. Think of the sound as presure wave perhaps.
And now at 3.3 mm, that is nuts.
Cannibalize a scope? Interesting approach. Don't know their digitization level, bur I vaguely remember numbers like 8 to 12 bits of ADC, good enough for the screen. Already have data sheets on several of the 10 bit
1GSs, but simulations predict need more bits for the design to work correctly, so willing to take a hit on the speed.Pricing I've gotten is in the range of $20,000 each. That's a bit significant already.
The market [translate that to supply] right now is indeed being driven by communications requirements. parbly because a scope doesn't require many bits.
If price is no object then this one will do dual 1 Gsps @ 16 bits. eg.
I don't recall seeing any PCI board level products spec'd for anything more than 18bits and I suspect the local electrical hash of switcher noise in a PC case would make digitising more finely a waste of time.
You can get 18bit off the shelf but how good the low bits are is open to question and earth loops. This ones spec is 500Msps quad @ 18bits.
There are a few USB based devices intended to be sampling scopes that could probably be subverted to do what you asked for but probably not with the number of bits resolution at full speed you think you want.
A pair of video digitiser cards might well do it if you don't have strict requirements for linearity and relax your bit width spec.
-- Regards, Martin Brown
Glenn,
Thank you for replying with EXACT URLs! Getting close to what I need. I may have read the Hermes spec incorrectly, but I thought I saw 4 channels of 12 bit each, which unfortunately won't accomplish what's needed. BUT! I like Open Source! and, the fact that people are actually building, circa
2011, instrumentation getting near to what's needed. So, there's hope to find an existing solution.
don't know what the "rtl sdr world" is.
Love that phrase, "...sounds like a load of..." with the key word, 'sounds'. As opposed to, "I calculated" or "I measured" let's all harken back to the times of those lengthy arguments over the origin of the Nile.
There's a tradeoff between number of bits and sampling rate. Is your signal really worth 20 bits at 100MS/s? I don't believe it. Show us why you think this is justified.
Just to set the scene, say we have a 2V full scale 20-bit ADC, so one LSB is 2uV. A 100MS/s ADC might have a bandwidth of 50MHz, although it's usually more. The Johnson noise of a 50 Ohm resistor in 50MHz is 6.4uVrms. Even a good ADC has a noise figure in the
25dB ballpark. You are asking far too much.Your turn.
Jeroen Belleman
What I mean is: Distortion is DIFFERENT than phase shifting.
It is my understanding that the ear is NOT very sensitive to phase shifts. Which makes sense, since nature provides ample opportunity for phase shift. However the ear *IS* very sensitive to frequency components, especially to those that do NOT belong to what is perceived as the 'original' voice. Distortion creates 'extra' frequency components, and the ear WILL hear those. And the mind knows what is supposed to be there, discovers 'extra', or weird, components, and discerns ...that doesn't sound quite right.
Any idea where to get a 16, 18, or 20bit 200MSs ADC system? without making it myself.
On a sunny day (Mon, 31 Mar 2014 07:45:40 -0700) it happened RobertMacy wrote in :
Yes, absolute phase shifting. But my youtube demo shows how a specific frequency will get ATTENUATED depending on position of observer, source, refecting object.
Fourier tells us that if you change the amplitide of one or more frequency components in a wave form, then the wave form changes, so _distortion_. * There is clearly audible difference between a signal with 1 kHz and 3.1 kHz tones, and one with 1 kHz only *, if by some reflection or whatever the 3.1 kHz was severely attenuated.
Right, for reason I mentioned it cannot be,
Nobody hears to 1.5 MHz or 1 MHz you claimed. And, as I pointed out, there is no transducer to deliver that spectrum,
Well, if you fork out enough I will build one. I still want to buy that ship.. sail around the world.
On a sunny day (Mon, 31 Mar 2014 07:35:46 -0700) it happened RobertMacy wrote in :
Maybe you should measure something yerself mate.
This sounds like audiophool stuff founded in random mysticism.
The ears are sensitive to two separate types of information.
Frequency domain from the resonant hairs in the cochlea which allows determination of pitch and amplitude (but is easily fooled) and time domain based on the transient response of impulses arriving at the ears.
A combination allows trained musicians to beat naive sampling theory in terms of recognising a note from a few cycles of short tone burst.
The phase/time delay allows you to place a sound around you with surprising accuracy (helped by the asymmetry of the ear). Binaural recordings used to exploit these features to give pseudo surround sound.
The ever rising tone and similar illusions show just how easy it is to trick the ear into "hearing" a lower frequency by playing the right combination of higher frequency harmonics.
18bit is available off the shelf but I am not convinced that inside a PC the least significant bits will be anything other than noise...-- Regards, Martin Brown
RTL2832U & R820T chipset Software Defined Radio. Nominally a TV/DAB radio tuner subverted by custom drivers.
-- Regards, Martin Brown
Thank you for your reply with these URLs. I signed up, and received data sheets. There is also a forum, which may prove useful, too.
So far this is the closest. Unfortunately, price is ALWAYS an object.
You mention noise inside a PC case. Surprisingly, Creative Labs EMU1212 and EMU0404, which plug INSIDE an extremely noisy PC case have very clean 2 channels of 24 bits at 192kS/s. Very close match, very repeatable. Very quiet, too.
The problem I'm running into is that most buyers/customers use ADC as real-time system where 'each' sample has some value. As you increase sampling rate, the white noise contribution comes up which starts to obliterate your sample. Thus, most suppliers perceive no benefit from having more bits at those high rates. You'll see that effect if you plot quantization noise with layers of that pesky square root of sampling rate times noise floor. Just comes up and swamps your sample's value.
So why don't you tell us a bit about this application and your arrival at this estimated bandwidth?
You're asking for a source noise temperature below 125 K (at 50 ohms). Let alone what the the preamp (if any), sampler and comparators can contribute.
There are few things that need that...and there's a good reason why...
Tim
-- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
Am not allowed to explain why the bits are justified. Suffice to say, they are justified. Your premise and what you say/present is correct. Just does not apply to this applicaiton.
Your presentation is EXACTLY why it is difficult to find what I need, because the suppliers are thinking the same way.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.