What he heck is burr-brown and what is 1-bit D/A converting??

[MrMatt]

I'm looking at new cd players, and the ones I'm looking at for my car have something called "1-bit D/A Converter:24-bit Burr-Brown" wtf is this?

 

[ElFenix]

http://en.wikipedia.org/wiki/Burr-Brown_Corporation

http://electronics.howstuffworks.com/question620.htm

 

[nutxo]

Burr browns = better sq. So I hear anyways.

 

[Leros]

A key part of the converter is a circuit called a delta-sigma modulator, which takes the binary signal (1s and 0s) from the CD and changes them to a steady pulse, called a pulse train. The pulse train contains an average of the change in the amount of energy represented in the sample. A low-pass filter removes all time-domain information and recovers only the average energy of the pulse train that feeds it.

Am I correct in understanding the a 1 bit D/A outputs a pulse width with a DC value equal to the desired analog value?

 

[MrMatt]

does anyone actually notice a difference with this?

 

[crashtestdummy]

does anyone actually notice a difference with this?

You might, in a quiet room with an excellent sound system. I have an external DAC on my laptop because the one it came with isn't great. I wouldn't bother for a car, though. The road noise alone will probably eliminate any difference you'd hear.

 

[krylon]

You might, in a quiet room with an excellent sound system. I have an external DAC on my laptop because the one it came with isn't great. I wouldn't bother for a car, though. The road noise alone will probably eliminate any difference you'd hear.

Not everyone drives a dump truck.

 

[evident]

Not everyone drives a dump truck.

most cars, even a lexus /mercedes, will have some type of ambient noise level that will eliminate any advantage of super duper HQ sound components unless if you're dead still....

with that said, it would probably make an effect on OP's choice if he had a decent sound system in his car, as in upgraded speakers running through an amp connected to the head unit.

 

[TecHNooB]

Am I correct in understanding the a 1 bit D/A outputs a pulse width with a DC value equal to the desired analog value?

yes. the filter is what changes the pulse widths into the waveform you expect.

the filter can be as simple as a resistor + capacitor in parallel.

the capacitor acts like a sliding window averager. key point is that it's a windowed average.

 

[MrMatt]

most cars, even a lexus /mercedes, will have some type of ambient noise level that will eliminate any advantage of super duper HQ sound components unless if you're dead still....

with that said, it would probably make an effect on OP's choice if he had a decent sound system in his car, as in upgraded speakers running through an amp connected to the head unit.

which I do; Infinity Kappas, and infinity reference series amps

 

[BTA]

Infinity gear can hardly be considered great, or even good (there are worse choices to be had, for sure)...however, any improvement in the signal chain can have benefits.

Personally I think you are wasting your money if you spend more specifically for burr brown DAC's with a basic setup like that. If the HU you are looking at has a bunch more features that you like as well as the BB DAC's, then by all means, it certainly wont hurt.

Put your money into processing capabilities (good EQ, xovers, T/A, etc) before you worry about what DAC's the unit has.

 

[TuxDave]

yes. the filter is what changes the pulse widths into the waveform you expect.

the filter can be as simple as a resistor + capacitor in parallel.

the capacitor acts like a sliding window averager. key point is that it's a windowed average.

Not really. He said does the pulse train vary in DC voltage based on the desired analog output. The sigma delta modulator outputs a pulse train of equal DC voltage. It's better to picture the output of the modulator as an "up" and "down" signal where if you get a pulse it means to integrate the value incrementally up and if you don't get a pulse it means to integrate the value incrementally down. As a result, the sigma delta modulator must sample data at rates much much higher than the nyquist frequency to work correctly.

 

[Leros]

I found a nice simple explanation of a 1 bit DAC
http://www.cs.tut.fi/sgn/arg/rosti/1-bit/

 

[Mark R]

Am I correct in understanding the a 1 bit D/A outputs a pulse width with a DC value equal to the desired analog value?

Essentially, that's right. You then place a crude analog filter on the output, and voila - you've got a nice analog output.

In practice a '1 bit DAC' (also called 'bitstream DAC' or 'multi-stage acoustic noise-shaping [MASH]') is a Delta-sigma modultator. This uses clever digital logic to create a high-frequency pulse-width modulated output from a digital input. (The naive method. Switch on output. Count from 0 to x. Switch off output. Count to max. repeat. Is impractical).

The advantages of the DS modulator approach are:
It's cheap - High-precision DACs must be built with tolerances equal to their precision - so a 16 bit DAC needs to be built with resistors accurate to within 0.001%. This means special manufacturing techniques - like laser trimming, and special design features like on-die sapphire heat-spreaders to ensure that all the DAC resistors are at the same temperature. The crude '1 bit DAC' needs no expensive factory adjustments, and the digital algorithm ensures absolute linearity and virtually zero distortion.
It 'spreads the quantization noise'. Any DAC/ADC introduces noise due to the 'steppiness' of the output - this noise is spread evenly over the frequency band of the DAC (e.g. 48 kHz). In a DSM (which may operate at many MHz), however, the digital processing 'shapes' the noise so that it mainly falls in the high frequencies, which are in the MHz range and totally inaudible, and leaves the audio range ultra pure. The result, is that a DSM DAC introduces virtually no noise in the audio range - a unique property of the DSM DAC.

The advantages of a DSM DACs/ADCs for audio are overwhelming, and they are now virtually the only audio technology available. However, many manufacturers have now dropped the '1 bit' design, and may use 2-4 bit DACs in their DSMs, as these higher performance components dramatically improve the performance of the DSM as a whole.

 

[DisgruntledVirus]

Infinity gear can hardly be considered great, or even good (there are worse choices to be had, for sure)...however, any improvement in the signal chain can have benefits.

Personally I think you are wasting your money if you spend more specifically for burr brown DAC's with a basic setup like that. If the HU you are looking at has a bunch more features that you like as well as the BB DAC's, then by all means, it certainly wont hurt.

Put your money into processing capabilities (good EQ, xovers, T/A, etc) before you worry about what DAC's the unit has.

This.

You'll notice the DAC if the rest of the downstream equipment is a high enough quality. Unfortunately, Infinity is not it. You would need to step up into Focal K2 series or something like that before maybe noticing the difference in a Burr Brown DAC vs other.

 

[TuxDave]

I found a nice simple explanation of a 1 bit DAC
http://www.cs.tut.fi/sgn/arg/rosti/1-bit/

Yup and the red lines is the output of the pulse trains where the peak of the pulse doesn't vary in amplitude.

Since we seem to be nerding it up, if you pay attention to the peaks of the sine wave where it's not really moving (and acting like a DC signal), a problem with a sigma delta modulator is that you'll start getting 0101010101 sequences as it tries to keep the signal stable. This causes a large tone that needs to get removed via dithering or other solutions where instead of a 01010101 stream, you get something like 001011000111 or something like that so you don't end up with a single large frequency signal at the end.

 

[Leros]

Yup and the red lines is the output of the pulse trains where the peak of the pulse doesn't vary in amplitude.

Since we seem to be nerding it up, if you pay attention to the peaks of the sine wave where it's not really moving (and acting like a DC signal), a problem with a sigma delta modulator is that you'll start getting 0101010101 sequences as it tries to keep the signal stable. This causes a large tone that needs to get removed via dithering or other solutions where instead of a 01010101 stream, you get something like 001011000111 or something like that so you don't end up with a single large frequency signal at the end.

You mean a high frequency tone not a large tone?

 

[TuxDave]

You mean a high frequency tone not a large tone?

It's a high frequency large tone.

 

[Leros]

It's a high frequency large tone.

I've never heard "large tone" before.

 

[TuxDave]

I've never heard "large tone" before.

How about a narrow high frequency high energy signal?

 

[TecHNooB]

Not really. He said does the pulse train vary in DC voltage based on the desired analog output. The sigma delta modulator outputs a pulse train of equal DC voltage. It's better to picture the output of the modulator as an "up" and "down" signal where if you get a pulse it means to integrate the value incrementally up and if you don't get a pulse it means to integrate the value incrementally down. As a result, the sigma delta modulator must sample data at rates much much higher than the nyquist frequency to work correctly.

Well, to be more precise, it's the integral of the current, not voltage.