AD1865 vs AD1955

[5th element]

Originally Posted by SoNic_real_one

Exactly! The DAC doesn't know when the music is about to start, but the mixing engineer knows. He raises the faders up right before that and we see the analog noise.
The 2 seconds "before" noise modulation is not happening in the DAC, is in the mixing console. DAC cannot hold that much data to anticipate for 2 seconds.

I think all the S-D DAC's do that - it is called "zero detect mute".

Well if the outputs are muted during zeros being sent to it, then quite clearly if abraxalito is hearing noise as soon as the DAC starts being sent data, then the data contains more then just zeros.

[ljm_ljm]

In my opinion the speakers than the amplifier, the amplifier is greater than the player.

I rarely hear the difference between the DAC and the DAC to bring our biggest different data.

I like to see the test. Like 120DB SNR of digital .. However, I just read and then continue to use the integrated sound card of my computer.

[chris719]

Originally Posted by ljm_ljm

In my opinion the speakers than the amplifier, the amplifier is greater than the player.

I rarely hear the difference between the DAC and the DAC to bring our biggest different data.

I like to see the test. Like 120DB SNR of digital .. However, I just read and then continue to use the integrated sound card of my computer.

You know, I tend to agree with you. The differences between speakers / headphones are massive compared to well implemented electronics.

[abraxalito]

Originally Posted by chris719

The differences between speakers / headphones are massive compared to well implemented electronics.

I agree with you too - poorly implemented electronics abounds.

[regal]

For those that don't consider a DAC an important part of the audio chain, you have to have a well put together system and room to understand .

In audio we are so close minded we think noise ends at the end of the FFT ~22khz or maybe 192khz. Yes NOS AD1865 will have more noise here. Noise is engergy and the first law of thermodynamics means it has to go somewhere. We all know that the AD1955 produces more noise if we have even a basic understanding of a modulating DAC so where is it?.

Take a look at the noise floor of an AD1955 up into the ghz, thats where the proplem is. All this energy isn't easily absorbed by ground or miller capacitance, it induces quick oscillations with our well designed subsequent analog stages. This is how noise shaping works, moves it from one region to another. And in the case of S-D DAC technology it just plain isn't working, the industry needs a reset.

Its time for folks to stop defending this cheap technology, I know thats hard for builders to do because they have a vested interest in designs that can't be based on obsolete or planned obsolete chips.

Thanks abraxalito for showing yet another flaw with the s-d concept, however I don't think its the bigest issue it certainly is another paper cut in a technology that is bleeding to death.

[SoNic_real_one]

Unfortunatelly... this is just speculation, this cheap thechnology works.
The actual DAC's are a combination of multibit and S-D (4-10 bits), using the filters specified by manufacturer after a S-D DAC keeps the noise due to shaping below -110..120dB in audio band, up to 40-50kHz. You don't have "GHz bandwidth" with those devices, gosh, it would be amazing if you manage to do that with those chips.
The multibit DAC's have more distortions at high level, and have other sources of noise at low level too.
Where is that noise that you are talking about in thise graph (CS4392):



Noise-shaping is not the problem in S-D concept.
Before somebody twists my logic again, let me clarify: I don't say there is no problem with S-D concept, just that noise-shaping is not the problem.

[regal]

Originally Posted by SoNic_real_one

Where is that noise that you are talking about in thise graph (CS4392):

Noise-shaping is not the problem in S-D concept.
Before somebody twists my logic again, let me clarify: I don't say there is no problem with S-D concept, just that noise-shaping is not the problem.

Yes I exaggerated but of course there is no "noise" in that plot, you aren't going to measure the noise ( interference) from a S-D dac with a ADC. An old Multibit DAC has a much lower bandwidth than a modern modulator DAC, how else would the modulators get 32 bits out of 6? (shifting the measured band of interest.) What I am saying is the things are akin to little radio transmitters only radiating 'inaudible" digital hash through the audio chain. Would you hard wire your wireless router's transmitter to your preamp ? No because the active devices would have fits (oscillations), fits that are too fast to measure with an averaging FFT from an ADC.

[SoNic_real_one]

The transistors inside the audio devices don't go that high (GHz range), they are way slower than LDMOS/Ga-As ones, have limited slew-rate. Are you building your amplifier using 2.4GHz router transmitters?

[abraxalito]

Originally Posted by SoNic_real_one

Where is that noise that you are talking about in thise graph (CS4392):

Where are the vehicles in this picture?

Noise-shaping is not the problem in S-D concept.
Before somebody twists my logic again, let me clarify: I don't say there is no problem with S-D concept, just that noise-shaping is not the problem.

How do you know it?

[SoNic_real_one]

Was the picture taken with the same exposure speed as the ear response? Then... there are no cars.
You use the argument with "the ear integrates the signal, there is no need for filters" when is in your interest (alias images at 22kHz). But when is about noise you say "woohoo, look at that noise in the MHz band!".
Maybe you can make up your mind.