[pibroch]

On the Dedicated Source Components forum someone is claiming that in a D/A converter the character of the sound is mainly due to the particular model of dac chip used. They say this is because the chip is the component that is actually rendering the waveform.

They agreed that (in their words) “the implementation, including filtering and analog output stage also have a significant effect on the sound.”

I had argued that “different chips enable different design options but a dac chip alone has no sound character as such”.

Would someone be able to provide a link to a good article which gives insight into these matters. In particular I would like to know how far off base I was about the sound character of the chip itself.

 

[NA Blur]

Actually a lot of the differences come from the filter system used, not the DAC chip itself which is an IC / resistor system.

 

[bigshot]

Most DACs are designed to be audibly transparent and exceed the range of human hearing. The only ones I know of that aren't are from the early 90s before they had oversampling.

 

[colonelkernel8]

Most DACs are designed to be audibly transparent and exceed the range of human hearing. The only ones I know of that aren't are from the early 90s before they had oversampling.

Well, most audio DACs anyway. There’s instrumentation DACs designed for measurements and other applications that may have considerably less bit depth (we’re talking like 8 or 10 or 12 bits) but sample at hundreds of megahertz and up. Oscilloscopes are an example of this application, but they integrate the DAC into an ASIC that’s coupled with buffers and voltage references and tons of other things because if they weren’t all on the same chip, they couldn’t achieve the necessary bandwidth. A digression. To answer the question, you can check out Wikipedia which has a great simple description of a DAC. Keep burrowing down through Wikipedia with terms like delta-sigma modulation or R-2R to get a greater idea of how many DACs work.

 

[colonelkernel8]

Actually a lot of the differences come from the filter system used, not the DAC chip itself which is an IC / resistor system.

Not always resistors. Some DACs output current and are essentially sets of cascaded transistors. This document has some great information on the history and architecture of DACs: http://www.analog.com/media/en/training-seminars/tutorials/MT-015.pdf

 

[bigshot]

Well, most audio DACs anyway. There’s instrumentation DACs designed for measurements and other applications that may have considerably less bit depth.

You can safely assume that whatever I say is referring to consumer audio intended for listening to music in the home.

 

[colonelkernel8]

You can safely assume that whatever I say is referring to consumer audio intended for listening to music in the home.

I know friend. Just playing

 

[bigshot]

Look around you though. There are people who are predisposed to think that. Your joke is taken seriously. I care more about the newbies than I do the experts. I'm not here to impress people (unlike some). I'm here to help people with things I know about clearly and with a practical application.

 

[ev13wt]

 

[old tech]

Most DACs are designed to be audibly transparent and exceed the range of human hearing. The only ones I know of that aren't are from the early 90s before they had oversampling.

Are you sure it was the early 90s? My first CD player, a 1985 Pioneer Elite, had oversampling. Sounded nice too.

 

[gregorio]

[1] Would someone be able to provide a link to a good article which gives insight into these matters. In particular
[2] I would like to know how far off base I was about the sound character of the chip itself.

1. Non-linearities on DAC chips, even cheap $2 ones are well below audibility and no one in the scientific community is going to waste time and money to demonstrate that inaudible non-linearities cannot be heard! You might try looking up the manufacturers specs for the chips and then cross referencing those specs to audibility levels. Not sure that will get you very far with most audiophiles though, because they don't believe in measurements, or believe their ears are better than measurements anyway. If someone comes up with reliable evidence that differences between chips can be heard, then the scientific community would investigate and of course the audiophile press would go nuts and have a massive party but there's never been any such reliable evidence I've ever heard of.

All you can do is ask that "someone" making the claim of audible DAC chip differences to back it up with some reliable evidence. But then you'll get into an argument about what "reliable evidence" means and you'll find yourself blocked or banned soon after!

2. What you've stated is a safe bet, safe from everyone except extremist audiophiles that is!

G

 

[JaeYoon]

That's the idea that keeps being spread around on head-fi.

Audiophile 1: "The burr brown dac chip has a nice warn around and rolled off treble"

Audiophile 2: "it depends on the implementation and the high quality parts used to make the sound quality better"

Audiophile 3: "the Sabre dac chip has a glare and sharp treble and makes me fatigued. I'm treble sensitive".

Audiophile 2: "there is a 1000 dollar player that uses both that dac chip and i didn't notice any glare. It's definitely the high quality parts used to make the sound"

I see this way too often in the source forum. I think imagination has more to do why people "hear" these huge differences between D/A converters.

If someone were to take two cheap $1-2 modern dac and make a diy board and let a sample of headfi users listen.

Put a sign that says ESS Sabre on it. Other one a burr brown. I guarantee they will start saying stuff like "the burr brown is very warm. It's very musical while the sabre has high treble".

People say stuff and spread it on head-fi and is repeated by many people.