Burr Brown 32 bit DAC

Onkyo TX-NR 5009, PR-SC 5509.
A 32-Bit DAC might be technically better, but if it really improves anything audibly against a 24-Bit DAC is not clear.

Burr Brown was purchased by TI some time back, so Burr Brown is TI brand name now.

There are a number of companies who make DACs. Cirrus comes to mind. Wolfson another brand. Analog Devices as memory serves.

I am sure they all make perfectly good DACs. There's different models from each manufacrturer, one of the differences being the SNR. You would need to know exact model numbers to know which exact DAC is in any given unit.

It's rather difficult to compare DAC chips directly...I have never seen any test I would call valid.

Personally, I don't worry about it.

The Integra and Onkyo 5009 have PCM1795 32/192 DAC from 2009 (not that new). A quick look at the following shows the SNR isn't the highest in 32-bit DACs and some 24-bit DACs have higher SNRs. My understanding of SNR is anything above 100dB or 110dB isn't really distinguishable so SNR isn't really that important, or the most important consideration. The 32-bit only means it accepts 32-bit input, not any implication of SNR at all. Being 32-bit is certainly not in my list of priority when picking AVR but that doesn't stop Oppo fans getting excited by the 32-bit Sabre. Other things could matter more - have a look at the data sheet.

SNR:
PCM1791 24-bit 192kHz 113dB
PCM1792A 24-bit 192kHz 132dB
PCM1793 24-bit 192kHz 113dB
PCM1794A 24-bit 192kHz 132dB
PCM1795 32-bit 192kHz 123dB
PCM1796 24-bit 192kHz 123dB
PCM1798 24-bit 192kHz 123dB
PCM5100 32-bit 384kHz 100dB
PCM5101 32-bit 384kHz 106dB
PCM5102 32-bit 384kHz 112dB

Is it true than that denon is using lowend DACs in the 4311 ?
AK5358B dac S/N: 102dB
Does this mean more hiss when the volume is turned up ?

Bump. I am wondering the same thing? What causes the difference in SQ between dacs that have similar or the same specs, such as 24/196 or 24/96? If two dacs use the same chip that is rated at 24/96, shouldnt they both sound the same? What I am wondering is, what makes a really expensive or good dac, sound different from a cheaper dac that has similar specs?

Accuracy on DACs would likely be measured as THD (total harmonic distortion.) The THD figures on DACs are well below THD levels considered to be audible.

If the distortion is inaudible, and their measured frequency response is ruler flat (as would be expected) then it's not clear why they would sound anything other than non neutral (not saying they could not have some chracteristic sound, but I would not expect that.)

I would hope by now the problems inherent in creating a properly behaving DAC were solved long ago.

The difference between DACs is vanishingly small. In a properly functioning system noise and distortion is inaudible except when one's ear is pressed to a tweeter. MiniDSPs for instance have a signal to noise ratio of just under 100db but unless there is a gain matching problem that noise is inaudible. A device with a SNR over 100db is indistinguishable from one with a 120db SNR. As someone here has said "just because it can be measured doesn't make it audible."

Slightly off topic, but now that I've connected my PC to a 2 channel receiver I've been curiously puzzled why I see so many references to external (and often expensive) DACs. My onboard sound card has DACs, my receiver has DACs. Am I looking at this wrong?

Only discussion I ever had on this with any references to measurements which made sense to me was that SNR could make a difference.

How 100 vs 110 or 120 could make a difference, I forget, because typical dynamic ranges when listening are narrower than 100 dB.

DACs maybe just for golden ears or poets,(these are the same people that write about wine and cigars) But every new issue of stereophile mag has something to say about all these new high priced black boxed dac add-ons, for the rich and soon tobe poor if they keep buying this BS.

Yes, my Essence STX sound card has interchangeable DACs and I've never been moved to try different ones. The noise floor with the stock one is below my hearing threshold.

Audible differences among DACs, unless the noise floor is very high, are more likely due to the distortion spectrum, which is related to their (differential and integral) linearity curves. You might see something in the ENOB, THD, IMD, or DNL/INL plots if you know what to look for, but by and large the average consumer does not have the background to read and understand all the techie details.

In terms of a DAC in a box, what most audiophiles mean by "DAC", the output buffers, power supply, and board design probably have more influence than the actual DAC chip itself.

All IMO, of course. - Don

I can interpret THD and IMD but the rest are foreign to me. The only things I see in most discussions of the various audio qualities of DACs has been totally subjective and soaked in snake oil.

Can you explain simply how ENOD adds to knowledge gleaned from THD and noise specs? Does it tell us something more or is it just another way to view the information on signal distortion?

I understand that converting a quantized signal to analog is non trivial - the method must not only smooth out the steps in the signal, but also not introduce any audible artifacts. But I thought I have seen scope images of post DAC output which looked rather smooth so I presumed that they do a very good job.

I also seem to recall, there's the possible clocking issues.

The effective number of bits (ENOB) includes all error sources (per the IEEE ADC Standard) so encompasses noise and all nonlinearity. It is derived from SINAD, the signal-to-noise-and-distortion ratio. Humans are generally more sensitive to IMD than THD or noise. DNL/INL are d.c. measurements that effectively determine the best the DAC will ever do. If your INL has a pattern (curve) and is more than an lsb, you'll lose a bit. Then you add noise and dynamic errors (distortion, jitter, etc.) and lose more bits...

The curves on a scope probably look smooth due to resolution and filtering. By eye, a mid-frequency 6-bit sine wave looks smooth. For that matter, most high-speed scopes only reach about 6 ENOB from their 8-bit converters (ADCs). By eye, about 1% is good enough to look perfect, and that's about the 7-bit level.

There are a myriad of issues, clocking among them...

Thanks, I will try to fully assimilate your answer

Oh, now we're bringing the Borg into this discussion?

heh

Funny, this article was repeated on NPR today.

http://www.npr.org/blogs/13.7/2012/0...e-what-we-like

"Maybe the evolutionary psychologists are right and our preferences are really complicated strategies to display wealth and win sexual partners."

Take out the "maybe" and I'd agree with this statement.

Obviously that plays a role in how we operate. But I think humans go beyond the basics of food and reproduction. We create works of art and science that I feel exceed those things.

This thread isn't about external DAC nor evolutionary biology so that is off-topic. As usual every DAC thread on AVS can't stay on track and civil and eventually turns sideways and into condescending remarks at the mild end or insults and personal attacks at the nastiest. They just can't help it somehow. If some people really think owning an external DAC is about wealth and sex they have a problem. MJH is a DAC sceptic but at least to his credit he is usually capable of engaging in a civilised conversation without an attitude towards others here.

I would just say that while there are differences between DACs, the difference is subtle. People are led to think that there are great differences. I don't think all external DACs are a waste of money as that subtle difference is of great importance to some. The Burr Brown (TI) 32 bit DACs are about as good as any IMHO.

In case the above tirade was directed at my posts, I must clarify that all DACs, whether in an AVR or an external box, do require power supplies, output buffers, and other supporting components, and are placed upon a circuit board, all of which can affect the sound. Or not -- you must decide.

Also, 6 bits looks smooth to the eye but the eye is not the ear. A perfect 6-bit DAC would have SNR around 38 dB and spur floor around 54 dB. At full scale even those numbers might not be audible, but as the signal level drops the relative SNR degrades.

Perhaps the biggest factor is the type of filtering applied, both in the digital domain and at the analog output. The relatively low sampling rate and steep high-order analog filters smeared the pulse response of early DACs (and ADCs). Oversampling has mostly fixed that problem, though a look at most any review with a pulse response will show the ringing added by digital filters before the DAC.

Finally, the shape (frequency spectrum) of the noise floor may also impact the sound. I cannot say if it really does or not (you must listen for yourself), but different DAC architectures do have different noise spectra, and some may be more objectionable (or us more sensitive to them, take your pick).

FWIWFM - Don

I apologize if what I posted came across as a tirade. The spirit behind the words was conciliatory. I just wanted to say that I didn't believe there were no differences between DACs.
Peace.