I just need to rant about this, 'cause nobody else seems to be doing it, and will understand if this gets deleted.
But, here goes:
WHY do government regulators continue to allow ESS, Burr-Brown, Asahi & etc. to sell "32-bit dac" chips, when this is PLAINLY OUTRIGHT FRAUD?
Just 'cause you cram a 32bit(allegedly) digital filter into the same chip package as an optimistically 24bit dac, HOW does that give anyone the proper right to call it a "32-bit DAC" chip?
How is that not fraud?
The deceiptful chip makers call them 32bit dacs, then the dac/player makers rely on that fraud to call their units 32bit, and the public buys them thinking they're getting 32bit performance, when they are damn lucky if they're getting even 22bit of dynamic range & snr.
How is that not fraud?
Why is there nobody holding the chip & gear makers accountable?
But, here goes:
WHY do government regulators continue to allow ESS, Burr-Brown, Asahi & etc. to sell "32-bit dac" chips, when this is PLAINLY OUTRIGHT FRAUD?
Just 'cause you cram a 32bit(allegedly) digital filter into the same chip package as an optimistically 24bit dac, HOW does that give anyone the proper right to call it a "32-bit DAC" chip?
How is that not fraud?
The deceiptful chip makers call them 32bit dacs, then the dac/player makers rely on that fraud to call their units 32bit, and the public buys them thinking they're getting 32bit performance, when they are damn lucky if they're getting even 22bit of dynamic range & snr.
How is that not fraud?
Why is there nobody holding the chip & gear makers accountable?
The ESS DAC chips accept 32 bit data and process that data as 32 bit. They include a 32 bit digital volume control too so you can attenuate the signal digitally without losing resolution.
Some DAC chips nowadays also integrate a DSP core within them and if these are operating at 32 bits then that also increases the overall precision of the calculations that they will perform.
Sounds like 32 bit is a reasonable declaration to make about said parts if the internal data processing is handled on a 32 bit level. Just because the analogue side of things is real world limited to 21-22bits doesn't mean that it's fraudulent to claim a DAC is 32 bit when a significant part of its internal architecture is working at 32 bits.
Some DAC chips nowadays also integrate a DSP core within them and if these are operating at 32 bits then that also increases the overall precision of the calculations that they will perform.
Sounds like 32 bit is a reasonable declaration to make about said parts if the internal data processing is handled on a 32 bit level. Just because the analogue side of things is real world limited to 21-22bits doesn't mean that it's fraudulent to claim a DAC is 32 bit when a significant part of its internal architecture is working at 32 bits.
Whilst I agree there are no 32bit DAC chips and the marketing is at best deceptive, I'm unclear how the number of bits a DAC is described to have is determined in cases such as these. With ladder DACs its fairly easy to see how many bits - just count the number of switches driven from the input data word, with S-D types its not at all easy.
It wouldn't be the first time something like this happened.
https://en.wikipedia.org/wiki/Audigy#First_generation
https://en.wikipedia.org/wiki/Audigy#First_generation
I don't see how you're making anything remotely like a justifiable argument. If you claim to make a 32bit dac, the only reasonable assumption a consumer can make, with expert knowledge, is that they will hear 32bit dynamic range. NONE of these fraudulent chips exceed 24bits of actual d/a conversion performance. The only way a consumer could discover that they have been misled is to look up each chip maker's datasheets, then look up on some esoteric source what actual 32bit dynamic range & s/n ratio specs would be, then compare. That's rather an unreasonable expectation.
Just because you can allegedly perform digital filter operations at 32bit level, if the actual dac portion of the chip is utterly incapable of outputting anywhere near 32bit, by what rationale is that a 32bit dac?
NO, it is NOT a reasonable declaration. If the 32bit digital filter was a separate chip in a dac unit, and the dac chip itself was only capable of 24bit, would it be anything but fraud for the unit maker to claim it's a 32bit dac unit? It would be reasonable to call it a 24bit dac that does digital signal processing at 32bit, but it sure as heck would NOT be reasonable or justifiable, or anything short of fraud to just outright declare it to be a 32bit dac unit. That dac unit maker would be universally ridiculed & condemned as a dishonest & deceiptful manufacturer. Major chip makers do EXACTLY the same thing, and nobody has a problem with it?
Last edited:
An s-d dac chip's bit depth capability, while not as clear cut as is the case with ladder dacs, is very directly connected to it's dynamic range & signal-to-noise performance. Before some dishonest chip maker(was ESS the first one?) decided they could get away with jamming their 32bit digital filter into the same chip as their optimistically-24bit dac, no chip makers were daring to call any dac chip more than 24bit capable, and the vast majority of those didn't make much better than 22bit actual performance. But once the first chip maker *got away with it*, the floodgates opened & every maker is jumping on the fraudulent bandwagon.
Since the dac/player makers must be assumed to have expert knowledge, I view EVERY one of them using these chips & calling their products 32bit as every bit as guilty of fraud as the chip makers.
To their great shame, Burr-Brown, hopefully due to orders from above from T.I., didn't even put their best effort into their "32bit dac". Rather than base it on their best, PCM1794, they cheaped out & just upped the onchip digital filter in the PCM1798 to 32bit to foist the PCM1795 on the world. This to me is pretty convincing proof that the chipmakers know darn well that they're getting away with fraud, being transparently obvious as "jumping on the bandwagon" practices.
Last edited:
In this case, yes, it can be argued as fraud, but at least, in this case, Burr-Brown was quite honestly attempting to make a 24bit chip, with actual 24bit circuit design. But I would also argue, personally, that one can get at least very near 24bit performance from the PCM1704 with the right design approach. B-B's test circuit for measurements is not an adequate approach, IMO.
This whole 32bit chip thing is just blatantly deceptive on it's face.
Last edited:
I do. Seems quite reasonable to me. If the DAC can accept and process at 32 bits, and attempt a 32 bit output, it's 32 bits. All be it with a high noise floor.
Go have a look at 24 speed bicycles to get an idea of misleading specs.
In response to both of the above responses, yes, it is about intention. The chip makers are DELIBERATELY deceiving the public, via gear makers they know are happy to collude with them, by making dac chips that they positively know they cannot legitimately claim are 32bit or remotely close.
If any of these chip makers had a dac chip that was capable of beyond 24bits, or could even be argued as such, then they would have produced it as a dac chip that is just a dac chip. It is convincing evidence, in my view, of deliberate fraud that all makers of these "32bit dac" chips have made nothing but combined digital filter + dac chips(+input receiver in ESS's case), as a deliberately deceptive marketing strategy.
If any of these chip makers had a dac chip that was capable of beyond 24bits, or could even be argued as such, then they would have produced it as a dac chip that is just a dac chip. It is convincing evidence, in my view, of deliberate fraud that all makers of these "32bit dac" chips have made nothing but combined digital filter + dac chips(+input receiver in ESS's case), as a deliberately deceptive marketing strategy.
Or look at the speedometer reading of a scooter. The number is based on the rotation of the wheel. Roll the scooter downhill and the maximum "speed" may reach 300 km/h.
There is nothing wrong with that. It has been naturally like that since the first day marketing is invented. Uneducated customers should improve their knowledge instead of complaining
In this and some of your other postings I think you are accusing the industry based on an unreasonable assumption of what a 32 bit DAC would be used for.
Lets say a 24 bit DAC has full dynamic range. That means, if you have a 0dB noise floor in your listening room (not likely) your systems capacity should be 144dB, and you should actually be using it. So lets take a more realistic scenario. You have a noise floor of 20dB (still pretty good) and your system has a 130dB SPL capacity. This gives you a maximum dynamic range of 110dB and your noise floor is inaudible. Then, if you turn your system up to 130dB, you will be far from able to hear anything at 40dB. This means you will never be able to actually hear anything that exceeds 90dB of dynamic range.
So that leaves us with another question. If you are a manufacturer, and you make a 32 bit DAC chip. What would you use the 17 bits you realistically can not hear for? How can you use a 192dB dynamic range? What would you use it for? Does it make sense to make DACs for a signal format that you can not get any content for?
I think you should answer that before you judge the choices manufacturers have done.
You are trying to be funny right?
There are 128 bit processors long time ago, just not used as no real advantage over 64.
And still doesn't have much to do with the current dilemma.
Deceiving those who choose to be, perhaps.
If it were such a large deception then 16/44.1 media would have given way en masse to the far superior 24/96 and 32/384 etc.
But you know, such hardware is really only revelled by perhaps less than 0.05% of the population that consume digital audio.
(Think of all those people who listen to FM broadcast on in their cars. Think of the automotive hardened tuner and 24 bit DSP stage and poor grounding architecture producing perhaps 14 bits effective at the speaker. Then deduct the road noise for about 40dB SNR at the ear the cabin. Where are your bits now? Doesn't it make you want to cry?)
If it were such a large deception then 16/44.1 media would have given way en masse to the far superior 24/96 and 32/384 etc.
But you know, such hardware is really only revelled by perhaps less than 0.05% of the population that consume digital audio.
(Think of all those people who listen to FM broadcast on in their cars. Think of the automotive hardened tuner and 24 bit DSP stage and poor grounding architecture producing perhaps 14 bits effective at the speaker. Then deduct the road noise for about 40dB SNR at the ear the cabin. Where are your bits now? Doesn't it make you want to cry?
)
Last edited:
I can't see how one could blame TI, ESS or the other hardware manufacturers. They don't market their chips directly to the public as their market isn't the end customers but hifi manufacturers. The engineers working for those manufacturers know perfectly well what a 32bit dac means: a dac able to accept 32bit data as input.
One could blame the marketing teams of those hifi manufacturers for taking "32 bits" out of context and using it to convince their customers of the superiority of their product. Nihil novi.
One could blame the marketing teams of those hifi manufacturers for taking "32 bits" out of context and using it to convince their customers of the superiority of their product. Nihil novi.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.