yes its like putting a speedo meter on the bike needle to 400 miles an hour and it fails to go beyond 150 miles an hour. So its cheating isnt it? what to do the unbelievable thing is that Texas Instruments going towards such cheaper business strategies.
Put a new name put a new architecture put everything and tell that the tech is different but dont tell that putting a 32 bit tag on it and which fails the dynamic range of even 22bit.
I think the days will come back when everybody will switch to complete discrete... i think its good idea lets give quality products to clients rather giving poor quality products.
So using a product where one claims 32bit and if it doesnt deliver as good as what a 22 bit can it means that the equipment manufacturers are cheating to customers? indirectly is that the meaning?
Put a new name put a new architecture put everything and tell that the tech is different but dont tell that putting a 32 bit tag on it and which fails the dynamic range of even 22bit.
I think the days will come back when everybody will switch to complete discrete... i think its good idea lets give quality products to clients rather giving poor quality products.
So using a product where one claims 32bit and if it doesnt deliver as good as what a 22 bit can it means that the equipment manufacturers are cheating to customers? indirectly is that the meaning?
All these comments about fraud seem excessive. We do not need 32bit DAC. 22 true bits is already much better than what we need on the analog side. We might however need DAC that accept 32bit inputs, to simplify the integration of those chipsets into a digital chain working at 24 or 32bits for good reasons.
Engineers designing devices around those new chips are well aware of this. The only fraud there might be is in the marketing directed towards the final users... not by TI but by their customers.
Engineers designing devices around those new chips are well aware of this. The only fraud there might be is in the marketing directed towards the final users... not by TI but by their customers.
I think you're missing the point. The PCM1795, or any other of the 32 bit DACs are only as fraudulent as their 32 bit spec itself, as in written on paper, goes. The real specs are quite clearly displayed as the available SnR within the datasheet. None of the DACs that are capable of accepting a true 32 bit signal try to claim that they can approach a real world SnR anywhere near what 32 bits could provide.
None of the real engineers designing the PCBs are going to think that the 32 bit DACs are going to provide that as an end result.
If being able to put 32 bit on an end product helps to sell said product by fooling the ignorant masses then so be it, this has been happening for years. PMPO power springs to mind.
What is interesting about a 32 bit product is the internal processing, where it adds in additional precision for any digital stuff that may happen to occur within. This is most certainly a big deal. The ES9018 and family are a prime example of this. Where yes, they are 32 bit and don't give you anything better than around 22 bit end performance, but the data pipe inside is 32 bits wide. When you then use the internal volume control you don't start throwing away resolution straight away because you've got 10 extra bits floating around with which to play with before that happens.
Now whether or not other 32 bit DACs do a similar thing I do not know. ESS make a big deal about promoting their advanced digital volume control, but if other 32 bit DACs started to do this, it would certainly make all 32 bit DACs better than their 24 bit counter parts. Sadly most manufactures don't say much beyond the fact their products can accept 32 bit data. Whether or not this is truncated before the digital processing occurs within we don't know.
Ironically all 24 bit DACs operate with a 64x bit clock anyway, they will work perfectly fine with a 32 bit wide input signal, it's just they throw away the other bits right at the input. If a 32 bit DAC threw away the additional bits before it processed anything then this isn't any different to the 24 bit DACs anyway.
None of the real engineers designing the PCBs are going to think that the 32 bit DACs are going to provide that as an end result.
If being able to put 32 bit on an end product helps to sell said product by fooling the ignorant masses then so be it, this has been happening for years. PMPO power springs to mind.
What is interesting about a 32 bit product is the internal processing, where it adds in additional precision for any digital stuff that may happen to occur within. This is most certainly a big deal. The ES9018 and family are a prime example of this. Where yes, they are 32 bit and don't give you anything better than around 22 bit end performance, but the data pipe inside is 32 bits wide. When you then use the internal volume control you don't start throwing away resolution straight away because you've got 10 extra bits floating around with which to play with before that happens.
Now whether or not other 32 bit DACs do a similar thing I do not know. ESS make a big deal about promoting their advanced digital volume control, but if other 32 bit DACs started to do this, it would certainly make all 32 bit DACs better than their 24 bit counter parts. Sadly most manufactures don't say much beyond the fact their products can accept 32 bit data. Whether or not this is truncated before the digital processing occurs within we don't know.
Ironically all 24 bit DACs operate with a 64x bit clock anyway, they will work perfectly fine with a 32 bit wide input signal, it's just they throw away the other bits right at the input. If a 32 bit DAC threw away the additional bits before it processed anything then this isn't any different to the 24 bit DACs anyway.
[First, a disclosure - I work for TI. I'm a PCM audio product definition engineer. I am completely biased.]
While they are based on a similar architecture, the PCM1795 and PCM1792A are different.
Looking historically, the PCM1796 and PCM1792 were based on similar designs, but the digital interpolation filter and current segment engine on the PCM1792 are larger and more complex than the 96. (hence it's 9dB better performance)
In 2009, we launched the PCM1795, which used the same analog section (current sources, voltage references etc) as the PCM1796, but with a new digital interpolation filter that could support 32bit data.
It was found that with an equal amount of averaging as a competitors device, that the analog back end in the PCM1796 could actually show analog content at levels equivalent to the 25th bit.
As a consequence, the PCM1795 was born as a potential replacement for the PCM1796.
s3tup also makes a good point, compared to the PCM1792, the PCM1795 is a lower cost product. That has been a strong part of it's success. At 123dB dynamic range, it is still considered an excellent DAC and remains a strong part of our portfolio.
(and now for the pitch... because you knew it was coming!
The PCM5142 was based on a similar architecture to the PCM1792. Whilst it has lower performance, it integrates WAY more than the PCM179x, and throughout the many, MANY hours of listening tests at our Golden Ear customers. (if I hear another Diana Krall song, I may go nuts), many of the experienced customers compared it tonally to the PCM1792.
I stop by this forum every once in a while. If I see more on this thread, I'll comment on my next pass through
Y'all are always welcome over on TI's e2e forum too. (TI E2E Community)
Cheers
Dafydd Roche
Any new ADDA stuff coming?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.