Digital, but not by the numbers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The relevance I see is that the engineers (or perhaps the managers) responsible appealed to their numbers to justify the safety of their system, rather than going on the evidence from patients. The parallel with appeals to THD and SNR to justify digital audio's superiority over analog was too obvious for me to ignore.
 
Software is different from hardware though, and is very hard to quantifyfor reliability or to measure and provide a metric for itsreliability. The only figures I have seen were USA millitary based/aero based and werefrightening, I belkieve it was as bad as 10%, ie every 10 lines of code had an error of some sort.
Hardware can be tested, but ifit is controlled by software you are at the mercy of the programmers. I can remember the days when a design dept. had maybe 10 engineers who did hardware and the tiny bit of software between them. These days you'll have a couple of hardware engineers and about 60 software engineers.The days of hardware based systems is diminishing, with hardware becoming more customisable as it has both emebdded software and often disc based software to controll it. Is this better,IMO no, look at NASA'a Apollo navigation system and compare it with todays software based systems, or OS9 (or Linux) operating system compared to Windows. There are numerous stories regarding software faux pas when controlling hardware that would make you go live in a deep cave, luckily most of them do get sorted out during tested, but as your example shows not all do.
 
That the metric happened to concern reliability is entirely incidental to my posting the story here. So its relatively OT to wander off into talking about software vs hardware reliability :D Its the human factors I'm concerned about, not the technical ones so much - even though I agree the technical ones are interesting too.
 
Last edited:
The relevance I see is that the engineers (or perhaps the managers) responsible appealed to their numbers to justify the safety of their system, rather than going on the evidence from patients. The parallel with appeals to THD and SNR to justify digital audio's superiority over analog was too obvious for me to ignore.

you don't build silicon by listening to it. Silicon is designed by engineers which apply sound engineering principles to meet design objectives. Sure the standards surrounding digital audio may have been based on engineering compromises and expected sound quality but other than that nobody designs silicon based on imaginary subjective qualities that can't be measured or translated into hardware. No silicon foundries work this way unless they are run by crackpots.
 
you don't build silicon by listening to it.

Some do - have you listened to Martin Mallinson's talk at RMAF last year?

Silicon is designed by engineers which apply sound engineering principles to meet design objectives.

The irony here amuses me no end :) Is the design objective end-user listening satisfaction or is it THD and SNR?

Sure the standards surrounding digital audio may have been based on engineering compromises and expected sound quality but other than that nobody designs silicon based on imaginary subjective qualities that can't be measured or translated into hardware.

Indeed - strawman.
 
I haven't found it to be a problem so far, so I'm giving it almost zero attention in this project. I suspect this could be because of my choice of multibit (rather than S-D) DACs. People who choose S-D DACs have to pay attention to jitter ISTM because those kinds of DACs work on the principle of generating lots of HF energy. Jittering the HF noise induces audio band noise modulation - coz jitter sensitivity scales with frequency and amplitude. At least that's my take on it - other ideas welcome of course :)
 
Some do - have you listened to Martin Mallinson's talk at RMAF last year?



The irony here amuses me no end :) Is the design objective end-user listening satisfaction or is it THD and SNR?



Indeed - strawman.

yes but when you say that some bit of kit sounds better and the differences can't be measured by normal means then how on earth do you translate that into a design brief ?? What possible directions would you give to a silicon designer ?
 
Last edited:
Here's an interesting (as in hilarious :D) post on Head-Fi which is pretty typical of the 'objectivist' position, if anyone's still in any doubt. The discussion is over a DAP (Charles Altmann's Tera Player) which subjectively raved about but has so far no published 'objective' measurements : Altmann Tera Player


Oh my, you really drank the Kool-Aid, didn't you?
 
Here's an interesting (as in hilarious :D) post on Head-Fi which is pretty typical of the 'objectivist' position, if anyone's still in any doubt. The discussion is over a DAP (Charles Altmann's Tera Player) which subjectively raved about but has so far no published 'objective' measurements : Altmann Tera Player

So when Sanken designs output transistors they design them based on immeasurable subjective design criteria :D LOL
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.