FPGA Central

Hello all . . .

The thread on standards got me thinking about the original Red Book
pre-emphasis/de-emphasis standard . . . the de-emphasis flag only exists on
very few recordings in my collection, and they're all pre-1990 classical
releases. I also recall that "back in the day" there existed many CD
players with (automatically switched) analog de-emphasis as part of the
re-construction filter, and now this is almost universally implemented in
the digital filter (strictly for legacy support, I suppose).

My questions are, what was the thinking behind the inclusion of this as part
of the standard? Other than the obvious reduction in headroom and higher
parts count (for analog de-emphasis) in the players, what were the
drawbacks? What would motivate a record producer/recording engineer to
specify this in their recording?

Any insight?

Kirk Patton

"Kirk Patton" <[email protected]> wrote in message news:<5zjGb.289$[email protected] gy.com>...
> Hello all . . .
>
> The thread on standards got me thinking about the original Red Book
> pre-emphasis/de-emphasis standard . . . the de-emphasis flag only exists on
> very few recordings in my collection, and they're all pre-1990 classical
> releases. I also recall that "back in the day" there existed many CD
> players with (automatically switched) analog de-emphasis as part of the
> re-construction filter, and now this is almost universally implemented in
> the digital filter (strictly for legacy support, I suppose).
>
> My questions are, what was the thinking behind the inclusion of this as part
> of the standard? Other than the obvious reduction in headroom and higher
> parts count (for analog de-emphasis) in the players, what were the
> drawbacks? What would motivate a record producer/recording engineer to
> specify this in their recording?
>
> Any insight?
>
> Kirk Patton


It's pretty simple. Back in the late 1970's, no one could make an
economical 16-bit converter that was anywhere near true 16 bits. So by
using pre/de-emphasis, you could "hide" all the bad things that bad
converters do (since the converters come AFTER the pre-emphasis and
BEFORE the de-emphasis). It counts on the fact that the typical music
spectrum falls off rapidly above 4 KHz or so.

As a designer of ADC/DAC chips, I have spent way too much time
designing digital filters to implement this function when probably 1
in 10000 CD's actually has the pre-emphasis enabled. If I had a time
machine I would go back and change this!


Bob Adams

"Liz" <[email protected]> wrote in message
news:[email protected] om...

> It's pretty simple. Back in the late 1970's, no one could make an
> economical 16-bit converter that was anywhere near true 16 bits. So by
> using pre/de-emphasis, you could "hide" all the bad things that bad
> converters do (since the converters come AFTER the pre-emphasis and
> BEFORE the de-emphasis). It counts on the fact that the typical music
> spectrum falls off rapidly above 4 KHz or so.

OK. Duh. I do recall seeing it used only on program material with
comparatively limited HF useage . . . I somehow had this impression that use
of the pre-emphasis would result in normalizing the audio down a bit and
making poor low-level performance even more apparent, but this probably
didn't happen in practice . . . I was also thinking about a correlation
between this and Sony's first approaches to analog post-filtering, which is
likely to be non-sequitur.

> As a designer of ADC/DAC chips, I have spent way too much time
> designing digital filters to implement this function when probably 1
> in 10000 CD's actually has the pre-emphasis enabled. If I had a time
> machine I would go back and change this!
>

FWIW, Bob, it seems that several DAC chip manufacturers are very proud of
your work! I still see lots of paper devoted to the graphing of the
accuracy of these legacy filters . . . even for fs=32kHz & 48kHz. Maybe
somebody should sue the RIAA over it for loss of revenue

Regards,

Kirk Patton