Jim Lesurf wrote:
In article , Mr.T
MrT@home wrote:


16 bits was an obvious choice because it's two bytes and provides a
sufficient degree of overkill. What you could also say is that not
for nothing was the early use and acceptance of 14 bit CD players,
when 16 bit converters were more difficult/expensive to make.

In fairness, I should point out, though, that the first generation
Philips '14 bit' chipsets for CD players actually used x4
oversampling. Thus - in principle at least - returned 16-bit
resolution.


Pray tell how oversampling increases resolution ? The reason for
oversampling was/is to make reconstruction filters easier to implemnt
without artifiacts of a steep slope. It's been a whil, have I forgotten ?

geoff

On 2006-10-29, Geoff wrote:
Jim Lesurf wrote:
In article , Mr.T
MrT@home wrote:
16 bits was an obvious choice because it's two bytes and provides a
sufficient degree of overkill. What you could also say is that not
for nothing was the early use and acceptance of 14 bit CD players,
when 16 bit converters were more difficult/expensive to make.

In fairness, I should point out, though, that the first generation
Philips '14 bit' chipsets for CD players actually used x4
oversampling. Thus - in principle at least - returned 16-bit
resolution.

Pray tell how oversampling increases resolution ? The reason for
oversampling was/is to make reconstruction filters easier to implemnt
without artifiacts of a steep slope. It's been a whil, have I forgotten ?

I have sometimes wondered about the Philips x4 upsampling DAC in early
CD players (I use "upsampling" here to distinguish from the use of
oversampling in the ADC case).

I assume (but have never looked for proof) that the conversion of a single
16-bit sample xx..xxYY (YY are the two LSBs) would be accomplished by
replacing the single 16-bit sample by four 14-bit samples as follows:

xx..xx00: xx..xx, xx..xx, xx..xx, xx..xx

xx..xx01: xx..xx, xx..xx, xx..xx, xx..xx+1

xx..xx10: xx..xx, xx..xx, xx..xx+1, xx..xx+1

xx..xx11: xx..xx, xx..xx+1, xx..xx+1, xx..xx+1

Or something similar. The DAC will effectively interpolate so the LSBs
are not lost. The noise floor will be right for 16 bits because of
the upsampling.

I wonder if the amplitudes of the preceding and succeding samples should
be taken into account to determine the right order of the +1s in the
interpolation? Probably not as I suspect the spectrum differences will
fall above the original Nyquist limit.

John
--
John Phillips

In article , John Phillips
wrote:
On 2006-10-29, Geoff wrote:
Jim Lesurf wrote:

In fairness, I should point out, though, that the first generation
Philips '14 bit' chipsets for CD players actually used x4
oversampling. Thus - in principle at least - returned 16-bit
resolution.

Pray tell how oversampling increases resolution ? The reason for
oversampling was/is to make reconstruction filters easier to implemnt
without artifiacts of a steep slope. It's been a whil, have I
forgotten ?

I have sometimes wondered about the Philips x4 upsampling DAC in early
CD players (I use "upsampling" here to distinguish from the use of
oversampling in the ADC case).

I'd prefer to call it 'oversampling' in both cases for various reasons. One
being that in some situations 'upsampling' may be a distinctly different
practice.

I assume (but have never looked for proof) that the conversion of a
single 16-bit sample xx..xxYY (YY are the two LSBs) would be
accomplished by replacing the single 16-bit sample by four 14-bit
samples as follows:

xx..xx00: xx..xx, xx..xx, xx..xx, xx..xx

xx..xx01: xx..xx, xx..xx, xx..xx, xx..xx+1

xx..xx10: xx..xx, xx..xx, xx..xx+1, xx..xx+1

xx..xx11: xx..xx, xx..xx+1, xx..xx+1, xx..xx+1

Or something similar. The DAC will effectively interpolate so the LSBs
are not lost. The noise floor will be right for 16 bits because of the
upsampling.

I wonder if the amplitudes of the preceding and succeding samples should
be taken into account to determine the right order of the +1s in the
interpolation? Probably not as I suspect the spectrum differences will
fall above the original Nyquist limit.

The above is essentially the same explanation that I would have given,
but since John puts it quite neatly, I need not bother. :-) A more
detailed explanation is given in the special issue of Philips Tech Rev
that was released at the same time as CD audio was launched, and
describes CD audio and the initial chipsets.

The samples are 'noise shaped'[1] by a process along the lines that the top 14
bits of each sample are DAC converted and fed out as an analog level, and
the 'unused' 2 LSB are fed back and combined with the next sample value.
The simplest method is the one described above, but alternative feedback
shaping processes can be used.

The output filter then acts to take a 'running average'. Four 14 bit values
then sum or average to give a 16-bit result in the passband of the analogue
filtering arrangement.

In principle, the behaviour is the same as when any 'low bit depth' DAC is
used (with oversampling and noise shaping) to get results with higher
depths.

Thus by using oversampling and noise shaping we can symultaneously ease the
burden on the analog reconstruction filter that follows DAC conversion, and
allow the use of a DAC with less than 16 bits. This also is the basis of
other methods like low-bit DAC delta-sigma, 'bitstream', and various other
commercial techniques which use the same general approach to obtain both
a shift of reconstruction images to higher frequencies (thus easing analog
filter requirements) and obtaining high resolutions.

Hence the original Philips 14-bit x4 oversampling system would be able,
in principle, to deliver full 16-bit resolution *if* the chips and the
associated electronics was made with suitable care. As usual, the practical
limits end up being determined by the care put into engineering the
actual implimentation. :-)

Slainte,

Jim

[1] I regret the term 'noise shaped' in this context since we are talking
about a deterministic process, but it became the standard term, so we
seem to be stuck with it!

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

On 2006-10-30, Jim Lesurf wrote:
In article , John Phillips
wrote:
I have sometimes wondered about the Philips x4 upsampling DAC in early
CD players (I use "upsampling" here to distinguish from the use of
oversampling in the ADC case).

I'd prefer to call it 'oversampling' in both cases for various reasons. One
being that in some situations 'upsampling' may be a distinctly different
practice.

I assume (but have never looked for proof) that the conversion of a
single 16-bit sample xx..xxYY (YY are the two LSBs) would be
accomplished by replacing the single 16-bit sample by four 14-bit
samples as follows:

xx..xx00: xx..xx, xx..xx, xx..xx, xx..xx

xx..xx01: xx..xx, xx..xx, xx..xx, xx..xx+1

xx..xx10: xx..xx, xx..xx, xx..xx+1, xx..xx+1

xx..xx11: xx..xx, xx..xx+1, xx..xx+1, xx..xx+1

Or something similar. The DAC will effectively interpolate so the LSBs
are not lost. The noise floor will be right for 16 bits because of the
upsampling. ...

The above is essentially the same explanation that I would have given,
but since John puts it quite neatly, I need not bother. :-) A more
detailed explanation is given in the special issue of Philips Tech Rev
that was released at the same time as CD audio was launched, and
describes CD audio and the initial chipsets.

The samples are 'noise shaped'[1] by a process along the lines that the top 14
bits of each sample are DAC converted and fed out as an analog level, and
the 'unused' 2 LSB are fed back and combined with the next sample value.
The simplest method is the one described above, but alternative feedback
shaping processes can be used.

The output filter then acts to take a 'running average'. Four 14 bit values
then sum or average to give a 16-bit result in the passband of the analogue
filtering arrangement.

In principle, the behaviour is the same as when any 'low bit depth' DAC is
used (with oversampling and noise shaping) to get results with higher
depths.

Thus by using oversampling and noise shaping we can symultaneously ease the
burden on the analog reconstruction filter that follows DAC conversion, and
allow the use of a DAC with less than 16 bits. This also is the basis of
other methods like low-bit DAC delta-sigma, 'bitstream', and various other
commercial techniques which use the same general approach to obtain both
a shift of reconstruction images to higher frequencies (thus easing analog
filter requirements) and obtaining high resolutions.

Hence the original Philips 14-bit x4 oversampling system would be able,
in principle, to deliver full 16-bit resolution *if* the chips and the
associated electronics was made with suitable care. As usual, the practical
limits end up being determined by the care put into engineering the
actual implimentation. :-)

I looked up the details of the chipset. It seems that the SAA7030 does
the interpolation and does a large part of the reconstruction filtering.
It has a 96-point FIR filter at 4x input rate with a 28-bit accumulator
(16-bit data, 12-bit coefficients).

The thing that struck me from the datasheet is that the 28 bit accumulator
seems to get truncated without dither to 14 bits and then sent to the DAC
(a TDA1540).

Am I correct in thinking this undithered truncation is likely to generate
significant quantization error?

--
John Phillips

"John Phillips" wrote
in message

I looked up the details of the chipset. It seems that
the SAA7030 does the interpolation and does a large part
of the reconstruction filtering. It has a 96-point FIR
filter at 4x input rate with a 28-bit accumulator (16-bit
data, 12-bit coefficients).

The thing that struck me from the datasheet is that the
28 bit accumulator seems to get truncated without dither
to 14 bits and then sent to the DAC (a TDA1540).

The filter acts like an integrator and averages 4 samples to create a sample
with higher accuracy. My understanding of how this works suggests that the
final result would have about the same dynamic range as an ideal 15 bit DAC.
IOW, the LSB is toggling noisily.

Am I correct in thinking this undithered truncation is
likely to generate significant quantization error?

No, because there is no truncation in the final result.

In article , John Phillips
wrote:
On 2006-10-30, Jim Lesurf wrote:
[snip]

Hence the original Philips 14-bit x4 oversampling system would be
able, in principle, to deliver full 16-bit resolution *if* the chips
and the associated electronics was made with suitable care. As usual,
the practical limits end up being determined by the care put into
engineering the actual implimentation. :-)

I looked up the details of the chipset. It seems that the SAA7030 does
the interpolation and does a large part of the reconstruction filtering.
It has a 96-point FIR filter at 4x input rate with a 28-bit accumulator
(16-bit data, 12-bit coefficients).

The thing that struck me from the datasheet is that the 28 bit
accumulator seems to get truncated without dither to 14 bits and then
sent to the DAC (a TDA1540).

My (unreliable) recollection is that noise shaping is employed, but (see
below) I can't recall the details off-hand.

Am I correct in thinking this undithered truncation is likely to
generate significant quantization error?

I'd need to re-read the references to be sure. My (unreliable) recollection
is that the bits not sent to the DAC are returned into the noise shaping.
(e.g. the top 14 bits are 'subtracted' once sent to the DAC, but the LSBs
are retained to combine with the next oversample value result.)

In general, any set of multiplication and addition processes like those
used in the filters, etc, might cause some truncation problems - if only
because both integer and float math uses values with a finite number of
quantised values. Hence, for example, the coefficients used for the
multiply may not be exactly the 'sinc' values if that was the intended
pattern. In practice, the aim would be to ensure that any such rounding
errors do not accumulate and remain tiny wrt the intended range. As might
reasonably be expect for 28 bit values being reduced down to 14 or 16.

Hence there is a distinction being quantization errors existing and being
'significant'. :-) Can't comment on that without looking at the details
again. All depends on the details.

The above is, IIRC, one of the arguments for applying dither to almost
*any* computation process which deals with serial pattern signals, save for
the most trivial computations.

I'll try an remember to dig out the relevant Philips Tech Rev, etc,
tomorrow, and have a look to check.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

"Jim Lesurf" wrote in message

In article ,
John Phillips wrote:
On 2006-10-30, Jim Lesurf
wrote:
[snip]

Hence the original Philips 14-bit x4 oversampling
system would be able, in principle, to deliver full
16-bit resolution *if* the chips and the associated
electronics was made with suitable care. As usual, the
practical limits end up being determined by the care
put into engineering the actual implimentation. :-)

I looked up the details of the chipset. It seems that
the SAA7030 does the interpolation and does a large part
of the reconstruction filtering. It has a 96-point FIR
filter at 4x input rate with a 28-bit accumulator
(16-bit data, 12-bit coefficients).

The thing that struck me from the datasheet is that the
28 bit accumulator seems to get truncated without dither
to 14 bits and then sent to the DAC (a TDA1540).

My (unreliable) recollection is that noise shaping is
employed, but (see below) I can't recall the details
off-hand.

I found the spec sheet for the TDA1540 and SAA 7030 online, and can confirm
that noise shaping is done in the SAA 7030.

The TDA 1540 spec sheet was found at the Signetics web site.

BTW the speced dynamic range of the TDA 1540 is 85 dB.

In article , Arny
Krueger
wrote:

I found the spec sheet for the TDA1540 and SAA 7030 online, and can
confirm that noise shaping is done in the SAA 7030.

The TDA 1540 spec sheet was found at the Signetics web site.

BTW the speced dynamic range of the TDA 1540 is 85 dB.

Not read the data sheets. However I've now had a chance to re-read the
special issue of Philips Tech Rev that includes

Digital-to-analog conversion in playing a Compact Disc.
Goedhart, et al.
Philips Tech Rev V40(6) 1982 pages 174-9

This paper outlines how the SAA7030 and TDA1540 operate as part of the
conversion system.

This confirms the noise shaping, essentially by the method of taking the
LSB portion of the 28 bit accumulator and employing it as a carry forwards
to combine with the next filter-computed oversample.

Although the dynamic range is around 85dB this is essentially for the x4
bandwidth, and the paper explains that the result should end up being more
like 97dB if the devices operate as intended.

Two reasons for this.

1) Even with a 'white' quantisation noise spectrum the final bandwidth only
covers a quarter of the oversampled rate bandwidth, so this would give a
6dB improvement.

2) The noise shaping actually generates a noise spectrum which rises with
frequency, thus the 85dB noise is predominantly above 22kHz. This improves
the result according to their analysis by another 7dB or so over what you'd
get for 'white' noise.

The results are broadly in line with the use of noise shaping in other,
more modern, oversampling systems that use lower bit-depths than the input
data.

Given that this was the first system Philips used, it still looks
remarkably 'fresh' in concept. Hardly surprising that some marketing types
and journalists have had to use the term 'upsampled' more recently to try
and pretend they have come up with a new idea, when this may not always be
so. :-)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

"Geoff" wrote in message

Jim Lesurf wrote:
In article
, Mr.T
MrT@home wrote:
16 bits was an obvious choice because it's two bytes
and provides a sufficient degree of overkill. What you
could also say is that not for nothing was the early
use and acceptance of 14 bit CD players, when 16 bit
converters were more difficult/expensive to make.

In fairness, I should point out, though, that the first
generation Philips '14 bit' chipsets for CD players
actually used x4 oversampling. Thus - in principle at
least - returned 16-bit resolution.

Pray tell how oversampling increases resolution ?

http://www.daqchina.net/daqchina/circuit/adpro.pdf

The reason for oversampling was/is to make reconstruction
filters easier to implemnt without artifiacts of a steep
slope.

That's one reason of several.

It's been a whil, have I forgotten ?

yep.