Digital BW, The Print

In audio not many people can determine the difference of a pure tone 
quantizied to 6 bits vs. 8 bits much less 16 bits. Actually I have heard 
claims that 4 bits is sufficient. A long time ago we did a study on 
this. We were remoting receivers back stateside and we were concerned 
with the minimum bandwidth we could get away with.

But if you now ask the question of a complex audio like a piece of music 
- while you might not be able to hear the quantization noise, you sure 
can tell the difference between Mozart played at 8 bits and Mozart 
played at 16 bits. The extra dynamic range shows up and by making the 
music much "fuller" and more robust - what ever the hell that means.

The same I think holds for our eyes when you consider a complex image 
like a B&W photograph.

Truman

Austin Franklin wrote:

> Hi Truman,
>
> > I believe that. Even the best digital recievers only use about 3/4s of
> > the dynamic range. Even at that we are taking 11 or 12 bits. If you scan
> > a 8 bits you probably get 6 bits dynamic range.
>
> Actually, if your scanner is, say a 12 bit scanner, and you scan at 8 
> bits,
> it still scans at 12 bits...then takes the setpoints and curves you 
> apply in
> the scanner interface software, and gives you adjusted 8 bit files, 
> that are
> actually full 8 bits of dynamic range.
>
> > But 876 to 2982 in your example is 11 bits quantization and is much
> > better than most photgraphic papers and inkjet printers are capable of
> > separating. So I would think that compressing the max amout of detail
> > into the dynamic range of the negative is still the way to go.
>
> Yes, you are absolutely correct!  That's one of the issues I have 
> always had
> with discussing things like this.  We humans can only see between 100 
> to 200
> distinct tones, but...as has been mentioned, you really need some larger
> number of tones to get smooth tonal transitions...though you can't
> "distinguish" them, they are still visible, and give that "smoothness" to
> the image.  Now, how many is debatable ;-)
>
> Regards,
>
> Austin
>