Is the algorithm for non uniform two bits quantization correct?

[johndoe7748]

[GerdW]

Hi John,

---

@johndoe7748 wrote:

missing text

---

Your message is rather short and misses a lot of information!

• What is your algorithm? There is no explanation on what you want to program… (Wikipedia entry?, link to a scientific paper?)
• There only is an image, but no VI. From your image I can tell the VI is broken and so it is "not correct"!

[johndoe7748]

I'm so sorry but can you show me the algorithm in a picture with a block diagram?

[johndoe7748]

- Adjust bit depth of single tone signal and save it in binary file format
- - Single tone : 1MHz
- 8 bits / 4 bits / 2 bits / 2 bits / 2 bits / 2 bits / 2 bits of non-form 2 bits each
Perform an operation that reduces depth
- Set one threshold for non-uniform 2 bits

Then, can you show me the exact algorithm in a picture using a block diagram?

[GerdW]

Hi John,

---

@johndoe7748 wrote:

I'm so sorry but can you show me the algorithm in a picture with a block diagram?

---

As I asked in my previous message you need to supply information on the algorithm you want to implement.

You did not provide any information so far except an image of an unfinished block diagram…

Edit:

There's a lot of Korean in your paper. I cannot understand Korean (and also forgot how to pronounce the chars).

Is there an English version of your presentation? (Don't forget you ask for help in a worldwide forum…)

[johndoe7748]

If you look up, they answered again

[johndoe7748]

this is english version!

i want non-uniform 2bit algorithm with block diagram picture! thank you 

[student7564]

Here's the picture of the algorithm I made. But I think it's wrong.
I attached a related information pdf file

[GerdW]

Hi John,

as far as I understand your paper/homework description your VI doesn't look right to me…

For all those quantization routines (from 8bit to 2bit (non) uniform) you need to define thresholds to determine each level. For uniform quantizations those thresholds are easier to determine, for the non-uniform quantization you could try to determine optimized thresholds.

In the end your VI should take an input array of samples, compare each sample to the thresholds and output the resulting level data (from 8 to 2 bits). Then you "just" need to pack those bits as described in your paper.

None of that can be seen in your image…

Suggestion to determine the quantisized level for a sample:

@Student7564:

• Please don't create duplicate threads for the same topic.
• Please don't create new users just to ask the same question multiple times…

[student7564]

Can you make a photo of the non-uniform 2bit quantization algorithm by referring to the pdf file in your way? I'm sorry I don't have a vi file right now