Compiler/OPT9221: Calculating depth from phase in de-aliasing mode

Suramya, Thank you for your answer!

for my case I set dealiased_ph_mask = 3.

But I got raw phase values from range 0 to 4095 it is seem ok.

In datasheet it is written that output phase data are 12-bit, I do so:

unsigned short phase = 0;

// some code


phase = phase <<4;
phase = phase >>4;

for erasing flags data.

But how can I see that raw phase data have valid values?

And Is it possible to use this formula for calculating depth:

depth = phase*R/(2*f_eff)?

if not, so why it is need to calculate firstly de-aliased phase, and only after depth?

As I understand dealiased_ph_mask operates with phase data into the chip...

Suramya, Thank you for tour answer!

May be it is more clear for me.

But for dealiased_ph_mask = 3,  max phase value is 2048

temp = (4095 * 2* 3)>> (5 - 3) = 6144 = 1100000000000 (in binary )

1100000000000 &111111111111 = 100000000000 =2048 (in decimal)

so, is it right conversion for raw phase data:  (1)

unsigned short phase = 0;

// some code


phase = phase <<4;
phase = phase >>4;

I checked values for register: 5C 02 0x101800; 5C 03 0x900C42    (2)

use_dealiasing_mode = true

ind_freq_data_en = false
ma = 2
mb = 3
ka = 2
kb = 1
dealiased_ph_mask = 3

Where do I make mistake (1), (2) or... ?

Suramya, Thank you for tour answer!

I done as you advised me.

unsigned short phase = 0;

// some code


phase = (phase&0xFFF);

but any way phase values are more 2048 and it is seem have range [0, 4095]

I checked my code and did not found anything wrong...

// some code
arrayDpt[i] = (unsigned short)phase;

// some code

    imgDpt = cv::Mat::zeros(f_size.height, f_size.width, CV_16UC1);

//some code

memcpy((char*)imgDpt.data + 640 * fragment_gl, arrayDpt, 320 * sizeof (unsigned short)); 

// copy frame fragment to openCV matrix

// I output in log some 1 per 100 frames 

printf("phase at cx, cy = %hu\n", imgDpt.at<unsigned short>(cy, cx) );

where can be mistake?

may be register description has mistake?

And while ind_freq_data_en = true, phase values can be found in range [0, 4095] for f1 and f2 frequencies?

for example, R = 7.5 m, dealiased_ph_mask = 3. (if all is OK with raw phase data)
so phase = 2047 corresponds distance 7.5 but using formula (26),
distance will be ~2.5 m but not ~7.5m

or may be Is there some intermediate formula when together with (26)  from phase = 2047 it is possible to get distance = 7.5m?

Suramya, Thank you for tour answer!

If I understood rightly, so:

1) And phase max defines number step bin for respective distance, Yes?

2) Are phase values from range [0, 4095] for ind_freq_data_en = true for each frequency?

3) I made nonlinearity and common offset calibrations with phase_mask = 3.

 Are calibrations values valid for case phase_mask = 2?

4) I have sufficiently big noise level on second frequency, but on first one noise level is lesser.

What can be the reasons of it?

Suramya, Thank you for tour answer!

I shared the raw data in other my question.

https://e2e.ti.com/support/sensors/f/1023/p/748490/2769914#2769914