[Terminal]: Screen Reader does not announce Pop-up name when its focus first jump inside pop-up.

### Windows Terminal version

1.12.3472.0

### Windows build number

10.0.22504.1010

### Other Software

**Test Environment:**
OS: Windows 11 Version Dev (OS Build 22504.1010)
App: Windows Terminal Preview
Screen Reader: Narrator

### Steps to reproduce

**Repro Steps:**
1. Open Windows Terminal.
2. Open Narrator using 'Win + Ctrl + Enter' keys.
3. Press 'Ctrl + Shift + space' keys to open new tab Dropdown pop-up, Observe the issues.
4. Press 'Ctrl + Shift + r' keys to open Set tab color pop-up, Observe the issues.

**User Experience:**
Screen Reader users UX will not be good as they will not be able to know the purpose pop-ups.

**Guideline Reference:**
https://www.w3.org/WAI/WCAG21/Understanding/info-and-relationships

**Attachments:**
[Screen Reader does not announce Pop-up name when its focus first jump inside pop-up.zip](https://github.com/microsoft/terminal/files/7783133/Screen.Reader.does.not.announce.Pop-up.name.when.its.focus.first.jump.inside.pop-up.zip)

### Expected Behavior

Screen Reader announces Pop-up name when its focus first jumps inside 'New Tab Dropdown' & 'Set tab color' pop-up.
i.e. When user open 'New Tab Dropdown' using 'Ctrl + Shift + space' using Screen Reader should announce 'New Tab Dropdown pop-up window ....' and When user open 'Reset tab color' using 'Ctrl + Shift + r' using Screen Reader should announce 'Set tab Color pop-up window....'

### Actual Behavior

Screen Reader does not announce Pop-up name when its focus first jumps inside 'New Tab Dropdown' & 'Set tab color' pop-up.

**Observation:** Same issue repro with NVDA and JAWS Screen Reader.

Start by reproducing the issue in Windows Terminal with Narrator using the listed shortcuts for the New Tab Dropdown and Set tab color pop-ups, then compare behavior with NVDA or JAWS. Done means the screen reader announces each pop-up's name when focus first enters it, including the expected names described in the issue.

Written by the indexing model from the issue text.