Assimilation, if it occurs, I think it will be apparent in the spelling. Just to give an example, consider the verbs scaled to افتعل:
- سمع ← استمع
- غرر ← اغترَّ
- لحف ← التحف
- عرف ← اعترف
- ضرب ← اضطرب
- ضرر ← اضطرّ
Here in the last two examples we see how the typical T sound in the scale of افتعل is converted to velar ط to be in line and harmonize with the velar ض.
This said, I still hear it as "D" in the audio. However, I did notice that the speaker in the audio, even though a machine, but it apparently is based on an Egyptian female voice - The tone of the speech is distinctive. In their dialect they do indeed have a tendency sometimes to velarize the "D". Personally, I can spell the syllables صد and صض quite differently and easily.
2idghaam oftentimes happens in two letters of two different words where the first letter is in sukun and produces some almost similar sounds compared to the second letter. Example: اركبْ معنا here ب sukun meet م fatHa.
In صديق, there is no sukun except in ي that have different sounds with د and ق. And it's only one word. In many places, 2idghaam in one word will be 2izh-haar.
Furthermore, 2idghaam is only practiced when we recite Quran.