Im not sure what code you are referencing but i would be calling that a reject, if one isnt out of spec you might be able to call aqumulations of imperfections. It is a discontinuity, as is anything in the weld that isnt metal. Have you tried any other inspections on the part in question?
Well, the first thing that jumps in my mind is the gain. jumping the gain up is never a good thing. Most codes have a specific gain level to run at, going any hotter gives you a good chance of false indications. If at the prescribed scanning DB, it is acceptable, that is usually it, it's acceptable.
Then there is the nature of the indication you refer to. Rounded indications are omni directional reflectors. Consider the 1.5mm hole in the IIW block.
It makes a nice target and reflection. If you had a pile of them in any weld, I'd find it hard to believe that each of them could be sized individually from a simple physics of the waveforms standpoint.
I suggest being careful with this one, you may be setting yourself up for failure.
Regards,
Gerald