I think the topic might be clarified by a bit of history.
Before modern methods of non destructive testing became practical, seasoned veterens of the foundry, rolling mill, and forge shops would gather late in the shift and "sound" the days's production by tapping them with hammesr. By listening to the ding or the clunk they could spot defective work items and even pinpoint the location of the defect.
If the work was suspended or damped in various ways, struck in miltiple locations, compared in riing to identical items a very thorough examinatio was possible. Thus railroad wheels and axles, steam engine cylinders, anchor chains, shafting, and structural beams could be rejected for cold shuts, pipes, laminations, shrinks, voids, cracks, and a dozens other metallurgical and manufacturing defect based in the asessment of wise old ears.
I'm not suggesting vibration stress relief is all BS. I'm only saying I've seen no - zero - unqualified successes. I suspect it has to be intelligently applied amd for that the people specifying its use and the techs using it need to be schooled just as a man with a hemmer.
This makes sense as the ear is an incredibly sophisticated sound analysis device that unfortunately cannot be calibrated or included into an enginerred quality assurance program.
Many times when work was suspended and vigorously rung it would change shape or demsnsion slightly indicating a change in the material possibly from a resolution or re-distribution of locked in stresses. It was ony a srep to conclude that by circulating strain enegy through a raw casting, forging, fabricating, whatever stresses could be relaxed somehow. In the 1970 a number of well intended but over-sold vibration stress relieving gimmicks were touted in the manufacturing shows around the country and since senior executives and shop managers are no less gullible to flashy presentation than persons with penile perfprmance problems many systems were sold only to be misapplied.
Shocks from hammer blows shrewdly applied to items having locked in stresses may result in stress reduction because the energy rate and location of shock can be precisely controlled. Vibration stress relief is limited in energy rate and magnitude to nodes where resonance stress circulates to a greater degree, section changes, intersections, corners etc ie: places where fatigue could be expected.
I submit a skilled man with a hammer and keen ears selected for his intuition could resolve more stress in a metal item in about 1/10 the time and 1% of the surgical risk as a vibration stress system run by factrory trained techs.
Hell, I've done it. I used to run a big plate planer in the ship fitter shop cutting weld preps in the edge of 20 ft long bars burned from plate to be welded into elevator rails. Remove a double J-weld prep from the edge of burned plate and the result is a 3" to 7" of banana shape. During machining, the bars would slip and twist in the weenie ar cylinder clamps of the plate planer.. You have to send it to the bumper for straightening 3 time for some bars.
If I gave the ends of the bars a dozen mighty whacks with a 16 lb sledge and a healthy bash on about one foot centers down each of the burned edges, the bars needed no straightening for machining and only a little prior to weld.