Noise limits in India and the IFC guidelines: which one applies

Two sets of numbers apply to almost every lender-financed plant in India. The Noise Pollution (Regulation and Control) Rules, 2000 fix ambient limits in dB(A) L_eq for four categories of area, and rule 4(1) says noise in any area shall not exceed them.¹ The IFC General EHS Guidelines set one-hour L_Aeq limits at the receptor and add a second test: no more than a 3 dB increase over background at the nearest receptor off site.² The tables look alike. They are not.

Indian and IFC noise limits side by side

Where the two disagree

The hour from 06:00 to 07:00 is day time under the Rules and night time under IFC.^1,2 A shift change, a delivery convoy or a boiler start-up at 06:30 is judged against 55 dB(A) at a home under the Rules and 45 dB(A) under the IFC table. An operating schedule written around Indian day hours can fail the lender test in that single hour.

Averaging time differs too. The Rules describe L_eq as an energy mean over a specified period, but the Schedule does not fix the period.¹ The IFC limits apply to each hour.² A 16-hour daytime L_eq can sit below 55 dB(A) while several individual hours exceed it, so report both from the same logged data.

Then there is classification. In India the State Government categorises areas, a mixed area can be declared as one of the four categories, and a silence zone covers not less than 100 metres around hospitals, educational institutions and courts.¹ IFC classifies by receptor: residences, schools, hospitals and places of worship are all points of reception.² A house inside a notified industrial area is therefore still a residential receptor at 55 and 45 dB(A) for the lender, not 75 and 70. Commercial receptors swing the other way, with 55 dB(A) at night in India against 70 dB(A) under IFC.

The 3 dB test bites in quiet places

IFC says noise impacts should not exceed the Table 1.7.1 levels or result in a maximum increase in background of 3 dB at the nearest receptor off site.² Take a village with a night background of 35 dB(A). A plant that lifts it to 44 dB(A) passes the table and raises background by 9 dB. The wording does not say which test prevails, so we report both at every receptor and state which one governs. Background must exclude highly intrusive events such as aircraft flyovers and passing trains.²

That puts weight on the baseline. IFC expects monitoring designed by trained specialists, with periods long enough for statistical analysis: typically 48 hours of continuous logging, or several days covering weekdays and weekends, using a Type 1 or 2 sound level meter about 1.5 m above ground and at least 3 m from reflecting surfaces.² One daytime spot reading per location will not support a 3 dB argument. Our note on baseline duration covers the wider monitoring question.

Will the IFC noise limits change?

Not yet. The World Bank Group is updating its EHS Guidelines in phases, each draft going to a 45-day public consultation.³ On 23 October 2026 the update page listed closed consultations for sector guidelines such as airports, solar power and health care facilities, but no draft of the General EHS Guidelines, so the 2007 noise table still applies.³ That table derives from the WHO 1999 Guidelines for Community Noise.² WHO's 2018 environmental noise guidelines address road, rail and aircraft noise, wind turbines and leisure noise.⁴ Their stated scope does not include fixed industrial plant, so they do not replace the IFC table for a factory or power station.

The cheapest fixes come at design stage: quieter equipment, fan silencers, exhaust mufflers, enclosures, and barriers without gaps and with a surface density of at least 10 kg/m².² We run ISO 9613-2 noise models against both sets of limits as part of our environmental modelling work, and can deliver the noise chapter of an EIA or ESIA.