Its not the case of capacity of MAF, air input or injector size but how air is measured. The MAF measures a unit of airflow velocity while ECU is looking for a value of air input Q. the Mass Volume of air input which is a function of the average velocity of slug of air, the cross section area of the slug of air, and the temperature of the air to calculate density to get the mass. The MAF samples the velocity and temperature of a piece of this cross section. But this is not the average velocity of the whole cross-section; the only way to get the average is through extensive testing of the flow and dynoing by the manufacturer. and building a correction table within the ECU. ie at velocity 100 actual average flow may be say 95, while at 150 actual may be say 138. (the point being that is not 1 for 1 and may not even be linear with change in flow rate.). Now changing the path of airflow into the MAF can most likely change this relationship, now the new MAF reading will either result in more or less volume of air that the computer things is incoming. If the flow has gotten more efficient a MAF reading of 100 may now mean actual average airflow is say 102... the ECU still thinks its 95 ... and sets fuel rate for such; however, actual air flow is 7% higher ... and thus mixture is 7% leaner ... oops.. The only solution is building a new table in the tuning is accessible, or adjusting fuel trims ... this requires actual dyne time is active wideband O2 sensing. My Lotus spent 12 hours of running dyne time to build a new intake box and combination of MAF tube diameter trials and flow straighter tests to get a produce that while the flow could go higher that the stock intake, the end result was that for a given measured flow reading its still matched the average cross section flow.