Impossible to seek frame by frame position with AVFoundation plugin for IOS

The function setPosition(qint64) cannot reach each frame of the video in mpeg4 h264, but only key frames.
I want to read a video in slow motion, frame by frame and I can read only one frame out of 30.

I find that the function SetPosition calls [playerItem seekToTime:newTime] in the plugin AVFoundation of QtMultimedia, in the class AVFMediaPlayerSession.
I need the capability to do [playerItem seekToTime:newTime toleranceBefore:kCMTimeZero toleranceAfter:kCMTimeZero] as indicated in the Apple AVFoundation coding guide, in order to achieve precise positionning.

I tried subclassing to solve my problem by following this link http://doc.qt.io/qt-5/multimediabackend.html, but i encountered d_ptr and private class QMediaPlayerPrivate that forbids me to subclass QMediaPlayer.

Can you give indications as how to solve my problem or make a patch please? Thanks.