Why different approach for setting filter channels?

C4W1 - In video “Convolutions Over Volume” It show we can use only same number of filter with Input channels. (i.e Input is 6x6x3 and filter is 3x3x3)
Please check image: https://i.imgur.com/3jyWm3m.png

While in other video of same week “Why Convolutions?” It show different filters than input channels, (i.e Input is 32x32x3 and filter is 28x28x6)
Please check Image: https://i.imgur.com/plRggNP.png

I have big confusion in that …

In the first figure, the emphasis is on number of channels of the filter = number of channels in input. The number of filters is a different parameter.
In the second figure, there are 6 filters with width = height = 5. Each of the 6 filters will have dimension 5x5x3. As a result, the output has dimensions 28x28x6.