Could you please clarify the thoughts behind this question, are you experienceing some kind of issue with the file size?
A chart object in general is not that space requiring that it should be a concern. The concern would more likley be the memory used when the chart is calculated, which is related to the dimensions and expresison used in the chart related to the data model.
Actually i faced this question in recent interview.
I told the same answer what u said. (Its depend on dimensions we taken n expression we write.) They said if we taken same dimension and expression in all charts which will take more space? I said it may be bar chart. Is it wright r wrong?
I think u get some clarification.
Although it is right that it all depends on expressions and dimensions, and the complexity of both, a pivot table will usually perform poorer (when a huge dataset is handled in the chart) than any other object. I don'f have some deep technical specifications but the pivot table allows you to pivot, what means that the same expression may return different results for each dimension, depending on how dimensions and expressions are pivoted.
In addition, you can expand and collapse by several levels which takes additional work to display the chart.
When huge datasets are handled, and performance matters, I'd recommend you to use straight tables when you need to display values instead of pivot tables.
Hope that makes sense and helps.