Dissecting bitfields larger than 32 bits

I am writing a dissector for a protocol which contains bitfields that are larger than 32 bits. For instance there is one bitfield that is 48-bits long, with other bitfields being 80-bits long. Setting the bitmask for the field in the header_field_info structure does not work since the bitmask is only 32-bits long.

I believe that there are a number of possible ways to proceed:

1. I can extract the bytes from the tvb into local variables and extract the relevant bits for each field (possibly using bit accessors such as tvb_get_bits64 where appropriate) and then use proto_tree_add_text to add the values to the display
2. Treat the bitfield in a similar way to compressed data and extract the bit fields into a new buffer as if they were byte aligned then use tvb_new_real_data, etc., to add a new data source and then dissect the new tvb buffer from offset 0

Neither method will be as nice or as easy as using the inbuilt bitmask method. It would be nice to display the individual fields as the standard bitmask dissection (with the bit pattern preceding the field title and contents) but I cannot think of a nice way of doing this.

What are the pros and cons of these two methods and which should I choose? Is there another, better way to proceed that I haven't thought of?