Elastic search Nested Documents and Terms Aggregation

# Given this index and mapping:

POST /test
{
    "settings" : {
        "number_of_shards" : 1
    },
    "mappings" : {
        "testdoc" : {
            "properties" : {
              "META": {
               "type": "nested",
               "properties": {
                "scheme_name": {
                  "type": "string",
                  "index": "no"
                },
                "scheme_data": {
                  "type": "nested",
                  "properties": {
                  "column_name": {
                    "type": "string",
                    "index": "no" },
                  "value": {
                    "type": "string"}}}}}}}}}



# And this test data:

POST /test/testdoc
{
  "path": "/matt/matt-pics",
  "file": "dog.jpg",
  "mime_type": "image/jpeg",
  "META": [
    {
      "scheme_name": "Pseudo_EXIF_Scheme",
      "scheme_data": [
        {
          "value": "Nikon",
          "column_name": "CAMERA_MANUFACTURER"
        },
        {
          "value": "D80",
          "column_name": "CAMERA_MODEL"
        }

      ]
    }
  ]

}

POST /test/testdoc
{
  "path": "/matt/matt-music",
  "file": "song.mp3",
  "mime_type": "audio/mp3",
  "META": [
    {
      "scheme_name": "Pseudo_MP3_Scheme",
      "scheme_data": [
        {
          "value": "Foo",
          "column_name": "ARTIST"
        },
        {
          "value": "Bar",
          "column_name": "SONGTITLE"
        }

      ]
    }
  ]

}

# I want to be able to do terms / bucket aggregations to get something like this (buckets over META.scheme_name
#<snip>
"buckets": [
        {
          "key": "Pseudo_MP3_Scheme",
          "doc_count": 1
        },
        {
          "key": "Pseudo_EXIF_Scheme",
          "doc_count": 1
        }
]
#<snip>

#and also over specific values in the nested doc eg for 'Artist'

"buckets": [
        {
          "key": "Foo",
          "doc_count": 1
        }
]

#But I'm totally stuck on the syntax for aggregating nested documents like this and have tried for several hours without any success.
#Thanks in advance for your help!