diff --git a/cmd/revid-cli/main.go b/cmd/revid-cli/main.go
index 1c2b9c90..ac02572c 100644
--- a/cmd/revid-cli/main.go
+++ b/cmd/revid-cli/main.go
@@ -37,6 +37,8 @@ import (
 	"time"
 
 	"bitbucket.org/ausocean/av/revid"
+	"bitbucket.org/ausocean/av/stream/mts"
+	"bitbucket.org/ausocean/av/stream/mts/meta"
 	"bitbucket.org/ausocean/iot/pi/netsender"
 	"bitbucket.org/ausocean/iot/pi/smartlogger"
 	"bitbucket.org/ausocean/utils/logger"
@@ -72,7 +74,14 @@ var canProfile = true
 // The logger that will be used throughout
 var log *logger.Logger
 
+const (
+	metaPreambleKey  = "copyright"
+	metaPreambleData = "ausocean.org/license/content2019"
+)
+
 func main() {
+	mts.Meta = meta.NewWith([][2]string{{metaPreambleKey, metaPreambleData}})
+
 	useNetsender := flag.Bool("NetSender", false, "Are we checking vars through netsender?")
 	runDurationPtr := flag.Duration("runDuration", defaultRunDuration, "How long do you want revid to run for?")
 
diff --git a/revid/revid.go b/revid/revid.go
index aa81e8fd..6d99ac76 100644
--- a/revid/revid.go
+++ b/revid/revid.go
@@ -51,7 +51,7 @@ import (
 
 // Ring buffer sizes and read/write timeouts.
 const (
-	ringBufferSize        = 10000
+	ringBufferSize        = 100
 	ringBufferElementSize = 150000
 	writeTimeout          = 10 * time.Millisecond
 	readTimeout           = 10 * time.Millisecond
diff --git a/revid/senders.go b/revid/senders.go
index a73523d5..e86cb24b 100644
--- a/revid/senders.go
+++ b/revid/senders.go
@@ -35,6 +35,7 @@ import (
 	"net"
 	"os"
 	"os/exec"
+	"strconv"
 
 	"bitbucket.org/ausocean/av/rtmp"
 	"bitbucket.org/ausocean/av/stream/mts"
@@ -172,7 +173,7 @@ func (s *httpSender) extractMeta(r string) error {
 		s.log(logger.Warning, pkg+"No timestamp in reply")
 	} else {
 		s.log(logger.Debug, fmt.Sprintf("%v got timestamp: %v", pkg, t))
-		mts.MetaData.SetTimeStamp(uint64(t))
+		mts.Meta.Add("ts", strconv.Itoa(t))
 	}
 
 	// Extract location from reply
@@ -181,7 +182,7 @@ func (s *httpSender) extractMeta(r string) error {
 		s.log(logger.Warning, pkg+"No location in reply")
 	} else {
 		s.log(logger.Debug, fmt.Sprintf("%v got location: %v", pkg, g))
-		mts.MetaData.SetLocation(g)
+		mts.Meta.Add("loc", g)
 	}
 
 	return nil
diff --git a/stream/mts/meta/meta.go b/stream/mts/meta/meta.go
new file mode 100644
index 00000000..66790315
--- /dev/null
+++ b/stream/mts/meta/meta.go
@@ -0,0 +1,185 @@
+/*
+NAME
+  meta.go
+
+DESCRIPTION
+  See Readme.md
+
+AUTHOR
+  Saxon Nelson-Milton <saxon@ausocean.org>
+
+LICENSE
+  meta.go is Copyright (C) 2017-2019 the Australian Ocean Lab (AusOcean)
+
+  It is free software: you can redistribute it and/or modify them
+  under the terms of the GNU General Public License as published by the
+  Free Software Foundation, either version 3 of the License, or (at your
+  option) any later version.
+
+  It is distributed in the hope that it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with revid in gpl.txt. If not, see http://www.gnu.org/licenses.
+*/
+
+package meta
+
+import (
+	"encoding/binary"
+	"errors"
+	"strings"
+	"sync"
+)
+
+// This is the headsize of our metadata string,
+// which is encoded int the data body of a pmt descriptor.
+const headSize = 4
+
+const (
+	majVer = 1
+	minVer = 0
+)
+
+// Indices of bytes for uint16 metadata length.
+const (
+	dataLenIdx = 2
+)
+
+var (
+	errKeyAbsent     = errors.New("Key does not exist in map")
+	errNoHeader      = errors.New("Metadata string does not contain header")
+	errInvalidHeader = errors.New("Metadata string does not contain valid header")
+)
+
+// Metadata provides functionality for the storage and encoding of metadata
+// using a map.
+type Data struct {
+	mu    sync.RWMutex
+	data  map[string]string
+	order []string
+	enc   []byte
+}
+
+// New returns a pointer to a new Metadata.
+func New() *Data {
+	return &Data{
+		data: make(map[string]string),
+		enc: []byte{
+			0x00,                   // Reserved byte
+			(majVer << 4) | minVer, // MS and LS versions
+			0x00,                   // Data len byte1
+			0x00,                   // Data len byte2
+		},
+	}
+}
+
+// NewWith creates a meta.Data and fills map with initial data given. If there
+// is repeated key, then the latter overwrites the prior.
+func NewWith(data [][2]string) *Data {
+	m := New()
+	m.order = make([]string, 0, len(data))
+	for _, d := range data {
+		if _, exists := m.data[d[0]]; !exists {
+			m.order = append(m.order, d[0])
+		}
+		m.data[d[0]] = d[1]
+	}
+	return m
+}
+
+// Add adds metadata with key and val.
+func (m *Data) Add(key, val string) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.data[key] = val
+	for _, k := range m.order {
+		if k == key {
+			return
+		}
+	}
+	m.order = append(m.order, key)
+	return
+}
+
+// All returns the a copy of the map containing the meta data.
+func (m *Data) All() map[string]string {
+	m.mu.Lock()
+	cpy := make(map[string]string)
+	for k, v := range m.data {
+		cpy[k] = v
+	}
+	m.mu.Unlock()
+	return cpy
+}
+
+// Get returns the meta data for the passed key.
+func (m *Data) Get(key string) (val string, ok bool) {
+	m.mu.Lock()
+	val, ok = m.data[key]
+	m.mu.Unlock()
+	return
+}
+
+// Delete deletes a meta entry in the map and returns error if it doesn’t exist.
+func (m *Data) Delete(key string) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if _, ok := m.data[key]; ok {
+		delete(m.data, key)
+		for i, k := range m.order {
+			if k == key {
+				copy(m.order[:i], m.order[i+1:])
+				m.order = m.order[:len(m.order)-1]
+				break
+			}
+		}
+		return
+	}
+	return
+}
+
+// Encode takes the meta data map and encodes into a byte slice with header
+// describing the version, length of data and data in TSV format.
+func (m *Data) Encode() []byte {
+	m.enc = m.enc[:headSize]
+
+	// Iterate over map and append entries, only adding tab if we're not on the
+	// last entry.
+	var entry string
+	for i, k := range m.order {
+		v := m.data[k]
+		entry += k + "=" + v
+		if i+1 < len(m.data) {
+			entry += "\t"
+		}
+	}
+	m.enc = append(m.enc, []byte(entry)...)
+
+	// Calculate and set data length in encoded meta header.
+	dataLen := len(m.enc[headSize:])
+	binary.BigEndian.PutUint16(m.enc[dataLenIdx:dataLenIdx+2], uint16(dataLen))
+	return m.enc
+}
+
+// ReadFrom extracts a value from a metadata string d, for the given key. If the
+// key is not present in the metadata string, an error is returned. If the
+// metadata header is not present in the string, an error is returned.
+func Extract(key string, d []byte) (string, error) {
+	if d[0] != 0 {
+		return "", errNoHeader
+	} else if d[0] == 0 && binary.BigEndian.Uint16(d[2:headSize]) != uint16(len(d[headSize:])) {
+		return "", errInvalidHeader
+	}
+	d = d[headSize:]
+	entries := strings.Split(string(d), "\t")
+	for _, entry := range entries {
+		kv := strings.Split(entry, "=")
+		if kv[0] == key {
+			return kv[1], nil
+		}
+	}
+	return "", errKeyAbsent
+}
diff --git a/stream/mts/meta/meta_test.go b/stream/mts/meta/meta_test.go
new file mode 100644
index 00000000..e1f9f3b7
--- /dev/null
+++ b/stream/mts/meta/meta_test.go
@@ -0,0 +1,167 @@
+/*
+NAME
+  meta_test.go
+
+DESCRIPTION
+  See Readme.md
+
+AUTHOR
+  Saxon Nelson-Milton <saxon@ausocean.org>
+
+LICENSE
+  meta_test.go is Copyright (C) 2017-2019 the Australian Ocean Lab (AusOcean)
+
+  It is free software: you can redistribute it and/or modify them
+  under the terms of the GNU General Public License as published by the
+  Free Software Foundation, either version 3 of the License, or (at your
+  option) any later version.
+
+  It is distributed in the hope that it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with revid in gpl.txt. If not, see http://www.gnu.org/licenses.
+*/
+
+package meta
+
+import (
+	"bytes"
+	"encoding/binary"
+	"reflect"
+	"testing"
+)
+
+const (
+	tstKey1  = "loc"
+	tstData1 = "a,b,c"
+	tstKey2  = "ts"
+	tstData2 = "12345678"
+	tstData3 = "d,e,f"
+)
+
+// TestAddAndGet ensures that we can add metadata and then successfully get it.
+func TestAddAndGet(t *testing.T) {
+	meta := New()
+	meta.Add(tstKey1, tstData1)
+	meta.Add(tstKey2, tstData2)
+	if data, ok := meta.Get(tstKey1); !ok {
+		t.Errorf("Could not get data for key: %v\n", tstKey1)
+		if data != tstData1 {
+			t.Error("Did not get expected data")
+		}
+	}
+
+	if data, ok := meta.Get(tstKey2); !ok {
+		t.Errorf("Could not get data for key: %v", tstKey2)
+		if data != tstData2 {
+			t.Error("Did not get expected data")
+		}
+	}
+}
+
+// TestUpdate checks that we can use Meta.Add to actually update metadata
+// if it already exists in the Meta map.
+func TestUpdate(t *testing.T) {
+	meta := New()
+	meta.Add(tstKey1, tstData1)
+	meta.Add(tstKey1, tstData3)
+
+	if data, ok := meta.Get(tstKey1); !ok {
+		t.Errorf("Could not get data for key: %v\n", tstKey1)
+		if data != tstData2 {
+			t.Error(`Data did not correctly update for key "loc"`)
+		}
+	}
+}
+
+// TestAll ensures we can get a correct map using Meta.All() after adding some data
+func TestAll(t *testing.T) {
+	meta := New()
+	tstMap := map[string]string{
+		tstKey1: tstData1,
+		tstKey2: tstData2,
+	}
+
+	meta.Add(tstKey1, tstData1)
+	meta.Add(tstKey2, tstData2)
+	metaMap := meta.All()
+
+	if !reflect.DeepEqual(metaMap, tstMap) {
+		t.Errorf("Map not correct. Got: %v, want: %v", metaMap, tstMap)
+	}
+}
+
+// TestGetAbsentKey ensures that we get the expected error when we try to get with
+// key that does not yet exist in the Meta map.
+func TestGetAbsentKey(t *testing.T) {
+	meta := New()
+
+	if _, ok := meta.Get(tstKey1); ok {
+		t.Error("Get for absent key incorrectly returned'ok'")
+	}
+}
+
+// TestDelete ensures we can remove a data entry in the Meta map.
+func TestDelete(t *testing.T) {
+	meta := New()
+	meta.Add(tstKey1, tstData1)
+	meta.Delete(tstKey1)
+	if _, ok := meta.Get(tstKey1); ok {
+		t.Error("Get incorrectly returned okay for absent key")
+	}
+}
+
+// TestEncode checks that we're getting the correct byte slice from Meta.Encode().
+func TestEncode(t *testing.T) {
+	meta := New()
+	meta.Add(tstKey1, tstData1)
+	meta.Add(tstKey2, tstData2)
+
+	dataLen := len(tstKey1+tstData1+tstKey2+tstData2) + 3
+	header := [4]byte{
+		0x00,
+		0x10,
+	}
+	binary.BigEndian.PutUint16(header[2:4], uint16(dataLen))
+	expectedOut := append(header[:], []byte(
+		tstKey1+"="+tstData1+"\t"+
+			tstKey2+"="+tstData2)...)
+
+	got := meta.Encode()
+	if !bytes.Equal(expectedOut, got) {
+		t.Errorf("Did not get expected out. \nGot : %v, \nwant: %v\n", got, expectedOut)
+	}
+}
+
+// TestReadFrom checks that we can correctly obtain a value for a partiular key
+// from a string of metadata using the ReadFrom func.
+func TestReadFrom(t *testing.T) {
+	tstMeta := append([]byte{0x00, 0x10, 0x00, 0x12}, "loc=a,b,c\tts=12345"...)
+
+	tests := []struct {
+		key  string
+		want string
+	}{
+		{
+			"loc",
+			"a,b,c",
+		},
+		{
+			"ts",
+			"12345",
+		},
+	}
+
+	for _, test := range tests {
+		got, err := Extract(test.key, []byte(tstMeta))
+		if err != nil {
+			t.Errorf("Unexpected err: %v\n", err)
+		}
+		if got != test.want {
+			t.Errorf("Did not get expected out. \nGot : %v, \nwant: %v\n", got, test.want)
+		}
+	}
+}
diff --git a/stream/mts/metaEncode_test.go b/stream/mts/metaEncode_test.go
new file mode 100644
index 00000000..e970b7c8
--- /dev/null
+++ b/stream/mts/metaEncode_test.go
@@ -0,0 +1,102 @@
+/*
+NAME
+  metaEncode_test.go
+
+DESCRIPTION
+  See Readme.md
+
+AUTHOR
+  Saxon Nelson-Milton <saxon@ausocean.org>
+
+LICENSE
+  metaEncode_test.go is Copyright (C) 2017-2019 the Australian Ocean Lab (AusOcean)
+
+  It is free software: you can redistribute it and/or modify them
+  under the terms of the GNU General Public License as published by the
+  Free Software Foundation, either version 3 of the License, or (at your
+  option) any later version.
+
+  It is distributed in the hope that it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with revid in gpl.txt. If not, see http://www.gnu.org/licenses.
+*/
+
+package mts
+
+import (
+	"bytes"
+	"testing"
+
+	"bitbucket.org/ausocean/av/stream/mts/meta"
+	"bitbucket.org/ausocean/av/stream/mts/psi"
+)
+
+const (
+	errNotExpectedOut = "Unexpected output. \n Got : %v\n, Want: %v\n"
+	errUnexpectedErr  = "Unexpected error: %v\n"
+)
+
+const fps = 25
+
+// TestMetaEncode1 checks that we can externally add a single metadata entry to
+// the mts global Meta meta.Data struct and then successfully have the mts encoder
+// write this to psi.
+func TestMetaEncode1(t *testing.T) {
+	Meta = meta.New()
+	var b []byte
+	buf := bytes.NewBuffer(b)
+	e := NewEncoder(buf, fps)
+	Meta.Add("ts", "12345678")
+	if err := e.writePSI(); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	out := buf.Bytes()
+	got := out[PacketSize+4:]
+
+	want := []byte{
+		0x00, 0x02, 0xb0, 0x23, 0x00, 0x01, 0xc1, 0x00, 0x00, 0xe1, 0x00, 0xf0, 0x11,
+		psi.MetadataTag,                                                               // Descriptor tag
+		0x0f,                                                                          // Length of bytes to follow
+		0x00, 0x10, 0x00, 0x0b, 't', 's', '=', '1', '2', '3', '4', '5', '6', '7', '8', // timestamp
+		0x1b, 0xe1, 0x00, 0xf0, 0x00,
+	}
+	want = psi.AddCrc(want)
+	want = psi.AddPadding(want)
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestMetaEncode2 checks that we can externally add two metadata entries to the
+// Meta meta.Data global and then have the mts encoder successfully encode this
+// into psi.
+func TestMetaEncode2(t *testing.T) {
+	Meta = meta.New()
+	var b []byte
+	buf := bytes.NewBuffer(b)
+	e := NewEncoder(buf, fps)
+	Meta.Add("ts", "12345678")
+	Meta.Add("loc", "1234,4321,1234")
+	if err := e.writePSI(); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	out := buf.Bytes()
+	got := out[PacketSize+4:]
+	want := []byte{
+		0x00, 0x02, 0xb0, 0x36, 0x00, 0x01, 0xc1, 0x00, 0x00, 0xe1, 0x00, 0xf0, 0x24,
+		psi.MetadataTag,                                                                     // Descriptor tag
+		0x22,                                                                                // Length of bytes to follow
+		0x00, 0x10, 0x00, 0x1e, 't', 's', '=', '1', '2', '3', '4', '5', '6', '7', '8', '\t', // timestamp
+		'l', 'o', 'c', '=', '1', '2', '3', '4', ',', '4', '3', '2', '1', ',', '1', '2', '3', '4', // location
+		0x1b, 0xe1, 0x00, 0xf0, 0x00,
+	}
+	want = psi.AddCrc(want)
+	want = psi.AddPadding(want)
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
diff --git a/stream/mts/mpegts.go b/stream/mts/mpegts.go
index 0bef80d2..08791af7 100644
--- a/stream/mts/mpegts.go
+++ b/stream/mts/mpegts.go
@@ -32,11 +32,41 @@ import (
 	"errors"
 )
 
+// General mpegts packet properties.
 const (
 	PacketSize  = 188
 	PayloadSize = 176
 )
 
+// Program ID for various types of ts packets.
+const (
+	SdtPid   = 17
+	PatPid   = 0
+	PmtPid   = 4096
+	VideoPid = 256
+)
+
+// StreamID is the id of the first stream.
+const StreamID = 0xe0
+
+// HeadSize is the size of an mpegts packet header.
+const HeadSize = 4
+
+// Consts relating to adaptation field.
+const (
+	AdaptationIdx         = 4                 // Index to the adaptation field (index of AFL).
+	AdaptationControlIdx  = 3                 // Index to octet with adaptation field control.
+	AdaptationFieldsIdx   = AdaptationIdx + 1 // Adaptation field index is the index of the adaptation fields.
+	DefaultAdaptationSize = 2                 // Default size of the adaptation field.
+	AdaptationControlMask = 0x30              // Mask for the adaptation field control in octet 3.
+)
+
+// TODO: make this better - currently doesn't make sense.
+const (
+	HasPayload         = 0x1
+	HasAdaptationField = 0x2
+)
+
 /*
 The below data struct encapsulates the fields of an MPEG-TS packet. Below is
 the formatting of an MPEG-TS packet for reference!
diff --git a/stream/mts/psi/crc.go b/stream/mts/psi/crc.go
index a361307d..e0ac7deb 100644
--- a/stream/mts/psi/crc.go
+++ b/stream/mts/psi/crc.go
@@ -34,16 +34,16 @@ import (
 )
 
 // addCrc appends a crc table to a given psi table in bytes
-func addCrc(out []byte) []byte {
+func AddCrc(out []byte) []byte {
 	t := make([]byte, len(out)+4)
 	copy(t, out)
-	updateCrc(t)
+	UpdateCrc(t[1:])
 	return t
 }
 
 // updateCrc updates the crc of bytes slice, writing the checksum into the last four bytes.
-func updateCrc(b []byte) {
-	crc32 := crc32_Update(0xffffffff, crc32_MakeTable(bits.Reverse32(crc32.IEEE)), b[1:len(b)-4])
+func UpdateCrc(b []byte) {
+	crc32 := crc32_Update(0xffffffff, crc32_MakeTable(bits.Reverse32(crc32.IEEE)), b[:len(b)-4])
 	binary.BigEndian.PutUint32(b[len(b)-4:], crc32)
 }
 
diff --git a/stream/mts/psi/descriptor_test.go b/stream/mts/psi/descriptor_test.go
new file mode 100644
index 00000000..94441277
--- /dev/null
+++ b/stream/mts/psi/descriptor_test.go
@@ -0,0 +1,322 @@
+/*
+NAME
+  descriptor_test.go
+
+DESCRIPTION
+  See Readme.md
+
+AUTHOR
+  Saxon Nelson-Milton <saxon@ausocean.org>
+
+LICENSE
+  descriptor_test.go is Copyright (C) 2017-2019 the Australian Ocean Lab (AusOcean)
+
+  It is free software: you can redistribute it and/or modify them
+  under the terms of the GNU General Public License as published by the
+  Free Software Foundation, either version 3 of the License, or (at your
+  option) any later version.
+
+  It is distributed in the hope that it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with revid in gpl.txt.  If not, see http://www.gnu.org/licenses.
+*/
+
+package psi
+
+import (
+	"bytes"
+	"testing"
+)
+
+const (
+	errNotExpectedOut = "Did not get expected output: \ngot : %v, \nwant: %v"
+	errUnexpectedErr  = "Unexpected error: %v\n"
+)
+
+var (
+	tstPsi1 = PSI{
+		Pf:  0x00,
+		Tid: 0x02,
+		Ssi: true,
+		Sl:  0x1c,
+		Tss: &TSS{
+			Tide: 0x01,
+			V:    0,
+			Cni:  true,
+			Sn:   0,
+			Lsn:  0,
+			Sd: &PMT{
+				Pcrpid: 0x0100, // wrong
+				Pil:    10,
+				Pd: []Desc{
+					{
+						Dt: TimeDescTag,
+						Dl: TimeDataSize,
+						Dd: make([]byte, TimeDataSize),
+					},
+				},
+				Essd: &ESSD{
+					St:   0x1b,
+					Epid: 0x0100,
+					Esil: 0x00,
+				},
+			},
+		},
+	}
+
+	tstPsi2 = PSI{
+		Pf:  0x00,
+		Tid: 0x02,
+		Ssi: true,
+		Sl:  0x12,
+		Tss: &TSS{
+			Tide: 0x01,
+			V:    0,
+			Cni:  true,
+			Sn:   0,
+			Lsn:  0,
+			Sd: &PMT{
+				Pcrpid: 0x0100,
+				Pil:    0,
+				Essd: &ESSD{
+					St:   0x1b,
+					Epid: 0x0100,
+					Esil: 0x00,
+				},
+			},
+		},
+	}
+
+	tstPsi3 = PSI{
+		Pf:  0x00,
+		Tid: 0x02,
+		Ssi: true,
+		Sl:  0x3e,
+		Tss: &TSS{
+			Tide: 0x01,
+			V:    0,
+			Cni:  true,
+			Sn:   0,
+			Lsn:  0,
+			Sd: &PMT{
+				Pcrpid: 0x0100,
+				Pil:    PmtTimeLocationPil,
+				Pd: []Desc{
+					{
+						Dt: TimeDescTag,
+						Dl: TimeDataSize,
+						Dd: make([]byte, TimeDataSize),
+					},
+					{
+						Dt: LocationDescTag,
+						Dl: LocationDataSize,
+						Dd: make([]byte, LocationDataSize),
+					},
+				},
+				Essd: &ESSD{
+					St:   0x1b,
+					Epid: 0x0100,
+					Esil: 0x00,
+				},
+			},
+		},
+	}
+)
+
+var (
+	pmtTimeBytesResizedBigger = []byte{
+		0x00, 0x02, 0xb0, 0x1e, 0x00, 0x01, 0xc1, 0x00, 0x00, 0xe1, 0x00, 0xf0, 0x0c,
+		TimeDescTag,                                                // Descriptor tag
+		0x0a,                                                       // Length of bytes to follow
+		0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, // timestamp
+		0x1b, 0xe1, 0x00, 0xf0, 0x00,
+	}
+
+	pmtTimeBytesResizedSmaller = []byte{
+		0x00, 0x02, 0xb0, 0x1a, 0x00, 0x01, 0xc1, 0x00, 0x00, 0xe1, 0x00, 0xf0, 0x08,
+		TimeDescTag,                        // Descriptor tag
+		0x06,                               // Length of bytes to follow
+		0x01, 0x02, 0x03, 0x04, 0x05, 0x06, // timestamp
+		0x1b, 0xe1, 0x00, 0xf0, 0x00,
+	}
+)
+
+// TestHasDescriptorExists checks that PSIBytes.HasDescriptor performs as expected
+// when the PSI we're interested in has the descriptor of interest. HasDescriptor
+// should return the descriptor bytes.
+// TODO: HasDescriptor also returns index of descriptor - we should check this.
+func TestHasDescriptorExists(t *testing.T) {
+	p := PSIBytes(tstPsi3.Bytes())
+	_, got := p.HasDescriptor(LocationDescTag)
+	want := []byte{
+		LocationDescTag,
+		LocationDataSize,
+	}
+	want = append(want, make([]byte, LocationDataSize)...)
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestHasDescriptorAbsent checks that PSIBytes.HasDescriptor performs as expected
+// when the PSI does not have the descriptor of interest. HasDescriptor should
+// return a nil slice and a negative index.
+// TODO: check index here as well.
+func TestHasDescriptorAbsent(t *testing.T) {
+	p := PSIBytes(tstPsi3.Bytes())
+	const fakeTag = 236
+	_, got := p.HasDescriptor(fakeTag)
+	var want []byte
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestHasDescriptorNone checks that PSIBytes.HasDescriptor behaves as expected
+// when the PSI does not have any descriptors. HasDescriptor should return a nil
+// slice.
+// TODO: again check index here.
+func TestHasDescriptorNone(t *testing.T) {
+	p := PSIBytes(tstPsi2.Bytes())
+	_, got := p.HasDescriptor(LocationDescTag)
+	var want []byte
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestProgramInfoLen checks that PSIBytes.ProgramInfoLen correctly extracts
+// the program info length from a PSI.
+func TestProgramInfoLen(t *testing.T) {
+	p := PSIBytes(tstPsi1.Bytes())
+	got := p.ProgramInfoLen()
+	want := 10
+	if got != want {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestDescriptors checks that PSIBytes.descriptors correctly returns the descriptors
+// from a PSI when descriptors exist.
+func TestDescriptors(t *testing.T) {
+	p := PSIBytes(tstPsi1.Bytes())
+	got := p.descriptors()
+	want := []byte{
+		TimeDescTag,
+		TimeDataSize,
+	}
+	want = append(want, make([]byte, TimeDataSize)...)
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestDescriptors checks that PSIBYtes.desriptors correctly returns nil when
+// we try to get descriptors from a psi without any descriptors.
+func TestDescriptorsNone(t *testing.T) {
+	p := PSIBytes(tstPsi2.Bytes())
+	got := p.descriptors()
+	var want []byte
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestCreateDescriptorEmpty checks that PSIBytes.createDescriptor correctly adds
+// a descriptor to the descriptors list in a PSI when it has no descriptors already.
+func TestCreateDescriptorEmpty(t *testing.T) {
+	got := PSIBytes(tstPsi2.Bytes())
+	got.createDescriptor(TimeDescTag, make([]byte, TimeDataSize))
+	UpdateCrc(got[1:])
+	want := PSIBytes(tstPsi1.Bytes())
+	if !bytes.Equal(want, got) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestCreateDescriptorNotEmpty checks that PSIBytes.createDescriptor correctly adds
+// a descriptor to the descriptors list in a PSI when it already has one with
+// a different tag.
+func TestCreateDescriptorNotEmpty(t *testing.T) {
+	got := PSIBytes(tstPsi1.Bytes())
+	got.createDescriptor(LocationDescTag, make([]byte, LocationDataSize))
+	UpdateCrc(got[1:])
+	want := PSIBytes(tstPsi3.Bytes())
+	if !bytes.Equal(want, got) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestAddDescriptorEmpty checks that PSIBytes.AddDescriptor correctly adds a descriptor
+// when there are no other descriptors present in the PSI.
+func TestAddDescriptorEmpty(t *testing.T) {
+	got := PSIBytes(tstPsi2.Bytes())
+	if err := got.AddDescriptor(TimeDescTag, make([]byte, TimeDataSize)); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	want := PSIBytes(tstPsi1.Bytes())
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestAddDescriptorNonEmpty checks that PSIBytes.AddDescriptor correctly adds a
+// descriptor when there is already a descriptor of a differing type in a PSI.
+func TestAddDescriptorNonEmpty(t *testing.T) {
+	got := PSIBytes(tstPsi1.Bytes())
+	if err := got.AddDescriptor(LocationDescTag, make([]byte, LocationDataSize)); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	want := PSIBytes(tstPsi3.Bytes())
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestAddDescriptorUpdateSame checks that PSIBytes.AddDescriptor correctly updates data in a descriptor
+// with the same given tag, with data being the same size. AddDescriptor should just copy new data into
+// the descriptors data field.
+func TestAddDescriptorUpdateSame(t *testing.T) {
+	newData := [8]byte{0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08}
+	want := PSIBytes(tstPsi2.Bytes())
+	want.createDescriptor(TimeDescTag, newData[:])
+	got := PSIBytes(tstPsi1.Bytes())
+	if err := got.AddDescriptor(TimeDescTag, newData[:]); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestAddDescriptorUpdateBigger checks that PSIBytes.AddDescriptor correctly resizes descriptor with same given tag
+// to a bigger size and copies in new data. AddDescriptor should find descriptor with same tag, increase size of psi,
+// shift data to make room for update descriptor, and then copy in the new data.
+func TestAddDescriptorUpdateBigger(t *testing.T) {
+	got := PSIBytes(tstPsi1.Bytes())
+	if err := got.AddDescriptor(TimeDescTag, []byte{0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a}); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	want := AddCrc(pmtTimeBytesResizedBigger)
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
+
+// TestAddDescriptorUpdateSmaller checks that PSIBytes.AddDescriptor correctly resizes descriptor with same given tag
+// in a psi to a smaller size and copies in new data. AddDescriptor should find tag with same descrtiptor, shift data
+// after descriptor upwards, trim the psi to new size, and then copy in new data.
+func TestAddDescriptorUpdateSmaller(t *testing.T) {
+	got := PSIBytes(tstPsi1.Bytes())
+	if err := got.AddDescriptor(TimeDescTag, []byte{0x01, 0x02, 0x03, 0x04, 0x05, 0x06}); err != nil {
+		t.Errorf(errUnexpectedErr, err.Error())
+	}
+	want := AddCrc(pmtTimeBytesResizedSmaller)
+	if !bytes.Equal(got, want) {
+		t.Errorf(errNotExpectedOut, got, want)
+	}
+}
diff --git a/stream/mts/psi/helpers.go b/stream/mts/psi/helpers.go
index 82d7ce72..b8bab6b5 100644
--- a/stream/mts/psi/helpers.go
+++ b/stream/mts/psi/helpers.go
@@ -64,15 +64,14 @@ func UpdateTime(dst []byte, t uint64) error {
 	for i := range dst[TimeDataIndx : TimeDataIndx+TimeDataSize] {
 		dst[i+TimeDataIndx] = ts[i]
 	}
-	updateCrc(dst)
+	UpdateCrc(dst[1:])
 	return nil
 }
 
 // SyntaxSecLenFrom takes a byte slice representation of a psi and extracts
 // it's syntax section length
-func SyntaxSecLenFrom(p []byte) (l uint8) {
-	l = uint8(p[syntaxSecLenIndx]) - crcSize
-	return
+func SyntaxSecLenFrom(p []byte) int {
+	return int(((p[SyntaxSecLenIdx1] & SyntaxSecLenMask1) << 8) | p[SyntaxSecLenIdx2])
 }
 
 // TimeFrom takes a byte slice representation of a psi-pmt and extracts it's
@@ -112,7 +111,7 @@ func UpdateLocation(d []byte, s string) error {
 	for i := range loc {
 		loc[i] = 0
 	}
-	updateCrc(d)
+	UpdateCrc(d[1:])
 	return nil
 }
 
@@ -127,7 +126,7 @@ func trimTo(d []byte, t byte) []byte {
 
 // addPadding adds an appropriate amount of padding to a pat or pmt table for
 // addition to an mpegts packet
-func addPadding(d []byte) []byte {
+func AddPadding(d []byte) []byte {
 	t := make([]byte, PacketSize)
 	copy(t, d)
 	padding := t[len(d):]
diff --git a/stream/mts/psi/psi.go b/stream/mts/psi/psi.go
index 09028ba6..c93d3011 100644
--- a/stream/mts/psi/psi.go
+++ b/stream/mts/psi/psi.go
@@ -26,11 +26,16 @@ LICENSE
 
 package psi
 
-const (
-	PacketSize = 184 // packet size of a psi.
+import (
+	"errors"
+
+	"github.com/Comcast/gots/psi"
 )
 
-// Lengths of section definitions
+// PacketSize of psi (without mpegts header)
+const PacketSize = 184
+
+// Lengths of section definitions.
 const (
 	ESSDDefLen = 5
 	DescDefLen = 2
@@ -40,13 +45,14 @@ const (
 	PSIDefLen  = 3
 )
 
-// Table Type IDs
+// Table Type IDs.
 const (
 	patID = 0x00
 	pmtID = 0x02
 )
 
 // Consts relating to time description
+// TODO: remove this, we don't do metadata like this anymore.
 const (
 	TimeDescTag  = 234
 	TimeTagIndx  = 13
@@ -55,6 +61,7 @@ const (
 )
 
 // Consts relating to location description
+// TODO: remove this, we don't do metadata like this anymore.
 const (
 	LocationDescTag  = 235
 	LocationTagIndx  = 23
@@ -62,10 +69,35 @@ const (
 	LocationDataSize = 32 // bytes
 )
 
-// Other misc consts
+// crc hassh Size
+const crcSize = 4
+
+// Consts relating to syntax section.
 const (
-	syntaxSecLenIndx = 3
-	crcSize          = 4
+	TotalSyntaxSecLen = 180
+	SyntaxSecLenIdx1  = 2
+	SyntaxSecLenIdx2  = 3
+	SyntaxSecLenMask1 = 0x03
+	SectionLenMask1   = 0x03
+)
+
+// Consts relating to program info len.
+const (
+	ProgramInfoLenIdx1  = 11
+	ProgramInfoLenIdx2  = 12
+	ProgramInfoLenMask1 = 0x03
+)
+
+// DescriptorsIdx is the index that the descriptors start at.
+const DescriptorsIdx = ProgramInfoLenIdx2 + 1
+
+// MetadataTag is the descriptor tag used for metadata.
+const MetadataTag = 0x26
+
+// TODO: get rid of these - not a good idea.
+type (
+	PSIBytes   []byte
+	Descriptor []byte
 )
 
 // Program specific information
@@ -135,8 +167,7 @@ func (p *PSI) Bytes() []byte {
 	out[2] = 0x80 | 0x30 | (0x03 & byte(p.Sl>>8))
 	out[3] = byte(p.Sl)
 	out = append(out, p.Tss.Bytes()...)
-	out = addCrc(out)
-	out = addPadding(out)
+	out = AddCrc(out)
 	return out
 }
 
@@ -205,3 +236,135 @@ func asByte(b bool) byte {
 	}
 	return 0x00
 }
+
+// AddDescriptor adds or updates a descriptor in a PSI given a descriptor tag
+// and data. If the psi is not a pmt, then an error is returned. If a descriptor
+// with the given tag is not found in the psi, room is made and a descriptor with
+// given tag and data is created. If a descriptor with the tag is found, the
+// descriptor is resized as required and the new data is copied in.
+func (p *PSIBytes) AddDescriptor(tag int, data []byte) error {
+	if psi.TableID(*p) != pmtID {
+		return errors.New("trying to add descriptor, but not pmt")
+	}
+
+	i, desc := p.HasDescriptor(tag)
+	if desc == nil {
+		err := p.createDescriptor(tag, data)
+		return err
+	}
+
+	oldDescLen := desc.len()
+	oldDataLen := int(desc[1])
+	newDataLen := len(data)
+	newDescLen := 2 + newDataLen
+	delta := newDescLen - oldDescLen
+
+	// If the old data length is more than the new data length, we need shift data
+	// after descriptor up, and then trim the psi. If the oldDataLen is less than
+	// new data then we need reseize psi and shift data down. If same do nothing.
+	switch {
+	case oldDataLen > newDataLen:
+		copy((*p)[i+newDescLen:], (*p)[i+oldDescLen:])
+		*p = (*p)[:len(*p)+delta]
+	case oldDataLen < newDataLen:
+		tmp := make([]byte, len(*p)+delta)
+		copy(tmp, *p)
+		*p = tmp
+		copy((*p)[i+newDescLen:], (*p)[i+oldDescLen:])
+	}
+
+	// Copy in new data
+	(*p)[i+1] = byte(newDataLen)
+	copy((*p)[i+2:], data)
+
+	newProgInfoLen := p.ProgramInfoLen() + delta
+	p.setProgInfoLen(newProgInfoLen)
+	newSectionLen := int(psi.SectionLength(*p)) + delta
+	p.setSectionLen(newSectionLen)
+	UpdateCrc((*p)[1:])
+	return nil
+}
+
+// HasDescriptor checks if a descriptor of the given tag exists in a PSI. If the descriptor
+// of the given tag exists, an index of this descriptor, as well as the Descriptor is returned.
+// If the descriptor of the given tag cannot be found, -1 and a nil slice is returned.
+//
+// TODO: check if pmt, return error if not ?
+func (p *PSIBytes) HasDescriptor(tag int) (int, Descriptor) {
+	descs := p.descriptors()
+	if descs == nil {
+		return -1, nil
+	}
+	for i := 0; i < len(descs); i += 2 + int(descs[i+1]) {
+		if int(descs[i]) == tag {
+			return i + DescriptorsIdx, descs[i : i+2+int(descs[i+1])]
+		}
+	}
+	return -1, nil
+}
+
+// createDescriptor creates a descriptor in a psi given a tag and data. It does so
+// by resizing the psi, shifting existing data down and copying in new descriptor
+// in new space.
+func (p *PSIBytes) createDescriptor(tag int, data []byte) error {
+	curProgLen := p.ProgramInfoLen()
+	oldSyntaxSectionLen := SyntaxSecLenFrom(*p)
+	if TotalSyntaxSecLen-(oldSyntaxSectionLen+2+len(data)) <= 0 {
+		return errors.New("Not enough space in psi to create descriptor.")
+	}
+	dataLen := len(data)
+	newDescIdx := DescriptorsIdx + curProgLen
+	newDescLen := dataLen + 2
+
+	// Increase size of psi and copy data down to make room for new descriptor.
+	tmp := make([]byte, len(*p)+newDescLen)
+	copy(tmp, *p)
+	*p = tmp
+	copy((*p)[newDescIdx+newDescLen:], (*p)[newDescIdx:newDescIdx+newDescLen])
+	// Set the tag, data len and data of the new desriptor.
+	(*p)[newDescIdx] = byte(tag)
+	(*p)[newDescIdx+1] = byte(dataLen)
+	copy((*p)[newDescIdx+2:newDescIdx+2+dataLen], data)
+
+	// Set length fields and update the psi crc.
+	addedLen := dataLen + 2
+	newProgInfoLen := curProgLen + addedLen
+	p.setProgInfoLen(newProgInfoLen)
+	newSyntaxSectionLen := int(oldSyntaxSectionLen) + addedLen
+	p.setSectionLen(newSyntaxSectionLen)
+	UpdateCrc((*p)[1:])
+
+	return nil
+}
+
+// setProgInfoLen sets the program information length in a psi with a pmt.
+func (p *PSIBytes) setProgInfoLen(l int) {
+	(*p)[ProgramInfoLenIdx1] &= 0xff ^ ProgramInfoLenMask1
+	(*p)[ProgramInfoLenIdx1] |= byte(l>>8) & ProgramInfoLenMask1
+	(*p)[ProgramInfoLenIdx2] = byte(l)
+}
+
+// setSectionLen sets section length in a psi.
+func (p *PSIBytes) setSectionLen(l int) {
+	(*p)[SyntaxSecLenIdx1] &= 0xff ^ SyntaxSecLenMask1
+	(*p)[SyntaxSecLenIdx1] |= byte(l>>8) & SyntaxSecLenMask1
+	(*p)[SyntaxSecLenIdx2] = byte(l)
+}
+
+// descriptors returns the descriptors in a psi if they exist, otherwise
+// a nil slice is returned.
+func (p *PSIBytes) descriptors() []byte {
+	return (*p)[DescriptorsIdx : DescriptorsIdx+p.ProgramInfoLen()]
+}
+
+// len returns the length of a descriptor in bytes.
+func (d *Descriptor) len() int {
+	return int(2 + (*d)[1])
+}
+
+// ProgramInfoLen returns the program info length of a PSI.
+//
+// TODO: check if pmt - if not return 0 ? or -1 ?
+func (p *PSIBytes) ProgramInfoLen() int {
+	return int((((*p)[ProgramInfoLenIdx1] & ProgramInfoLenMask1) << 8) | (*p)[ProgramInfoLenIdx2])
+}
diff --git a/stream/mts/psi/psi_test.go b/stream/mts/psi/psi_test.go
index e437066e..7e3a3104 100644
--- a/stream/mts/psi/psi_test.go
+++ b/stream/mts/psi/psi_test.go
@@ -282,7 +282,7 @@ var bytesTests = []struct {
 func TestBytes(t *testing.T) {
 	for _, test := range bytesTests {
 		got := test.input.Bytes()
-		if !bytes.Equal(got, addPadding(addCrc(test.want))) {
+		if !bytes.Equal(got, AddCrc(test.want)) {
 			t.Errorf("unexpected error for test %v: got:%v want:%v", test.name, got,
 				test.want)
 		}
@@ -301,7 +301,7 @@ func TestTimestampToBytes(t *testing.T) {
 func TestTimeUpdate(t *testing.T) {
 	cpy := make([]byte, len(pmtTimeBytes1))
 	copy(cpy, pmtTimeBytes1)
-	cpy = addCrc(cpy)
+	cpy = AddCrc(cpy)
 	err := UpdateTime(cpy, tstTime2)
 	cpy = cpy[:len(cpy)-4]
 	if err != nil {
@@ -343,7 +343,7 @@ func TestLocationGet(t *testing.T) {
 func TestLocationUpdate(t *testing.T) {
 	cpy := make([]byte, len(pmtWithMetaTst1))
 	copy(cpy, pmtWithMetaTst1)
-	cpy = addCrc(cpy)
+	cpy = AddCrc(cpy)
 	err := UpdateLocation(cpy, locationTstStr2)
 	cpy = cpy[:len(cpy)-4]
 	if err != nil {