The following text is an extract from the Book by Lama Anagarika Govinda, titled “Foundations of Tibetan Mysticism” p287 – 288.

represents in Sanskrit the voiceless sound of breathing out, and is called visarga. In Tibetan the sign corresponding to it is not spoken and merely serves as basis for a vowel or as an indication that the ‘a’ sound inherent in the previous consonant is lengthened into a full ‘ā’ (transcribed as aḥ).

(only in Sanskrit and Pāli), called anusvāra, nasalises the preceding vowel, and is pronounced like the English ‘ng’ in ‘long’ or as the humming after-sound of ‘m’ (as for instance in OṀ).

corresponds to the English ‘ng’ (as above). In Sanskrit and Pāli it is used only within the word, in Tibetan also as termination.

In all aspirates the ‘h’ following the consonant is audibly pronounced. th should never be pronounced like the English ‘th’, but as two distinct sounds, like ‘t’ and ‘h’ in ‘rat-hole’. Similarly: ph like in ‘sap-head’, dh like in ‘mad-house’, kh like in ‘block-head’, jh like in ‘sledge-hammer’.

c corresponds to the English ‘ch’, like in ‘church’.
ch to ‘ch-h’ in ‘match-head’.
j corresponds to the the English ‘j’, as in ‘jar’.

ñ corresponds to the initial sound in ‘new, or the Spanish ‘ñ’ as in ‘mañana’.

The cerebral consonants ṭ, ṭh, ḍ, ḍh, ṇ require the tongue to be placed against the roof of the mouth, while in the case of the dentals (t, th, d, dh, n) the tongue touches the teeth.

ś represents a sharp palatal like a forcefully pronounced ‘sh’.
is a soft cerebral ‘sh’, while s corresponds to a sharp ‘s’ or ‘ss’, as in ‘cross’.