Sanskrit ,Hindi , Telugu ,etc. Pdf to speech (mp3) converters
Munagala Sivaprasad
BVK Sastry (G-S-Pop)
Namaste Munagala Sivaprasad
On < I want to know whether there exists a technology of such as Pdf(texts) to audio format or read aloud format for Indian languages like sanskrit , hindi , tamil , telugu etc. >
You have three complex questions bundled in one line.
1. ‘Technology Existence’ is different from (a) Technology Availability at Affordable cost (b) ‘Design- Delivery Ingenuity in solution delivery.
Fallacy of current solutions being delivered are anchored to the faulty assumption : < English language / foreign language model Tech- Soft-Solutions for Text –readers/ Speech to Text serve equally good for Indian Languages like Sanskrit…... >.
In other words The LLM used for Samskruth and Indian Language Tech solutions suffer from deep rooted fallacies. given to TECH DEVELOPERS by Colonial Academia, as ‘Tower of Babel- Linguistic History’. :
a) ‘ Sanskrit is one more human historic social- societal language like English and other foreign languages.
[List of ISO 639 language codes - Wikipedia - https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes ]
b) Sanskrit SPOKEN LANGUAGE Solutions are deliverable from DEVANAGARI- SCRIPT VISUAL MAPPED TO ROMANIZED TRANSLITERATION SCHEMA and IPA VOICE Repository ( Tech locked in ISO 639).
For more information, explore links: Text to Speech | Web Accessibility Initiative (WAI) | W3C -
https://www.w3.org/WAI/perspective-videos/speech/
2. The question addressed, solution paths, market and expected goals are different .
Again, on ‘whether there exists a technology’ – part of your post: ‘Solutions available as <collated -co-existence of scripted and voiced text >.
Here, we see two solution models (used for making Manuscript digital editions; similar to ‘subscripting video with different language texts).
2.1. Tagged- File- Solutions: Human Read file format tagged to digital visual script: Programmer is tagging human read text to the visual script.
This approach is Mainly seen in archives – where one can here the text and see the text in two separate –disconnected lanes.
This is currently adapted as ‘ Easy Cheap’ solution’ in many ‘School Texts’ as a part of Digital Literacy !
Easy/ Cheap/ Free - Practical Solution is TOTALLY DISCONECTED APPROACH from LINGUISTICALLY RIGHT /RELEVANT/APPROPRIATE – Solution.
It is for user to figure out the Visual script and Voice connection. These are simple ILLUSIONS OF Tech-Solution Illusions
[How the programmer links single word or letter or sentence for VOICE – Hypertext linking is contextual. But basic logic is same.
Screen –Read solutions are NOT HUMAN- PROCESS PROGRAMMING BY HUMAN LANGUAGE MODEL.
The option to live by conscious choice of ‘Easy/ Cheap/ Free - Practical Solution has its consequences on downgrading Human assessed
Language Standards Quality. If One keeps eyes open, this is seen all around. [ A warning which Patanjali seems to allude in the statement:
‘adyatve tu ….. tvaritaa – vaktaarah….].
2.2. Generated Output - File- Solutions: Cross lingual Text Readers (aka Screen Readers): Machine Read OUTPUT programmatically generated
from a digital Character sequence. = Voicing the Key stroke ( from an old generation technology used in calculators).
Programmer generate a Primary ‘Pre-programmed Voice for a given script- character sequence ; Machines Output Phoneme-Voice Tagged to
The character. Tech-Solution Illusions for HUMAN LANGAUGE is created by speed of machine, large data base, and word- sentence Voice
Tagging.
Technology wise- WORD-MEANING issue is not in this domain of Solution design.
3. Coming to last part : < read aloud format for Indian languages like > . There seems to be no solutions of right model, which implement recognizing the ‘ LANGUAGE – PHONEME – VOCING DIFFERENCE’ by LANGAUGE SHIFT, when ‘SAME TEXT-SCRIPT ( Say in Devangari, Tamil or Hindi or Bengali) is Machine Read.
In the current model ( going by framework marked above as ‘Tagged or Generated’), MANY INDIAN LANGUAGE SOLUTIONs are delivering ‘Anglicized Machine Generated Phonemes’ and claiming that to be A.I – Indian Languages, derived from ‘Sanskrit’ .!!??
The delivered Solutions Serves as a < Easy/ Cheap/ Free - Practical Market -Solution > for laity;
But result in systematically damaging LINGUISTICALLY < RIGHT /RELEVANT/APPROPRIATE Human Classical Language Standards > OF Text
READING. There is a compromised situation, Systemically and Systematically, pushing ‘Statistical usage proliferation’ over Classical (Mantra-
Yoga) Brahmi Language ‘Shaastra-Standard’ .
Statistical usage proliferation= Praakruta – jana- bhashaa Vyavahara- bhooyishtataa ;
Classical (Mantra- Yoga) Brahmi Language Standard = Human Systemic – Semiotic Standard of Thought - Articulation = Shaastra-Standard = Samskrutha – Uccharanam = Vak- Yoga -Darshanam, Varna-Aksharam = ‘Upadeshaah – Shuddhah – Pathayante’ as Patanjali says.
Regards
BVK Sastry
--
You received this message because you are subscribed to the Google Groups "भारतीयविद्वत्परिषत्" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bvparishat+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/bvparishat/CAAYz2oOg1LaOzVboosFZxbK9cR7XH5hLdHhfRVbo9s_Fn%3DungQ%40mail.gmail.com.