Re: Around India with a Canute: Transcribing Braille into 12 Indian languages in Pune
30 views
Skip to first unread message
E M Rogers
Sep 13, 2017, 9:44:00 AM9/13/17
to Braillists Forum, BBT Reports
% Journal of WCMT supported India trip, 4th entry \\
Around India with a Canute: \\
Transcribing Braille into 12 Indian languages in Pune
% Ed Rogers, Bristol Braille Technology
% 13th of September, 2017; from Anna Centenary Central Library, Chennai
Concerns meeting with Modular Infotech in Pune, over a week ago.
# Background
Briefly deviated from meetings in Bangalore to fly to Pune. These
flights were very generously paid for by Modular Infotech.
I had been talking by email with Raghunandan Joshi, one of the company
founders, for many years, but this was our first face to face meeting,
where we were joined by Shekhar Jadhav and their fellow directors.
He and I had always been enthusiastic about finding a way of combining
our work; Modular specialise in Indian language support in Braille and
screen reader technology.
# Screen reading
MI have developed a Windows screen reader that supports 12 regional
Indian languages. It is based off NVDA by NV Access
# Braille transcription and embossing
They have also developed solutions to transcribe and emboss Braille
paper Braille called Shree-Lipi Braille.
Again, they have their own custom software which generates Braille in
any of the 12 Indian languages, plus English, several varieties of
Arabic and a few more.
This is edited in a conventional looking Braille editing software. They
then sell imported embossers from the States.
They also have developed an optical Braille recognition software, for
scanning in Braille books. This is no mean feat, as BBT learnt in 2014,
when we developed our own basic Braille scnning software.
# Braille displays
MI make and sell two Braille displays; the Mitra 16 and the Mitra 32.
Both single line (I'll leave you to guess the number of cells) and using
Metek piezo-electric cells from Germany.
The form is a metal box about 1" by 10" by 10". The Braille cells are
along to top half of the machine.
It is mains only through a DIN plug on the front. They told me they
thought they would have to make a battery version, but it turns out
people are happy with mains. This challenges an assumption I had about
selling Braille devices in India, which was that internal batteries
would be vital given the frequency of power cuts. And, frankly, it
contradicts my own experiences of the last two weeks, when power cuts
have sometimes been as frequent as six in a day/night.
The only controls are a horizontal row of ten small buttons above the
cells, which one uses to programme in book numbers (books are numbered
and are selected by their number) and page numbers. They also have
specific modal functions for different menus, I gather. Then there are
two more of the same type of buttons, one on either side of the line,
for panning. These buttons have great tactile feel with at least 3mm of
travel.
The design of the user interface software is ---very broadly--- similar
to that taken by the Canute UI team; i.e. a focus on loading in
pre-prepared material and cycling between those files as a stand-alone
reading device. The interesting distinction is that they chose to
navigate the book selection by catalogue number rather than flat file
list, as on the Canute.
However I cannot recall whether it also converts plain text files on the
fly.
The Mitra 32 sells for something like £1,800, which is competitive, but,
given its use of metek cells, not remarkably affordable, especially for
India.
It should be pointed out that MI do not seem to be making a great deal
of profit off these, but rather are doing it to further their aims of
supporting Braille and regional Indian languages.
# Braille libraries
The Mitras are designed to be used in a 'Braille Library' set up. By
this I don't mean any old Braille library, but a specific concept by MI
which they call a Braille Library; that is, five or so Mitras in a room,
with a central computer with a scanner and MI OCR and transcription
software. The PC produces the content, which is then loaded into each of
the Mitras in turn.
They have set up five of these libraries so far. I've not seen a MI
Braille Library in action, though at the moment I am sitting in the
Braille section of the Anna centenary Central Library in Chennai, and
they have a Mitra 32 here. (More on Anna Centenary Library later in the
month.)
# Canute and other foreign refreshable Braille devices in India
MI estimate, given their experience of selling the Mitra, that a Canute
selling for 50,000 Rupees (around £600--700) would sell somewhere in the
region of a dz units a year into India.
Greater sales could be made with a concerted marketing effort, but at
that price it is outside of not only individuals, but even many Govt
schemes to provide equipment for individuals, and the generosity of
various corporate sponsors.
Which is not to say those sponsors couldn't be found for a series one
50kINR Canute, but that it would be far easier to get corporate
sponsorship and onto the books of various Govt schemes if the price were
somehow to drop by around two fifths: 30kINR.
Therefore we discussed likely ways of distributing new Braille
technology into India, whether Canute or other;
- Short term: Import devices to India for high prices, finding small
but influential users in universities &c.
- Medium term: Move production to India (for the Indian market models
at least), thus allowing prices to drop significantly due to absence of
tariffs and lower labour costs, establishing significant presence.
- Long term: Reconsider design to produce a custom device especially
for the Indian market, with the focus on coming close to 10kINR, thus
allowing widespread distribution, even to individuals.
# Specific requirements for Indian Braille readers
Even in the instance of the long term, MI insisted on the importance of
the devices being stand-alone, so they don't require a PC. Makes sense,
but I wonder whether reliance on a mobile phone would be acceptable..?
Something which struck me at some point between meeting MI and today, so
I might as well put it here; so far as I can tell there is little
benefit in 40 cells in the Indian market. There doesn't seem to be
back-catalogues of BRF files, which is the only reason for it. So for
the 'longer term' discussed above, the fictional device could be 20, 24
or 28 cells.
Only thing to counter that is that some Indian regional languages have
much longer words than contracted English Braille; sometimes in the
region of 20 cells.
# Future
I look forward to talking with MI when back in the UK about how their
expertise in Indian languages, and existing refreshable Braille
distribution in India, can be successfully coupled with BBT's technology
and the Braillists Foundation's aims.
MI were interested in having the Mitra tested by the Braillists in the
UK. I will look into this further when home.
Also, I want to emphasise again the generosity of MI for paying for my
flights to and from Pune, and for unexpectedly picking me up from the
airport. Thanks!
⠠⠃⠗⠊⠌⠕⠇⠀⠠⠃⠗⠇⠀⠠⠞⠑⠡⠝⠕⠇⠕⠛⠽⠀⠠⠞⠑⠇⠀⠼⠚⠛⠊⠚⠓⠑⠋⠊⠃⠁⠙
| Bristol Braille Technology CIC | Tel: +447908 569 214 |
Around India with a Canute: \\
Transcribing Braille into 12 Indian languages in Pune
% Ed Rogers, Bristol Braille Technology
% 13th of September, 2017; from Anna Centenary Central Library, Chennai
Concerns meeting with Modular Infotech in Pune, over a week ago.
# Background
Briefly deviated from meetings in Bangalore to fly to Pune. These
flights were very generously paid for by Modular Infotech.
I had been talking by email with Raghunandan Joshi, one of the company
founders, for many years, but this was our first face to face meeting,
where we were joined by Shekhar Jadhav and their fellow directors.
He and I had always been enthusiastic about finding a way of combining
our work; Modular specialise in Indian language support in Braille and
screen reader technology.
# Screen reading
MI have developed a Windows screen reader that supports 12 regional
Indian languages. It is based off NVDA by NV Access
# Braille transcription and embossing
They have also developed solutions to transcribe and emboss Braille
paper Braille called Shree-Lipi Braille.
Again, they have their own custom software which generates Braille in
any of the 12 Indian languages, plus English, several varieties of
Arabic and a few more.
This is edited in a conventional looking Braille editing software. They
then sell imported embossers from the States.
They also have developed an optical Braille recognition software, for
scanning in Braille books. This is no mean feat, as BBT learnt in 2014,
when we developed our own basic Braille scnning software.
# Braille displays
MI make and sell two Braille displays; the Mitra 16 and the Mitra 32.
Both single line (I'll leave you to guess the number of cells) and using
Metek piezo-electric cells from Germany.
The form is a metal box about 1" by 10" by 10". The Braille cells are
along to top half of the machine.
It is mains only through a DIN plug on the front. They told me they
thought they would have to make a battery version, but it turns out
people are happy with mains. This challenges an assumption I had about
selling Braille devices in India, which was that internal batteries
would be vital given the frequency of power cuts. And, frankly, it
contradicts my own experiences of the last two weeks, when power cuts
have sometimes been as frequent as six in a day/night.
The only controls are a horizontal row of ten small buttons above the
cells, which one uses to programme in book numbers (books are numbered
and are selected by their number) and page numbers. They also have
specific modal functions for different menus, I gather. Then there are
two more of the same type of buttons, one on either side of the line,
for panning. These buttons have great tactile feel with at least 3mm of
travel.
The design of the user interface software is ---very broadly--- similar
to that taken by the Canute UI team; i.e. a focus on loading in
pre-prepared material and cycling between those files as a stand-alone
reading device. The interesting distinction is that they chose to
navigate the book selection by catalogue number rather than flat file
list, as on the Canute.
However I cannot recall whether it also converts plain text files on the
fly.
The Mitra 32 sells for something like £1,800, which is competitive, but,
given its use of metek cells, not remarkably affordable, especially for
India.
It should be pointed out that MI do not seem to be making a great deal
of profit off these, but rather are doing it to further their aims of
supporting Braille and regional Indian languages.
# Braille libraries
The Mitras are designed to be used in a 'Braille Library' set up. By
this I don't mean any old Braille library, but a specific concept by MI
which they call a Braille Library; that is, five or so Mitras in a room,
with a central computer with a scanner and MI OCR and transcription
software. The PC produces the content, which is then loaded into each of
the Mitras in turn.
They have set up five of these libraries so far. I've not seen a MI
Braille Library in action, though at the moment I am sitting in the
Braille section of the Anna centenary Central Library in Chennai, and
they have a Mitra 32 here. (More on Anna Centenary Library later in the
month.)
# Canute and other foreign refreshable Braille devices in India
MI estimate, given their experience of selling the Mitra, that a Canute
selling for 50,000 Rupees (around £600--700) would sell somewhere in the
region of a dz units a year into India.
Greater sales could be made with a concerted marketing effort, but at
that price it is outside of not only individuals, but even many Govt
schemes to provide equipment for individuals, and the generosity of
various corporate sponsors.
Which is not to say those sponsors couldn't be found for a series one
50kINR Canute, but that it would be far easier to get corporate
sponsorship and onto the books of various Govt schemes if the price were
somehow to drop by around two fifths: 30kINR.
Therefore we discussed likely ways of distributing new Braille
technology into India, whether Canute or other;
- Short term: Import devices to India for high prices, finding small
but influential users in universities &c.
- Medium term: Move production to India (for the Indian market models
at least), thus allowing prices to drop significantly due to absence of
tariffs and lower labour costs, establishing significant presence.
- Long term: Reconsider design to produce a custom device especially
for the Indian market, with the focus on coming close to 10kINR, thus
allowing widespread distribution, even to individuals.
# Specific requirements for Indian Braille readers
Even in the instance of the long term, MI insisted on the importance of
the devices being stand-alone, so they don't require a PC. Makes sense,
but I wonder whether reliance on a mobile phone would be acceptable..?
Something which struck me at some point between meeting MI and today, so
I might as well put it here; so far as I can tell there is little
benefit in 40 cells in the Indian market. There doesn't seem to be
back-catalogues of BRF files, which is the only reason for it. So for
the 'longer term' discussed above, the fictional device could be 20, 24
or 28 cells.
Only thing to counter that is that some Indian regional languages have
much longer words than contracted English Braille; sometimes in the
region of 20 cells.
# Future
I look forward to talking with MI when back in the UK about how their
expertise in Indian languages, and existing refreshable Braille
distribution in India, can be successfully coupled with BBT's technology
and the Braillists Foundation's aims.
MI were interested in having the Mitra tested by the Braillists in the
UK. I will look into this further when home.
Also, I want to emphasise again the generosity of MI for paying for my
flights to and from Pune, and for unexpectedly picking me up from the
airport. Thanks!
⠠⠃⠗⠊⠌⠕⠇⠀⠠⠃⠗⠇⠀⠠⠞⠑⠡⠝⠕⠇⠕⠛⠽⠀⠠⠞⠑⠇⠀⠼⠚⠛⠊⠚⠓⠑⠋⠊⠃⠁⠙
| Bristol Braille Technology CIC | Tel: +447908 569 214 |
Graham Page
Sep 13, 2017, 3:15:44 PM9/13/17
to brail...@googlegroups.com
Interesting but I can't really understand why anyone would want to scan a Braille book and waste all the time necessary to make this work. I can't imagine that accuracy would be any better than patchy as it would be tricky to achieve any kind of decent contrast. Wouldn't it be more cost effective just to obtain the print book and scan that?
Regards
Graham
--
If you have any problems posting/replying to the Braillists Forum please email he...@braillists.org with a clear description of the issue and forward on any error messages you may have received.
---
You received this message because you are subscribed to the Google Groups "Braillists" group.
To unsubscribe from this group and stop receiving emails from it, send an email to braillists+...@googlegroups.com.
To post to this group, send an email to brail...@googlegroups.com.
Visit this group at https://groups.google.com/group/braillists.
To view this discussion on the web, visit https://groups.google.com/d/msgid/braillists/alpine.DEB.2.20.1709131657210.1921%40m-ThinkPad-X230.
For more options, visit https://groups.google.com/d/optout.
Regards
Graham
If you have any problems posting/replying to the Braillists Forum please email he...@braillists.org with a clear description of the issue and forward on any error messages you may have received.
---
You received this message because you are subscribed to the Google Groups "Braillists" group.
To unsubscribe from this group and stop receiving emails from it, send an email to braillists+...@googlegroups.com.
To post to this group, send an email to brail...@googlegroups.com.
Visit this group at https://groups.google.com/group/braillists.
To view this discussion on the web, visit https://groups.google.com/d/msgid/braillists/alpine.DEB.2.20.1709131657210.1921%40m-ThinkPad-X230.
For more options, visit https://groups.google.com/d/optout.
Sean Randall
Sep 13, 2017, 4:20:49 PM9/13/17
to brail...@googlegroups.com
brailled documents may be a different story though - A computer might translate hand-brailled material into print far quicker than a person. So for a non-print reader producing braille, ...
The issue of contrast is interesting,. Surely braille - being finite in pattern and very narrow in spec would be an ideal OCR candidate if the resolution were good enough? Not my field by a long way.
Sean Randall
IT and Accessibility Specialist
Email Con...@SeanRandall.me
Phone +44 (0) 1905 692280
Or visit My LinkedIn Profile for my blog posts, areas of interest and qualifications at http://UK.LinkedIn.com/in/AccessibleSean
IT and Accessibility Specialist
Email Con...@SeanRandall.me
Phone +44 (0) 1905 692280
Or visit My LinkedIn Profile for my blog posts, areas of interest and qualifications at http://UK.LinkedIn.com/in/AccessibleSean
To view this discussion on the web, visit https://groups.google.com/d/msgid/braillists/00d901d32cc5%243750ad90%24a5f208b0%24%40useit.plus.com.
E M Rogers
Sep 14, 2017, 2:17:45 AM9/14/17
to brail...@googlegroups.com
Far greater use of slates and stylus here than in UK. So as Sean points
out, good for personal documents.
If you get the lighting right then you can ensure the dots cast a
shadow which makes it easier to read. However I know nothing about their
technology in general.
Ed
On Wed, 13 Sep 2017, Sean Randall wrote:
> brailled documents may be a different story though - A computer might translate hand-brailled material into print far quicker than a person. So for a non-print reader producing braille, ...
>
> The issue of contrast is interesting,. Surely braille - being finite in pattern and very narrow in spec would be an ideal OCR candidate if the resolution were good enough? Not my field by a long way.
>
> Sean Randall
>
out, good for personal documents.
If you get the lighting right then you can ensure the dots cast a
shadow which makes it easier to read. However I know nothing about their
technology in general.
Ed
On Wed, 13 Sep 2017, Sean Randall wrote:
> brailled documents may be a different story though - A computer might translate hand-brailled material into print far quicker than a person. So for a non-print reader producing braille, ...
>
> The issue of contrast is interesting,. Surely braille - being finite in pattern and very narrow in spec would be an ideal OCR candidate if the resolution were good enough? Not my field by a long way.
>
> Sean Randall
>
>> On 13 Sep 2017, at 8:19 pm, Graham Page <gp...@useit.plus.com> wrote:
>>
>> Interesting but I can't really understand why anyone would want to scan a Braille book and waste all the time necessary to make this work. I can't imagine that accuracy would be any better than patchy as it would be tricky to achieve any kind of decent contrast. Wouldn't it be more cost effective just to obtain the print book and scan that?
>>
>> Regards
>>
>> Graham
>>
>> -----Original Message-----
>>
>>
>> Interesting but I can't really understand why anyone would want to scan a Braille book and waste all the time necessary to make this work. I can't imagine that accuracy would be any better than patchy as it would be tricky to achieve any kind of decent contrast. Wouldn't it be more cost effective just to obtain the print book and scan that?
>>
>> Regards
>>
>> Graham
>>
>> -----Original Message-----
>>
Reply all
Reply to author
Forward
0 new messages