SC Voice – the road to v1


Ah OK.

I think in both cases she doesn’t pronounce the letters separately. Let’s wait what Bhante says.


Here’s a new clip for Bhante @sujato and you with an accentuated break combined with a vowel change. The apostrophe stresses the second a which is a slightly different vowel: əcchəɾɪjə’abbʰʊt̪əsʊt̪t̪ə

Since Pali was essentially the written down Esperanto of ancient times, perhaps correctness should also consider the ability of different native listeners to distinguish the sounds. Our current global languages may introduce a need to amend traditional Asian Pali pronunciation slightly so that we don’t mis-hear the Pali.


It’s the second one, right? In case this is correct that the two a’s should be pronounced separately this comes indeed closer.


Ok. I shall add this to v0.9.2 unless Bhante objects.:white_check_mark: Thank you!


When this is fixed, does this mean that all similar instances are automatically identified, or should I report other ones when I come across?


The fix is to replace “aa” with “ə’a” everywhere. This may (?) cause new problems if there are exceptions to this rule. Fixes that are word-specific are undesirable given the vast number of rules that would be needed.


SuttaCentral Voice release v0.9.2 UTC 23:40 (~5m) RELEASE COMPLETED

Thanks to Bhante Sujato. Ang. Sabbamitta, and SCMatt, in v0.9.2 we have a much improved Pali pronunciation as voiced by Aditi, our chosen AWS Polly voice. Aditi is a bilingual voice capable of Hindi as well as English.

Thanks to Aminah, we have a much simplified SCV page that should work much better with mobile phones and smaller screens.

Here are the release details.
And plans for the next release.


Happy Holidays and best wishes for 2019!


My goodness, Karl. That’s one mighty Christmas present! Thank you so much :anjal:


Karl, this is simply fantastic!

(I was already silently weeping at the thought that this release would come while I am away… :cry:)

But now… :joy: :heart_eyes: :star_struck: :smiling_face_with_three_hearts: :joy: :joy:


@blake I am currently updating SCV sutta content via SC api. This will hammer the SC servers. If you have any concern please reply and I will shut down the job. I’m not sure how long it will take. Perhaps an hour or two. It’s doing about 4 suttas/second.

Apparently Bhante Sujato has been quite busy these past few months. My git commit list is getting quite long and has already reached 919 files. :open_mouth:

Update completed. 3753 files changed by Bhante. My goodness, Bhante!


Congratulations to everyone involved! The page is quite elegant in its design. I’m looking forward to more thoroughly investigating.

:confetti_ball: :balloon: :confetti_ball: :balloon: :confetti_ball:


In SuttaCentral Voice v1.0, you will be able to download any search result as a single MP3 file less than 3 hours long. You will have your choice of languages:

  1. Pali/English
  2. Pali
  3. English

Here is SN12.23 spoken by Aditi/Amy as a sneak preview (audio updated 1/11 due to software bug that truncated file):

For example, if you want two suttas, you will be able to type something like: “an3.76, an3.77” or even “an3.76-77”. Etc.

I’ve been trying out this feature and I’m finding it quite compelling to hear Pali followed by English for each text segment. The neurons are clicking and making connections I haven’t had before.

Questions and comments welcomed. It’s easier to change the code now rather than later.


Hi @karl_lew, thank you for the sample!

With the download as is now, that’s English only, I think the introductory part could perhaps do with some improvement. It says:

SuttaCentral Voice recording, AN 8-dot-6-slash-en-slash-s-u-j-a-t-o

The name of the translator is spelled out letter by letter, and it actually took me a few times listening until I understood what Amy is saying here. Maybe it would be nicer if the name could just be pronounced as such.


Interesting! I think I had given up on a general implementation initially, but since translators names are typically Pali, we can now simply ask the software to speak the word. I’ve added this as an optional v1.0 feature “below the line”. If it isn’t easy enough to get into v1.0, we can tuck it into a future release. Thank you!

For example, your own name would be pronounced thus:

it’s not quite right, but seems better than s-a-b-b-a-m-i-t-t-a?


I just realized that it is Jan. 11 with a release promised on Jan. 15. The current release status is posted here. As you can see, I have been aggressive at downsizing v1.0 commitments, but even with that Jan. 15 is about 50% with Jan. 19 looking much better.

The download work was a bit trickier than anticipated.



Just had a look at the plan, looks like you’re making pretty fine progress and that you’re pretty much on course to me.

I notice there was a question mark over one issue, sorry that was my addition, I made a quick note of it when listening one day, but never came back to explain. I haven’t checked again to confirm this bug, but as noted when I was listening to a sutta with multiple sections the player stopped reading after the first section and didn’t complete the sutta (I did reproduce this with another multi section sutta, but if you can’t replicate I can check again).


No need to replicate, I just haven’t had time to diagnose. I’ll keep plugging along. Sounds like the possible delay is fine. Thanks!


Well, of course it is, but looking at the plan and your report everything seems on track. :slight_smile:


Erm. Actually, I cannot reproduce it. I searched for “root of suffering” and then selected Play MN1 from the search results. The bell sounded and it played section 1 and then moved onto section 2. This worked with Eng/Pali as well as just Eng.



Seee! Flying though it all! No, I can’t reproduce either. I’ve only spent 2secs on it, so I don’t know if there’s more testing to do, but I used the same sutta and it’s fine. :white_check_mark: :slight_smile: