Steve Gregg Bible Commentary

[darinhouston]

There are a few services and software... Dragon Naturally Speaking is probably the main one for desktops.

Google podcast transcription services -- here's a couple of examples...

http://accentance.com/transcription.htm
http://www.automaticsync.com/captionsync/
http://www.transcriptionstar.com/mp3-po ... ption.html

Software...
https://www.cpcweb.com/youcap/

[Perry]

I don't believe that software you linked actually does speech to text. It seems to be more about adding captioning to video after you've already digitized the text.

Of the three services you linked I only could discover the prices from one of them. found here.

At $1.35/minute it would be just over $1,900.00 to pay that service to transcribe "Authority". I don't think that's a viable option.

Speech to text software is very difficult, and the tech just isn't here yet. I've seen it sort of work in specialized fields. I have a friend who integrates the medical module of Dragon into parts of the medical software he sells. Even in a specialized field it's only so-so.

BTW, I've completed the first lecture. It is very raw, and I haven't re-edited it yet. I don't intend for this to be the end product (even after I edit it). I'm not sure that this is going to continue to be my working procedure on this project. The bolded numbers for section headings are time stamps in the audio file. So for example, if you go to times 26:46 in the audio file, you should hear Steve say "The hippies were a rebellious people." (I promise I picked that example at random). It may look a bit goofy, but this method serves as well as any other for segregating the raw text.

[john6809]

I agree, I entered a 900 minute teaching series for quote and got a $900 estimate - ouch. My wife just downloaded a free app from Dragon - I will play with it and see what the results are. I love listening to the teachings and don't mind correcting errors caused in conversion. Would probably be faster for me. BTW - nice work Perry! You work quickly.

[Perry]

Versions of windows 7 and later have some speech recognition software built right into it. All speech software that I've seen requires that you "call out" punctuation. So comma for instance comma I would talk like this period. There may be a way to have the built-in speech recognition software read straight from and MP3 file, I'm not sure, but, for fun, I simply put a mike by a speaker and played the first minute of the second lecture into the speech recognition software. Here's the result:

It will be the only event live lecture and an infant care of it on an acting along planning and buying or when ordering an attitude about his relationship to knowing that for the rock quarry and on through and I think most people know that whatever does they know one thing that I know would not identify and why do you base your knowledge of wrath and that's what has given mall in the rounds in all one room beneath you know and the ideas I've added that where is your one goes on in your mind when you don't win on the white kids weren't very good as Perry and her as a native of it.

I thought it was interesting that my name was in there. Probably my customized dictionary did that somehow.

Pretty useless.
However, this has me thinking of a new idea. I'm going to see about listening and "re-speaking" into the speech recognition and seeing if that works any better.

You may, for instance, find it interesting to know the last paragraph of this message was dictated by using the speech recognition software. I was able to do it without a lot of hassle it didn't require a lot of the editing (it did require some), and it was probably a bit faster than I could have typed it. Also, the speech recognition software is a far better speller and I am. I'm a lousy see speller.

[jarrod]

I agree about the speech-to-text technology.

Perry, great starting work! I didn't mean to "change" your work... I was just trying to provide clarity with the organization. That is the cool thing about the wiki... you can always change it back or disagree because it tracks all of the changes.

I will commit to doing the new Kingdom of God series from his latest Great Commission School recording found here: http://theos.org/media/category/127/.

I hope through this we find a good organizational structure so that it is not only good for reading/publishing, but searching. I would like to have "tags" associated with sections such as verses (i.e. when Steve jumps around between verses) so that we can search for specific verses and be able to display the varied topics and teaching steve has done on that verse.

Jarrod

[jarrod]

Perry,

I see how difficult it is to transcribe these lectures and I am not even reformatting like you did, but just trying to type as fast as I can (I never realized how fast Steve speaks... lol). I definitely have improved my time after trying a few methods. I did a couple of things: 1) Slowed down the playback speed of the audio 2) loaded into iTunes and used the pause/play button on my mac. This combination allowed me to get through a section much, much quicker.

I know it is a lofty goal to think about transcribing the audio, but if it is worth the time, I want to keep plugging away. Our strength would obviously be in numbers... if it was just you, or just you and I doing it, it would take forEVER (<-- that forever can be interpreted literally).

I can see now why dictation machines have foot pedals. If my i-touch had a foot pedal it would speed things up. I'm going to at least finish transcribing the first lecture... (maybe the first two lectures) mainly because I'm kind of stubborn.

[darinhouston]

Do we need word for word? Or might paraphrase be of use?

[Perry]

it would take forEVER (<-- that forever can be interpreted literally).

Yes, it can be interpreted literally, unless, of course, we can get Steve to shut up! Which seems unlikely.

I did do the things you suggested, and saw similar improvements. I started playing back on my itouch rather than using my computer to play back. That way I wasn't having to constantly change focus back and forth between the text editor, and the play back controls. The itouch has a half-speed playback option that I'm using, which also helps a lot. Finally, the itouch has a skip-back-30-seconds function, which is much easier than try to tickle the positional slider into just the right place on a file that's typically 90 minutes long. All of those things made using my itouch to for playback a great improvement.

Lately, I've been playing around with speech recognition. It's novel enough to keep me interested at the moment, but that novelty is already wearing thin. It leads to some pretty funny mistakes. "Sin" constantly comes out "CNN" which is probably not too far off base. "A.W. Tozer" came out as "eight to be toes are." Those are the ones I've caught. It's constantly making mistakes between these words, "to, two, too", "their, there, they're", "its, it's". I'm certain I'm not catching and correcting all of those. I'm sure there are some funny errors slipping through.

I will say this, using the voice recognition, even though it's not necessarily a time saver, is a good way to break the monotony.

[Perry]

Do we need word for word? Or might paraphrase be of use?

I have been spending a lot of time thinking about this. There are absolutely up-sides and down-sides. These are my current thoughts on the pros and cons of continuing plug away, as best I can, at word-for-word.

Pros:
1). Steve often uses some pretty clever turns of phrase that I don't fully appreciate till I focus on them enough to transcribe them. “A good vocabulary promotes an economy of effort in communication.” is an example of one. That's actually a paraphrase of what he actually said, but I still missed it.
2). Better search capabilities when we have the full transcript more or less verbatim.
3). (Most important), I can be sure I'm not misrepresenting what he's said.
4). Having everything transcribed gives me something concrete to work with. It's my plan to create an edited version of the transcript that's more fitted to the written word. By having the raw transcript, I have something to begin cutting and pasting, culling and refining. That's when it will be more fun, and not just the drudgery of transcribing.

Cons:
1). Labor intensive.
2). Steve is, after all, just Steve. It's not like we should hang on his ever single word as tho they're strung down from heaven on a golden thread... some of them aren't, why spend all this labor on them?
3). Spoken word and written word are different mediums. And, Steve, I've learned, begins almost ever sentence with a conjunction, usually, “and”, “now”, or “but”, and they sometimes, I mean, Steve, you know, in speaking, there is this tendency, with the different subjects, and some of them string on and on, and move off into other subjects kind of like Paul did sometimes, which isn't so bad when you have vocal inflection for emphasis, and you begin to wonder if there is a subject and verb, and did this sentence end up where it started and is it a question or a statement or what?
4). He sometime interacts with the audience and makes reference to what are obviously things that are local to the room he's occupying at the time of the talk.

I'm still a bit uncertain whether to try to transcribe all sixteen of the “authority” lectures. I think somewhere there's a happy sweet spot between notes that are more extensive and less personal than I usually take, and the verbatim transcribing I'm doing now.

[jarrod]

Honestly, I started out by trying to collapse the verbosity of Steve's lectures, but I realized that some things I was doing might twist Steve's original meaning. Because of this alone, I decided to stick with word-for-word (mostly). Then, once in the wiki, I figured we could get editors to clean it up. Making edits in the wiki save a complete history of all changes as well as the author of the changes. Plus, it was easier to just type it word for word and not attempt to interject my own interpretation.

Do we need word for word? Or might paraphrase be of use?

Pros:
3). (Most important), I can be sure I'm not misrepresenting what he's said.