-
OT: Digital Heaven SpeedScriber is now in beta
-
not sure if it does — i know on line it syncs to the video you uploaded…….but that’s good enough for the documentary work i do…..with timestamps from :00:00 at the start of the video — what’s amazing is its accuracy. Used to (try to) Use the transcription function in earlier versions of Premiere and it was hit or miss (mostly miss)
Andy Field
FieldVision Productions
N. Bethesda, Maryland 20852 -
[Andy Field] “i know on line it syncs to the video you uploaded”
SpeedScriber does that as well. I don’t want anyone to have the impression it doesn’t sync to the video you upload.
Thinking out load that if FCPX ever supported .srt or any other caption format, especially on a source clip bases, that would be viable.
[Andy Field] “Used to (try to) Use the transcription function in earlier versions of Premiere and it was hit or miss (mostly miss)”
I don’t doubt that was one of the motivators for Digital Heaven to enter into this market.
-
[Craig Seeman] “If we export each transcribed sentence as a marker, then you can’t search for a word in FCP and see where the word is.”
Exactly. Which is why I have yet to understand why people always insist on using MARKERS. Weird. What you need to be doing is defining and naming FAVORITES which are basically range based markers! Bingo… you get what you’re looking for. And on top of that, you aren’t flooded by a countless number of keywords. Personally I think that’s the worst, most cumbersome choice possible.
Make them favorites, switch to favorites only, enter search term, bang. You have not only the clip with that search term, but the that exact section… only.
– RK
____________________________________________________
Deutsch? Hier gibt es ein umfassendes FCP X Training für dich! -
[Craig Seeman] “Keep in mind every single word needs to point to a specific spot on the clip.”
I’d say that is complete overkill. Having favorites (as described) for every sentence would more than suffice imho. A single word is of little to no use to me if I don’t know the context.
– RK
____________________________________________________
Deutsch? Hier gibt es ein umfassendes FCP X Training für dich! -
[Andy Field] “not sure if it does — i know on line it syncs to the video you uploaded”
Then I’d say you’re comparing two services with entirely different focus and levels of usability. Only having the transcript in online or Word form would make it about 10% as useful for me and a comparative PITA. In which case I might as well just pay a one-time fee and get SoundBite or the likes. MUCH cheaper over time, but still the same, continuous jumping back-forth-in-out dance. The exact thing I’ve wished I could finally do without for ages, which this would appear to offer.
– RK
____________________________________________________
Deutsch? Hier gibt es ein umfassendes FCP X Training für dich! -
[Craig Seeman] “Keep in mind every single word needs to point to a specific spot on the clip.”
[Robin S. Kurz] “I’d say that is complete overkill. Having favorites (as described) for every sentence would more than suffice imho. A single word is of little to no use to me if I don’t know the context.”
Surely the “databasiest” of all NLEs can handle a little metadata, right? Wink?
Kidding aside, knowing when an individual word occurs in video, rather than a sentence, allows you to cut from the script. Cutting that way brings its own set of problems, for sure, but being able to mark AV ins and outs directly in the transcript was more useful than I expected back when Premiere had it. I’d also suggest, perhaps outside the scope of this conversation, that knowing when an individual word occurs is important in an automated archival or analytical system.
If you know when the words occur, and if you know where the sentence boundaries are (not always easy from raw speech), then creating the favorite sentence ranges you describe is an easy problem to solve. I think it’d make a great feature request and I hope you submit it.
Walter Soyka
Designer & Mad Scientist at Keen Live [link]
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
@keenlive | RenderBreak [blog] | Profile [LinkedIn] -
[Walter Soyka] “the “databasiest” of all NLEs”
Genius, Apple marketing should use this!
-
[Steve Connor] “Genius, Apple marketing should use this!”
My NLE is databasier than yours!
Walter Soyka
Designer & Mad Scientist at Keen Live [link]
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
@keenlive | RenderBreak [blog] | Profile [LinkedIn] -
[Walter Soyka] “knowing when an individual word occurs in video, rather than a sentence, allows you to cut from the script.”
I don’t see how hearing just a single word isolated from a sentence somehow poses some sort of “editing advantage”, sorry. If that’s what you’re describing. Aside from the fact that with what I was talking about you can still search for a single occurrence, too. If you need the COUNT on the occurrences, then have them generate a Word document along side.
Put in a slider to define the “size” of regions from a single word up to a whole paragraph per range and you’re good. But considering the way favorites work and their limitations (no overlap for one), they probably wouldn’t actually lend themselves very well to that small of a region such as a single word. But for my purposes I would find that to be complete and utter overkill anyway.
[Walter Soyka] “I think it’d make a great feature request and I hope you submit it.”
For which app now?
I have had various discussions with people that could make something like this happen either inside and/or outside of FCP, yes. Unfortunately e.g. Philipp Hodgetts and I are completely contrary opinion on the use of favorites in general, let alone in this way. Because IA would be one example with their “magic keywords” feature in Lumberjack, which unfortunately (for me) only works as keyword ranges as well.
– RK
____________________________________________________
Deutsch? Hier gibt es ein umfassendes FCP X Training für dich! -
[Robin S. Kurz] “I don’t see how hearing just a single word isolated from a sentence somehow poses some sort of “editing advantage”, sorry. If that’s what you’re describing.”
I was trying to describe how Premiere used speech analysis. Premiere had a window which displayed the transcript. Selecting a word in the text there would jump the video to that point. You could set in/out points right in the transcript and Premiere would set the markers before the in word and after the out word. Basically, there was a link between the metadata (the transcript) and the data (the video); you could think in terms of the metadata and Premiere would act in terms of the actual data. When the speech analysis worked, using the results worked really well. It wouldn’t have been so useful without the temporal resolution in the metadata down to the word.
As for the feature request, I was suggesting SpeedScriber, since that’s the subject of the thread.
Walter Soyka
Designer & Mad Scientist at Keen Live [link]
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
@keenlive | RenderBreak [blog] | Profile [LinkedIn]
Reply to this Discussion! Login or Sign Up