Joseph Owens
16
COW Coins
16
DaVinci Coins
Forum Replies Created
-
[Louise Macaulay] “starting with a 1920×1080 Pro Res 422 LT file that was shot at a wide angle to leave room for cropping.”
Overshooting… One issue will be that you will have to be careful about all your other element placement.
Maybe a 300×250 matte would help… or whatever will assist in predicting where your ultimate crop will be.Dropping the 1080 material onto a smaller timeline should not affect the overall quality if you are expecting to do a pixel-for-pixel extraction. Had you noticed if the timeline render status changed color? Perhaps the degradation was simply due to a Final Cut lo-res proxy preview.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
You should try the FCPX Techniques forum.
My bet would be that most of the contributors here don’t even want to know about FCPX.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
[Bret Williams] “They only need to change 1 word of the audio,”
Pretty expensive word.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
[Travis Roesler] “I don’t know what interlace and deinterlace means.”
Desktop/streamed, disk-based, non-broadcast video wants to be “progressive”. That means the image is one single raster-frame of pixels that starts at one corner; some methods at the top, some at the bottom, and is continuously, contiguously mapped serially, row by column.
Interlace is a broadcast strategy that was implemented at the dawn of television to cope with the very low bandwidth allocation per channel. Each 29.97 fps “frame” is split into two “fields”, and the “lines” that make up each field are both physically and temporally separate — they are scanned into a home receiver to re-create the whole image; but they are separate entities, and this accounts for any number of motion artifacts.
“De-interlacing” is the process of converting an interlaced format to a progressive format. It is not necessarily completely straightforward. Screw up anything, like the dominance and you are done.toast.hooped.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
Bitrate is a big factor, yes… and in context with the codec you are using… mpg2? H264?
Normally, a DVD (for example) will be encoded at anywhere between 3 and 6 Megabits (3,000 – 6,000 kbits) per sec.
Take a look at some of the presets in Compressor in the “Quality” tab.jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
Drop Frame only exists in the 29.97 world, but there are variations in 23.98 to try to predict program segments/runtimes as they are edited for broadcast — to hit the 2997 drop frame targets, but they are not SMPTE standards per se.
It depends on whether your deliverable client is going to use your document to generate a foreign language version exactly the way you have edited them in your timeline or if someone is going to convert them into closed captioning or something else. It doesn’t sound as if you’re supplying a transcript.
I generate these files on occasion, and normally what I do is export a text file from the captioning software, in my case MacCaption. What has been accepted in the past is to place each caption on one line with the timecode “in” in the first column, and then the text Line 1=Column 2, Line 2= Column 3, etc., if there is more than one line of text in the caption.
Closed captions will contain different information than a transcript, though — sound cues, for example, and a big difference is that a good transcript will identify all the speakers with their individual lines, but won’t necessarily break up the individual’s spoken words into the phrases that captions require for pacing– but CC requires that an off-screen speaker be identified by name or other descriptor like “Narrator”.
Personally, I like to match the text with the dialog, and try to avoid barely missing edit transitions — having the text “pop” is distractin, so text on/off is better matched with the cuts. For a couple of projects, I created a Final Cut “Text Only” sequence and sent that — the prj is small enough to email, and the advantages of having a ready-made timeline are obvious, I think — just as long as the font libraries match!
When in doubt, I apply the rules of closed captioning, which you can read sometime. Its a book. But generally its about readability and grammatical clarity.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
[Esther Casas] “what do you mean with drop-frame??”
There are two flavours of SMPTE time code that run at 29.97 fps. Drop-Frame and Non-Drop Frame (DF and NDF). Its a broadcast thing.
Drop frame was invented to provide a time code display that ran more closely in sync with “clock time” so that after an hour of real-time video playback, a clock and the tape timer would agree, showing one hour, EXACTLY. The issue is that at 29.97, video is actually playing a little bit slower than 30 fps, but the frame count is still predicated on 30 fps. So it takes longer, in real time, for the counter to hit 1 hour… or 108,000 frames… which actually takes one hour, three seconds and 18 frames (108 frames too many, how about that?)
So SMPTE came up with an algorithm that DOES NOT ACTUALLY drop video frames… they just made the time code counter skip 2 frames every minute except on the “10’s”. That works out to: 120-12=108. Math, eh?
For example, the video frame you would be parked on at one-minute (NDF) into the program used to be at 1:01:00:00… but that number does not exist in Drop Frame… it will be displayed as 1:01:00;02. You will see the time code roll over from 1:01:59:29 to 1:01:00;02. Note the semicolon — it is one important difference in how DF is flagged. There is also a “bit” in the timecode digital “word” that is set so that modern equipment can tell what kind of code is in the stream after reading just one frame. But you would see 1:10:00;00, which is where DF “recorrects” itself.Why the time code skips two frames, and why NTSC video runs at 19.97 is peculiar to the way NTSC video used to handle color encoding. But to say that Drop Frame is more precise than Non-Drop is not exactly true on a frame-by-frame basis… if you were to pay attention to every frame, you’d see that DF is only actually completely accurate every ten minutes. But it works out over an hour and everybody’s happy except the people who are anal about hitting 00’s.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
[irene litapop] ” Can anyone help clarify this?”
Use “copy” and not recompress and you will be cloning your media, even if you choose not to copy the entire master clip… that is: only the selected source I/O trims plus handles. It simply truncates the original Quicktimes.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
[Jamie deRooy] ” I may be sending the file out-of-house to get closed captioned. In that case, I wonder if it’s best to give them the 23.98 Master Quicktime and have them do any conversions from there.”
Closed captioning (708) only works at 2997 so the conversion will have to be done… and if the outsource is doing all the transcriptions and time-stamping, then they might be just as happy with the 23.98 master so they can get it right for themselves. It is not something that Final Cut can do on its own.
jPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.
-
YOu might also enjoy:
Read this on The Globe and Mail
Is Twitter turning TV into social media’s hot new hub?
A half-century after Marshall McLuhan heralded TV’s participatory potential, live tweeting may finally be transforming the idiot box into a linchpin of interactivity. As Nielsen and Twitter announce a joint venture to produce Twitter TV ratings, Robert Everett-Green ponders the fallout – empowering and otherwise – of TV’s next actjPo
“I always pass on free advice — its never of any use to me” Oscar Wilde.