[Andrew Rendell]

FCP can only move the tracks in inuts of 1 frame, i.e., a thirtieth of a second, and that is too coarse an amount to line up the waveforms exactly, so you have to choose the closest you can get. Unless your cameras are Gen-Locked together (which is a way of synchronising them) this is normal.

In general you choose which audio track you’re going to use and people won’t notice a very slight difference in the timing of the vidion against te sound (as long as it’s within a frame). In real life the speed of sound is much slower than the speed of light so we do actually see things a fraction of a second earlier than we hear them, so you’re taking advantage of that facility of the human brain to percieve things correctly (but you’ve got to be very close to get away with it).

If you need to mix the tracks from different cameras you’ll have to export the audio to a sound editor to shift the waveforms by the right amount, otherwise you’ll get nasty sounding “phasing” effects.

[Andrew Rendell]

I do like to keep previous versions, particularly when I’ve exported copies for, say, the producer to view (that way I can go back to them if I need to respond to a comment about a shot, say, 6 minutes in and the timing has changed in the intervening time, so that shot is no longer at 6 min). But I agree with keeping your main project tidy – make another project for those backups and copy your sequences into it when you want to hold that version. I’ve been known to save my sequence in that way daily when I’m getting near the end of a big edit, and although it’s rare that I’ve needed it, it’s the occasion that you don’t bother that it turns out that it would have been really useful to have done so…

I prefer to do rough cuts of each scene as separate sequences before putting them together, as my most frequent employment is docs and we often restructure and swap the order of scenes in the edit so it’s handy to have the sequences separate until the structure is agreed. I’ll then do the fine cut on a complete timeline.

I do sometimes have one extra sequence somewhere where I can chuck “useful” or “interesting” bits and pieces. It can come in handy when doing extras for a web site or DVD. Again, have it as another project otherwise it’ll distract you when you should be concentrating on other stuff.

[Andrew Rendell]

Originally all standard def cameras had 4:3 shaped sensors and the camera would read the pixels out to create a 4:3 electronic picture.

Widescreen material, which meant feature films, was either transferred to video as letterbox, where you saw the whole picture but the top and bottom of the screen had black bars, or pan&scan, where the screen was filled with only a section of the film and you lost a large amount of the original film frame.

In the early 1990s widescreen tv was invented as a compromise. It allowed widescreen pictures to be displayed almost fully by stretching the same number of recorded pixels horizontally across the screen. This arrangement is called anamorphic and effectively means that if you imagine the picture as a series of pixels, those pixels are spread apart further in the horizontal direction than in the vertical direction (known as having non-square pixels). So compared to having a letterbox of your wide screen pictures in a 4:3 display, which reduces the number of pixels in the vertical direction, you use all of the available pixels in the vertical direction, so although you now have an imbalance between the resolution of detail in the horizontal and vertical directions, basically more pixels = sharper pictures. This 16:9 anamorphic tv transmits down the same broadcast path as 4:3 and is still standard def.

The early 16:9 widescreen tv pictures were principally made by adjusting the scans on a telecine transfer from film to tv, and tv cameras weren’t ready for the change when widescreen tv took off. Initially what the cameras did was to ignore the top and bottom of the sensor chip, so they’re looking at a 16:9 shaped bit of the sensor and interpolating the number of pixels they could read up to the required number, effectively creating a letter boxed image in the camera and then doing an electronic zoom to increase the pixel count to the full height. After a while the manufacturers came out with wider sensor chips so that that bodge wasn’t necessary, by which time Digibeta was the normal full quality tv format, so the Digibeta camcorders got the new (and expensive) technology and it took a few years for the wider chips to replace the older (and cheaper) 4:3 chips in less expensive formats. As an example, the Sony PD100 was an excellent piece of kit in many ways but the softness of it’s pictures compared to Digibeta when you made a 16:9 show, which was routine at the time (in the UK, at least), meant that it was treated (somewhat unfairly in my opinion) with derision by the engineers at some broadcasters.

I hope that made sense 🙂

[Andrew Rendell]

That’s quite an expensive way of having a second screen attached to a MBP…

[Andrew Rendell]

Thanks for the info, those look really helpful.

[Andrew Rendell]

Well, yes, but all the drop out compensator does is simply repeat the line information from the last good line until it gets a good signal again. We used to get a lot of minor dropouts which would be only a block of lines in one field, so you could replace them using the other field of that frame – a very quick and easy fix but one that you have to do manually.

As I recall, the BVW series were better engineered all round than the cheaper models, but the main difference between them and the PVWs was slightly worse signal-to-noise when making recordings and no audio channels 3 and 4, so although I’d avoid the UVW series for this kind of thing, I reckon that the difference between a BVW and PVW deck for it would be pretty minimal (unless you need audio tracks 3&4).

[Andrew Rendell]

SDI is a digital feed of 4:2:2 sampled YUV, with no data compression. I once did a test where I took a picture out of a digibeta deck, into a vision mixer and back onto digibeta, then repeating that. At 40 passes you could see some changes, but that was due to the vision mixer (an Abekas A84) interpolating the bit depth internally, not the SDI links.

Really your test is one that would show whether your system is working properly, rather than to discover anything meaningful about SDI.

[Andrew Rendell]

A couple of little things concerning analogue recordings:

Always use the component connections (don’t even think about composite if you have a regard for quality) and make sure that the cable lengths are the same for all 3 and use a good quality B&B generator for Gen Lock.

Be sure to clean the tape path regularly and keep the VTR serviced – Beta SP does not have data correction (not being digital) so the slightest imperfection in the tape surface causes drop outs.

[Andrew Rendell]

Peaking around -5 to -9 is a good healthy level (to be honest, I usually record a little bit lower than that, but I’m paranoid about distortion) so I’d just use it on just the one track.

[Andrew Rendell]

It’s because the producer can do it himself, therefore saving money by only employing an editor and a sound editor for one day (the digitising is done by the studio engineer). Basically, by having a rough assembly of the bulk of the show already done the fine cut and packaging can be done easily in a day. The radio version has a different running time to the TV version, no commercial break and things like people’s details (what their job is, etc) are done as supers on TV but have to be spoken on the radio so if the editor exported a finished programme there would still be a pile of work for the radio guy to do (and the radio TX is first).

The reason I’m thinking about this is not to do the same thing as that, but we have a stack of interviews shot on location (some single camera but mostly 2 camera) and what the producer wants to do is go through them and pull out sections for the edit, therefore give me a 10 – 15 minute assembly to cut into a 4 minute item, therefore saving me from spending my time going through an hour or more of interview to find the appropriate sections.

It occurs to me that getting timecoded audio clips out of the FCP into Sadie can be done by omfi rather than WAV… doh!

I don’t know if this’ll ever happen, but if it does I’ll be glad to have thought it through first. Thanks guys for your help.