[Mark Spano]

Stavros,

You may have seen the solution posted elsewhere as we just figured it out today. My coworker and I discovered this after noticing different behavior than normal when working in Tiger vs. Leopard with the Xsan. Deleting files on Tiger system would go to the trash, while deleting files on Leopard system would force the user to delete immediately. Found that it was a different way that Tiger and Leopard write the .Trashes files on the Xsan. Deleting them for each user and remounting/rebooting the clients solved that, and incidentally solved the ‘make offline’ issue as well. I’m copying it here so that it’s on the Cow for posterity:

1. Unmount Xsan volume from client.
2. Mount Xsan volume on MetaData Controller.
3. In Terminal on MDC, go to /Volumes/Xsan/.Trashes.
4. In Terminal on MDC, enter sudo ls

The output should be UIDs. Find the UID for the user having this problem, and then type sudo rm -R ### where the ###’s = the UID.

Have the client restart, then mount the volume, and try this media offline workflow. It should be back to normal…

[Mark Spano]

For SD masters, I would think original recipe ProRes 422 would be just fine. It is a 10-bit codec and a high enough bit rate to be virtually lossless. I would have no reservations archiving to ProRes 422.

[Mark Spano]

It’s a soundalike of the Beyonce/Jay-Z song “Crazy in Love”. I’ve heard this on a stock music CD but can’t remember which. Most likely it’s an APM track – you might have success forwarding this link to one of their music searchers or someone from their website. https://www.apmmusic.com/

[Mark Spano]

And based on what Jerry says, if you really want the resulting output to look that way (with only one field out of two represented), nest your sequence and throw a “Deinterlace” filter on it (in Video Filters > Video > Deinterlace). Set to whichever field is dominant for your sequence (usually lower for SD and upper for HD) and then output. Basically you will be throwing away half of the resolution of your sequence, but you’ll get a Quicktime that looks pretty close to what FCP is showing you in the sequence viewer.

For a much better deinterlace, you can use Compressor and its Optical Flow deinterlacer. Export a same-as-source Quicktime from your original sequence (yes, this will be interlaced). Drag that into Compressor and create a preset with your desired codec and frame size, and in the Frame Controls tab, choose Output Fields = Progressive and Deinterlace = Best. Compressor will choose the good parts of both fields and output you a progressive frame Quicktime.

I also recommend JES Deinterlacer for these kinds of tasks.

[Mark Spano]

[Phil Hawes] “So if my sequence is ProRes 60i and I have a 24p clip or a 50i clip, I will modify the clip to be 60i”

Sure – that is fine. What you are talking about is resolving differences in frame rates.

[Phil Hawes] “has anyone had problems mixing non-drop frame and drop frame clips on the same sequence?”

Answer: no. No problems whatsoever, since here there are no differences in frame rates.

[Mark Spano]

What all these people are trying to tell you is that your source clips’ timecode is completely independent of your sequence timecode. Whatever timecode you set your sequence to (drop/non-drop) is all that matters in your case. Leave all your source clips intact with whatever timecode they were captured in. No reason to change that. You can edit any drop or non-drop source clips into your sequence just fine.

[Mark Spano]

Much easier on your footage to add frames than to get rid of them. I would argue that converting your HD 23.98 footage to SD 29.97 first would allow you to preserve each bit of footage’s natural look. Since you are delivering 29.97 and some of your footage is 29.97, you should get everything to 29.97 before you edit. This is my opinion, and some people will tell you otherwise and have good reasons for it too, but this is what I would do in your scenario.

You can use a file based transcode or a hardware downconvert to get your HD 23.98 footage to SD 29.97. Transcoding is easy with Compressor – you create a preset that will output to the desired frame size (720×486), codec, and frame rate (29.97), with frame controls set to fast retiming, best resizing, and lower field first output. Hardware downconvert is also easy with the Kona – you make a sequence that is HD 23.98 and drop your footage onto it, then set the secondary output of Kona to SD NTSC (so that the Kona hardware does the pulldown insertion and downconvert). Then you can do an edit to tape of your footage out to a DigiBeta and capture it back in as SD from tape. That’s if you have a deck handy…

If you want to edit everything at 23.98, you can, just be aware that when you upconvert SD 29.97 to HD 23.98, you lose frames and perceived resolution. Your original SD footage is also not going to look so nice by the time you later downconvert your whole sequence.

[Mark Spano]

I’ve had good results with original recipe Apple ProRes as well as ProRes LT. HDV is a 25 Mbps codec, so the 100 or 145 Mbps flavors here should keep as much of the quality of the original and give you I-frame goodness and 10-bit headroom nicely.

[Mark Spano]

Shane is right. Take heed!

That is all.

[Mark Spano]

Cut video in FCP with your sequence settings set to 6 mono discrete channels for audio. When you export, do a video only quicktime export using current settings and do a separate audio export to AIFF of your sequence. That will give you an AIFF of each track. Go to Compressor and choose an AC3 setting and duplicate. Set it up for a 2/2 (quad) channel layout with LFE turned on, like this:

Leave all the rest of the settings intact and save this setting. Then, in the project window, click “Add Surround Sound”. Here you assign a surround sound item the channels you’ve exported from FCP. Click each speaker and select the file you want to play there. Any channels you don’t have, leave them empty. Apply your preset to this surround sound item and process. This will give you an AC3 file ready to mate with your MPEG-2 video for authoring.