Blog

• The mix&match feature of the Avid (e.g. MC4) assumes bringing in footage in corresponding projects. After that, you can instantly access that footage from any other project (or project type), and playback in real-time
• http://lfhd.net/2009/10/01/avid-mc-4-0-inside-look-at-mix-match
  • Your sequence setting is what you tell it.  1080i, 720p, 525i…whathaveyou.  And whatever clip you add to that that ISN’T that format, gets scaled to that format…using a filter called a MOTION ADAPTER.  This add interpolation to match the sequence settings, and this is added automatically when you add new footage that doesn’t match.  And there are all sorts of interpolation modes…these are all user selectable.  AND you can change your sequence settings to match something else later.
    
  • If you want the interpolation to better then you can “promote” the motion adapter to a full blown TIME WARP (that has been there for many years) and the footage will benefit more.
  • Works in software-only (no Mojo required) and takes advantage of multi-core (e.g. 8 core)
  • Avid’s ‘open timeline’ implementation is much better than FCP’s.  Avid MC automatically adds a plugin that is designed to do this upscale in very smart ways. It isn’t just scaling it and then repeating a frame.
  • The editor does need to have certain “switches “ turned on to see the highest quality output, such as: -Full Quality 10bit output, -HQ RT Scaling Decoder, -Advanced Polyphase image interpolation.
• For example: “you will have to import NTSC clips in an NTSC project and 720p60 clips in a 720p60 project. If you try to import 720p60 files into a 30i project, you will be downconverting upon import, which is not as nice, and will not be able to handle certain metadata correctly”
  

Amazon lets a system admin type person rent time/space/bandwidth on Amazon’s Cloud Computing network, “Elastic Compute Cloud”, EC2 (which also sounds like “Easy-To”).  It costs e.g. about 10 cents per hour (depending on selected level of service).

• Maybe useful for CGI generation?

Setup is fiddly, but this tutorial explains how to do it from command-line, including some pragmatic tips:

• s3.amazonaws.com/AmazonEC2Tutorial/AmazonEC2Tutorial.html

Here’s an easier way, via web-based GUI:

• http://www.youtube.com/watch?v=bBajLxeKqoY
  • Shows how to use via EC2 FireFireFox UI plugin, a much nicer way to work.

Generic intro to Cloud Computing:

• http://www.youtube.com/watch?v=XdBd14rjcs0&feature=related

 Automatic

Discovered at: http://prepshootpost.blogspot.com/search?q=

The Levelator is a freeware application (for various OS)  that automatically evens out multiple audio sources:

“It’s software that runs on Windows, OS X (universal binary), or Linux (Ubuntu) that adjusts the audio levels within your podcast or other audio file for variations from one speaker to the next, for example. It’s not a compressor, normalizer or limiter although it contains all three. It’s much more than those tools, and it’s much simpler to use. The UI is dirt-simple: Drag-and-drop any WAV or AIFF file onto The Leveler’s application window, and a few moments later you’ll find a new version which just sounds better.”

“So how do we calculate levels and process audio for The Levelator?  We first isolate segments that are silent and remove them from the calculations. We define silence as audio segments which have no subsegments of 50 ms or more where the RMS is greater than -44.0dB. We then compute the RMS value of the remaining segments and normalize them to our target RMS level of -18.0dB.

The above is actually a drastic simplification of The Levelator’s processing, which takes into account a number of subtleties when dealing with certain real-world sources. For example, the silence threshold of -44.0dB is not reasonable if the audio before normalization is already very quiet. The -44.0dB value is therefore used only after the overall RMS is first normalized to near that target. This requires an iterative calculation. The Levelator processes an entire audio file, not a continuous stream, so we have the advantage of infinite lookahead and the ability to make multiple passes over the data in large and small chunks.”

Mac/FCP add-on to automatically synchronize multicam clips based on their audio content.

Discovered in article at: http://prepshootpost.blogspot.com/search?q=

The article says: “…PluralEyes … syncs up multi camera footage without use of timecode. It’s in beta and you can download and try it out yourself.  (In the article author’s tests) it worked on about two out of every three clip pairs. When it worked, it was perfect, effortlessly lining up and then converting clips into a multicam clip in the browser.”

Video: http://www.singularsoftware.com/autosync/

”

System Requirements:

• OS X 10.4.11 or later
• Final Cut Pro 5.1.4 or later
• PluralEyes™ analyzes the audio content, so all clips to be synced need to have an audio track.

Once captured, the next problems are backup and archive.  Here is one man’s solution:

• We capture to a 2TB drive and edit from it as well. Then once it starts to get full, we use Retrospect Backup software to move files from the 2TB drive to smaller removable drives. Retrospect keeps track of the drives and once they fill up, you can just add another one. The main reason we do it this way is that Retrospect keeps track of all the data so that it’s very easy to find and restore projects. You can view them by the date you saved them or you can simply do a search…even for individual files. Plus, because they are firewire 800 drives, the restore process is very quick (at least it is right now with standard def DV footage….HD footage will take up more space)
• [http://promaxsystems.blogspot.com/2008/01/promaxs-xdcam-ex-workflow-video.html] 
  • It also has other useful info e.g. the transfer rates experienced in practice.

Forum thread on how to ingest EX footage into FCP (via ShotPut and XDCAM Transfer).

•  [http://www.dvinfo.net/conf/sony-xdcam-ex-cinealta/136492-xdcam-ex-fcp-xdcam-transfer-workflow-2.html].

It includes a cool screencast video tutorial by Matt Davis:

• [Batch naming clips in XDCAM Transfer] 

The point of doing it his way is that it makes thinks more foolproof than the more basic “just load it in” approach and does so in a fashion that is semi-automated.

According to CineForm NEO HD FAQ [http://techblog.cineform.com/?cat=69] as of 2009-05-05:

“Color does NOT support QuickTime codecs. Instead, Color only allows
certain codecs that have been compiled into the code to be used.  Apple
is aware of this limitation, but they haven’t been to quick to solve
the problem.”

Whaaaaat!?!?!   If so then that’s astonishing(-ly uncool).

On the Mac, Cineform uses the QuickTime wrapper, hence:

“CineForm files currently do NOT work in Apple’s Color application.”

“Final Cut 6 (with an update) will recognize XDCAM footage more or less in it’s native format. You import the footage using the Sony Transfer software and it merely puts a QuickTime “wrapper” around the XDCAM footage. It’s still Long GOP like HDV but a better codec. You can render it into discrete frame codec as you say but it is not entirely necessary. I sometimes do and often don’t. I do a final render out in a full frame/intraframe file and then send it to compressor to munch it into whatever final form I need it in.” [http://www.dvinfo.net/conf/sony-xdcam-ex-cinealta/144941-new-avid-release-could-revolution-xdcam.html]

I like to have a feel for bandwidths - roughly how big these things are and why.  Especially when I start running out of storage space and having to purchase additional disk drives etc…   Also I like to get a handle on how adequate my RAID will be for capturing uncompressed HD.  Google-search & calculations:

• In the case of HD-SDI (1080i50), bitrate is around 1.5 Gbps [http://www.filmmakermagazine.com/fall2006/line_items/hdmi.php]
• In our case we have 720p50, but that apparently has about the same uncompressed bandwidth as 1080i50, namely 1.5Gbps [http://www.dtvforum.info/index.php?showtopic=9100]
• My system has a RocketRaid+ProAVIO RAID5 arrangement.  I have not yet tested its uncompressed capture ability (I havea BlackMagic card), but it (if I’ve calculated it right), it looks like it should handle it easily.
  • RocketRaid themselves report tests [http://www.highpoint-tech.com/PDF/RR2522/RR2522_Performance_Report.pdf] involving an AJA capture card and a RAID5 array (as is mine) giving read and write speeds at about 400MBps (=3.2Gbps, because B i.e. Bytes = 8*b = Bits). 
    • If I understand this right, my system has a capture bandwidth twice that an uncompressed HD signal.  Plenty of headroom then.
• ProRes’s 123 Mbps is not far off the “100Mbps” figure often mentioned in relation to the Convergent Design flash memory recorder, at which level the recording, compressed (as is ProRes) in a variant of MPEG2, is regarded as almost indistinguishable from uncompressed.

Suppose you added an effect to the timeline and as a result it shows a red “Unrendered” (or “Needs Rendering”) line above it.  You press Command-R (say) and rendering happens and now you can play it in real-time.  But what exactly is this rendering?  What format is it in and where does it go, i.e. what file(s) in what folder(s) ?

•  The rendered file appears in a project-specific subfolder of the Rendering folder of FCP’s Scratch area.  In my case I defined the latter to be on the RAID.
  • Location was: [RAID_ProAVIO/App-Specific/Final_Cut/ FCP_Scratch/Render Files/Voda 2009 Estab 001/]
  • Example file name is: “Sequence 1-WP1-00000001“
• The coding format is ProRes 422 - Standard, not High Quality (which would be overkill).  This is good news!
  • VideoSpec displays its FourCC code to be “apcn”
  • The four-character codes “apch” and “apcn” denote ProRes 422 HQ and ProRes 422, respectively [http://blogs.adobe.com/toddkopriva/2009/03/prores_422_colors_in_after_eff.html].
  • The rendered file’s bitrate, according to VideoSpec, is an average of 123Mbps, around four times that of the HDCAM EX high-quality mode (35Mbps).  Nevertheless a great reduction on the uncompressed bitrate (around 1.5 Gbps) and broadly comparable to Convergent Design’s 100Mbps, judged by some to be virtually indistinguishable from uncompressed quality.  So it bodes well.