Archive for June, 2014

FPS, Sample Rate and Pitch Correction

Sunday, June 29th, 2014

calculator.gif

A few weeks back a question was posed on one of the industry forums about to sync recorded audio to a shot that was intended for slow motion. From that, I created a spreadsheet that would do the calculations including semitones needed to keep original pitch. The percentage or semitones can be used in pitch shift audio plug-in if desired.

There is a difference on how the WAV file declared sample rate in order for this to work in Media Composer.  For example, if recording 24.000fps on set with the intent to post at 23.976,the recorder would be set to 48.048kHz. But if the files have already been recorded at 48.000kHz, then you need to reset the sample rate to 47.952kHz using an application like Wave Agent from Sound Devices. The pullup or down would then be performed by the BWF import process. This previous blog shows sample rate to project for “normal” shooting rates. This will not work with the BWF/iXML AMA plug-in, as that does not support pullup/down workflows.

This spreadsheet calculates those values as well as the more offbeat rates that may be used when shooting for slow motion or speed up with different frame rates than than the intended playback rate. Download the Microsoft Excel spreadsheet here.  For those that do not have Microsoft Excel, this should open fine in Google docs or open source alternatives.

Don’t mess with the formulas!

Update 6/30/2014: You can now access the spreadsheet as a Google Sheet. You may need to sign in with a Google account to work with the spreadsheet online.

Update 10/26/2014: It has been brought to my attention that some field are not properly displaying when opened in “Numbers“. Specifically the sample rate fields. Make sure that these fields are set to 5 digits.

Adobe CC 2014 DCP Creation

Saturday, June 28th, 2014

ame.png

 Update 2/6/2015: Latest updates to Wraptor have eliminated this behavior to be in line with other applications as it relates to frame rates and audio sample rate.

With the latest release of Adobe CC 2014, Adobe added the ability to easily create DCP directly from Premiere Pro CC or Adobe Media Encoder CC. The Wraptor plug-in is provided by Quvis. This is exciting news for indie filmmakers looking to create a DCP screening copy for festivals or screen a work in progress in a theatre. Being mainly a Media Composer user, this was great news and DCP output from Media Composer has been a long time request from the Avid community. I would guess there are probably a fair amount of Media Composer users who also have an Adobe Cloud subscription for Photoshop and After Effects, they now have a solution for making a DCP from the Avid timeline.

Because the Wraptor plug-in is limited as to what controls are available, making the DCP creation process very easy to do. This is both good and not so good.  It pretty much is a drag and drop process after selecting aspect ratio. The output will always be 24.000fps but can take in 23.976 fps and 25.000 fps programs. It always assumes Rec.709 video (and levels) as an input and will properly apply XYZ and DCP gamma. For Media Composer users, do a mixdown or a render with either DNxHD or ProRes (OS dependent) and export a “same as source”.mov file. Open that with Adobe Media Encoder, select Wraptor, then aspect ratio of content, and you’re good to go. Adobe Media Encoder will transcode and create the DCP package with proper XML and MXF wrappers as defined by the DCI specification but no control over file DCP naming conventions.

The “not so good” side of the Wraptor encoder that comes with Adobe Media Encoder CC 2014, is that it does a frame rate conversion where program duration is maintained. So if working in 23.976p or 25p project types, it does what Avid Mix & Match does which is add or remove frames to maintain duration, rather than a frame-for-frame conversion where the program duration will change; .1% faster for 23.976 sources, and 4.1% slower for 25p sources. The advantage of frame-for-frame is much better overall quality as the motion remains as originally shot and mastered, be it camera moves, moving objects, or both. Of course the audio would need to be sample rate converted to maintain 48kHz when changing playback rate, but it is far easier for the eye to see motion artifacts than it is for the ear to hear a one-pass sample rate conversion.

For my test, I created a 1080p/25 timeline that was 00:01:00:00 in duration and created a DCP with Adobe Media Encoder. As you can see in the screenshot, the frame count of the EasyDCP Player (bottom left) does not match the burn-in frame counter of the program. At Frame 356, there is already a 12 frame difference being compensated for. Also, if you look at the bottom right, the total program duration at 24fps is still listed as 00:01:00:00. Click frame to see frame actual size:

wraptor-dcp.png

I then used FinalDCP as a comparison, as it does support frame-for-frame conversion as part of its feature set. As you can see in the following screenshot, for the very same frame in DCP Player, the frame count is the same, and the bottom right program duration is longer, as I would expect when slowing down 4.1% and since audio plays in sync, there was a proper sample rate conversion done to maintain sync at 48kHz. Click frame to see actual size.

final-dcp.png

I think it’s great to have a DCP encoder that can be used for quick screenings or festivals as part of a suite many of us may already have, but I would not recommend it for final delivery and distribution due to the quality of motion artifacts than will happen. This can be overcome in Premiere Pro by taking your final program output and using “interpret as” 24.000fps as it will create a frame-for-frame version. Then deal with converting audio tracks in an audio application for sync and sample rate conversion. For Media Composer users, I wrote up a step by step in my first blog entry:
File Based Universal Mastering

I will be reaching out to Quvis, to see if they have an upgrade to the free version has these types of controls with Adobe Media Encoder and will update once I get a response.

Update 6/29/2014:

  • Job te Burg mentioned in another thread that other DCP applications do the frame-for-frame conversion as well. Thanks for the heads up, I was only using FinalDCP as an alternative example of , and in my opinion, better method for higher quality DCP creation.
  • Oliver Peters emailed and asked about padding or scaling of the 1920×1080 sources. I admit my tests were focused on frame rate conversion and did not check to see about padding or scaling but will try to get to that in a future test.

Update 6/30/2014:

  • I just tried the free DCP-0-matic and it does the proper frame rate conversion, allows for padding or scaling of 1920×1080 to met the 2K DCI spec, and has stereo 3D support. Comes in all flavors of OS: OS X, Win 32bit, Win 64bit, and Linux.

 

BWF/iXML AMA Update (v8)

Friday, June 13th, 2014

ixml-logo-web-lg.jpg

There has been a thread on the Avid-L discussing BWF import and AMA linking and I was surprised to read in that thread that the AMA BWF/iXML Media Composer 7.0.4 now supports monophonic BWF files as a single clip. before this release it was only supported by import. I did not see this in any of the READ ME’s but is a nice add to the AMA functionality.

But… there are still trade-offs to consider between AMA and import that need to be considered when dealing with monophonic files. The following graphic shows the different results when importing or linking and how support for metadata changes depending on method chosen (click for full size)

bwf-mono_poly_ama_import.gif

The first thing is a bug that has cropped up in v8 where importing monophonic files and creating a single clip upon import does not have the right duration. Basically the imported duration (X) end up being: X = (BWF duration)/n  where n is the number total number of tracks being imported. As seen in the import, what should have been 2:27:00 in duration ended up being 18:09. While we can expect this to be fixed in a updated version, hopefully it will also include the track metadata that gets lost in tracks 5 and higher when importing polyphonic files.

AMA linking to the monophonic files works as mentioned, but does not parse any of the track metadata on and of the tracks. There is no workaround for this as there as  an option to link and maintain individual tracks does not exist with AMA as there is with “Import”. There is also some inconsistencies with source metadata between Source File and Tape ID on the AMA linked version that may affect EDL and such downstream.

audio-import-options.gif

For now, the best method for monophonic files is to import them as individual tracks, then select all the tracks in the bin and select AutoSync as seen in the bottom bin of the above screenshot. No need to AutoSequence them first as mentioned in the Avid-L thread, and multiple takes can all be done in one pass. The one downside to this method, is using matchframe to get back to alternate tracks from the sequence. For example, the 8 track clip in the example has a mix track on 1 and 2. AutoSync allows for any consecutive span of tracks to be used in the resulting .sync clips such as 4-6, or in this case 1&2. What is very helpful when editing is the ability to do a double matchframe back to the 8 track clip to use an ISO track instead of the mix track. This is done by match framing on the sequence which loads the .sync clip. Then, from the sequence side, turn off the V track and matchframe again on the source side and it will load the multitrack audio. When synced from a polyphonic file, or a single clip created from multiple mono tracks, it gets loaded at the same position for easy track and position selection. In this scenario, it will only load the original single mono track and not the synced group defeating the purpose of this feature.

Another workaround is to combine the monophonic files into a single multitrack polyphonic file. There are several applications that can do this, but this is easily done with Sound Device’s free application Wave Agent  available for both Windows and OS X. Here is the result of combining them, and then accessing the file via the BWF/IXML AMA Plug-in. (click to enlarge):

wave-agent-combine.gif

As always, plan your workflow accordingly. If there is no additional track metadata logged or needed,  then AMA may be the best route to take, if there is, import is the better way to go. When it comes to audio, it is so fast to import, that the instant access is not that much of a benefit as found with video formats. I would love to see a metadata view available in the AMA window selecting the BWF files displayed in the window, then do an “import” versus a link at that time rather  than the potential additional steps of transcode or consolidate with more clip management needed. Then create a “container” type file that managed all the tracks with the ability to define as mono, or stereo (or more). Once edited in the timeline, it would be a simple right-click to activate the ISO track(s) needed for that event. This would offer the best of both worlds.

Update 8/9/14:  Version 8.1 has fixed the duration issue when AMA linking to monophonic BWF files, but track metadata is not supported at all.