Jump to content

Something I didn't know about PluralEyes...


RPSharman

Recommended Posts

I'll start by saying I have never done a DSLR project, and only have knowledge of assistant editors successfully using Sync n Link on some RED projects.

But there were some behind-the-scenes guys hanging around on set the other day, and they had a DSLR.  We chatted a bit about syncing, and they mentioned that sometimes PluralEyes gets confused in narrative work.  The issue, I guess, is that with multiple takes of essentially the same performance, the waveform recognition software can have multiple matches.

Anyone here experience this on their DSLR shoots?

Robert

Link to comment
Share on other sites

No personal experience, but that is a very interesting observation. Just as we often do a wild line reading to try and replace an on camera sync delivery, it is nor surprising that PluralEyes could have a problem like this --- multiple line readings that are essentially the same. I do suppose, though, that the linear time frame for multiple lines of dialog with similar wave forms, would be known by PluralEyes, lessening the chance of mistakes.

Link to comment
Share on other sites

The answer is yes, PluralEyes can confuse takes and sync incorrectly. I recently did both production sound and post editing/mixing on a project where PluralEyes was used. It was a multicam RED shoot and I sent solid guide tracks to both cameras for dailies and syncing.

I found at least a dozen takes where PluralEyes picked the wrong take for the video. In the end I simply checked with the video editor regarding which take was used and I pulled it from the source files.

It was a hassle because after I found the first few incorrect files I began second guessing a lot of other clips. It definitely added time to the audio editing process.

So, again, something where time is saved in the video side of things it adds more time on to audio! 

Link to comment
Share on other sites

PE works best if you can lay out all the clips in a row in a sequence with all the audio as well.  For projects where there is repetition of dialog (ie commercials and to a less extent drama) good slates and notes are still vital (as well as having the person slating at least announce "marker one" etc).  As was said, PE works better on verite/free form/non-repetitive sorts of material, esp if there are no slates.

phil p

Link to comment
Share on other sites

Haven't posted much, as I enjoy learning from other more experienced people on the forum, but think i may have something to contribute here.

I have used PE to sync footage with DSLR, RED, and DV projects. I have found that you can alleviate confusion by having the slate markings read aloud by the clapper operator in drama situations.

The take numbers will generate discrete wave forms at the head of each file, and thus give a unique to each take file for the program to use. 

I also note that making sure that the scratch audio is clean helps PE a LOT.....

YMMV, but I am happy with this so far...

Chris M in atlanta..

Link to comment
Share on other sites

Second that.

When I get a problem like that I just run a batch with TRY HARDER enabled.

Takes more time.

PE works best if you can lay out all the clips in a row in a sequence with all the audio as well.  For projects where there is repetition of dialog (ie commercials and to a less extent drama) good slates and notes are still vital (as well as having the person slating at least announce "marker one" etc).  As was said, PE works better on verite/free form/non-repetitive sorts of material, esp if there are no slates.

phil p

Link to comment
Share on other sites

...I have used PE to sync footage with DSLR, RED, and DV projects. I have found that you can alleviate confusion by having the slate markings read aloud by the clapper operator in drama situations.

The take numbers will generate discrete wave forms at the head of each file, and thus give a unique to each take file for the program to use... 

Excellent tip.  Thank you. 

I always prefer having the CLAPPER/AC do a voice slate, but sometimes productions are "too cool" to fully embrace "old school" methods.  This gives a good, solid reason why it's as important as ever.l

Link to comment
Share on other sites

Hi folks,

Bruce from Singular Software here.  With respect to PluralEyes syncing to the wrong take when there are very similar takes, we are working on some things to improve that.  If it happens on your project, feel free to contact us at support at singularsoftware dot com and we'll help resolve it.

Bruce

Link to comment
Share on other sites

Hi folks,

Bruce from Singular Software here.

Welcome, Bruce. I saw that you signed up and I'm sure there are many here who will be pleased that you will be around to answer questions directly about your incredibly useful software.

Regards,  Jeff Wexler, CAS

Link to comment
Share on other sites

Welcome, Bruce!!

Your software has been talked about a GREAT deal on this site, and is clearly a fantastic tool.  Congratulations on developing such a useful product.

Robert

Hi folks,

Bruce from Singular Software here.  With respect to PluralEyes syncing to the wrong take when there are very similar takes, we are working on some things to improve that.  If it happens on your project, feel free to contact us at support at singularsoftware dot com and we'll help resolve it.

Bruce

Link to comment
Share on other sites

Yes indeed, welcome Bruce. Glad you made the effort to join jwsound and participate and educate us. I am a user and fan of PluralEyes and have never run into the issue. I do Docs n Internet Context/Product videos when I use your software. We often work off scripts but the readings vary a lot. This is a side business for me and my neighbor. PluralEyes has saved us a lot of time syncing our sound and 5D video. It is really a great product. Thanks.

My main money making job is recording sound for commercials and I see how they could be very similar in their waveform patterns, but I would think my voice slate at the head of the take and I voice tail slate as well, would vary the waveform of each take enough that this issue wouldn't show up. I'm sure whatever issues are at hand the Singular Team will fix it soon.

CrewC

Link to comment
Share on other sites

I would think my voice slate at the head of the take and I voice tail slate as well, would vary the waveform of each take enough that this issue wouldn't show up. I'm sure whatever issues are at hand the Singular Team will fix it soon.

CrewC

I think the problem there would be that head and tail IDs on sound are often spoken when the camera has not yet rolled and/or has already cut, at least in my case.  But on a DSLR (PluaralEyes) show, I suppose it would be easy to adopt the practice of speaking the head ID as the slate is being captured (perhaps at the expense of the clap).  Our show is operated by "film" guys who turn the cameras on at the last second and cut very quickly.  Although this practice (along with waiting for "sound speed" or bothering to cut at all), is becoming more and more rare.

Robert

Link to comment
Share on other sites

When feeding a scratch track to camera, at the beginning of the file, i send a short blast of tone..  This provides PE  with a very clear waveform signature.  I vary the number of blasts and duration just to make each take different. My colleague dave winston suggested this and it really does help!

john.

Link to comment
Share on other sites

When feeding a scratch track to camera, at the beginning of the file, i send a short blast of tone..  This provides PE  with a very clear waveform signature.  I vary the number of blasts and duration just to make each take different. My colleague dave winston suggested this and it really does help!

john.

This also is a very handy visual ref for take starts if you've been rolling wild beforehand (as in verite), and is a really good device for checking sync--esp if the wfms end up being different because the scratch mix is very diff from your production mix.

phil p

Link to comment
Share on other sites

  • 1 month later...

When feeding a scratch track to camera, at the beginning of the file, i send a short blast of tone..  This provides PE  with a very clear waveform signature.  I vary the number of blasts and duration just to make each take different. My colleague dave winston suggested this and it really does help!

john.

Nice one.That's a handy tip!

Link to comment
Share on other sites

Just finished a 30 day documentary shot on a DSLR. I lock my lock it and put it on the shoulder rig for the DSLR. I got a bnc (tc out of lock it) to minijack cable (mic in DSLR) and the camera records the audio timecode. Editor is very pleased. This is the way I shoot on DSLR shoots. We also did a boom clap for safety. The only downside is when also sending a mixdown to the dslr (for example mixdown = left and tc signal = right) that there is some TC bleeding into the mixdown. My TC only cable (bnc to minijaxk) is -20db. And my bnc to stereo minijack is -40 db. I always try to work this way and sort out the sync during the shoot not after!

Link to comment
Share on other sites

In fact, I think this problem could be used in a positive way to match some parts af audio (be it tiny or longer) on recorded parts that aren't clean enough to the editor (honk, bird chip, etc...).

For these times when "yeah, I understand but we have to move on because there's not enough time", you know ? ("have a happy ediitng" is usually my line in these situations) ;-)

So, along with the "try harder" function, there could be a "match something there" function !

Link to comment
Share on other sites

  • 4 weeks later...

i have just completed a job with x2 5d's and a 7d. I used x3 G3 units with hot shoes for the camera mounts. A+B camera where on medium primes and C camera was on a longer lens most of the time. We recorded in double system mode. Most of the performance was not scripted and A+B camera handheld. I recommended that a clapper board be used with no stop starts within the rolls/ takes. a good ID and pip either @ start or end created enough elements to use for syncing the pictures. I even use this process on RED cameras. Also setting the code to time of day and have it appear on the PDF reports helps the offline "track" the rushes with the script supervisor's day log.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...