CS6 / AVCHD / Spanned Clip Bug / What Cameras And Media Types Are Affected?

[joe_bloe_premiere]

This known bug in CS6 has been well documented regarding the potential symptoms, known issues and transcoding workarounds.

Audio and Video glitches | AVCHD footage

It would also be very helpful for CS6 users to be able to specifically identify the problematic cameras and/or media types in order to make informed decisions when purchasing a camera, or when receiving media files with the intent of editing in Premiere CS6.

If there is a definitive list of specific camera manufacturers / models, and/or media types that are affected by this bug, I have not been able to find it.

Of course this bug has been resolved in the June 17th release of the Creative Cloud Subscription update. It remains to be seen if the fix will be back-ported to CS6.

Thanks.

Cameras confirmed as not exhibiting the bug:

Panasonic GH2

Sony HDR SR-12

Cameras confirmed that do have the bug:

Canon Vixia HF G10

[jasonvp]

That's why I asked for model numbers to be posted

once the clip spanning bug can be confirmed.

I'm sorry I wasn't clear in my first post: it's 100% confirmed on both of those models.  I just don't remember what they are at the moment.

jas

[joe_bloe_premiere]

Thanks for the clarification, Jason.

[jasonvp]

Stan Jones wrote:

Well, I have several Handycams thaq shoot Digital 8!

And obviously, those Handycams are too old to matter for this discussion.  Some level of assumption is made when replying here. 🙂

As I noted above, Sony has added weird stuff at times that does not follow.

Perhaps, but look at it from a money perspective: if the code is written for AVCHD in one of their cameras, why on Earth would they change it for another camera?

Further, Sony is one of original creators of the AVCHD standard.  So it's perfectly safe to assume that all of their AVCHD-supporting cameras are identical.

jas

[Stan Jones]

And obviously, those Handycams are too old to matter for this discussion.  Some level of assumption is made when replying here. 🙂

Yep; it has me giggling again!

if the code is written for AVCHD in one of their cameras, why on Earth would they change it for another camera?

Further, Sony is one of original creators of the AVCHD standard.  So it's perfectly safe to assume that all of their AVCHD-supporting cameras are identical.

First, just as in the PDX10 scenario I cited, Sony has made changes, screwed up, and changed back. For creating an authoritative list, I just don't think we should make that assumption (yet).

Second, we don't know what causes this problem. While we can assume that Sony makes all its AVCHD cameras to conform to their specs, it is not clear that they have not, from time to time, modified anything in their AVCHD cameras that might interact with Adobe's code. We really don't even know that the fact that not all AVCHD spanned clips have problems is related to specific cameras/media types.

Given the range of claims in many posts, I will not be at all surprised to see someone list a camera as showing the bug, and another user listing the same camera as having no problems.

Kevin, can you provide the bug number for this problem? Can you get a response from the software engineers as to the believed source of the bug?

[joe_bloe_premiere]

Kevin, can you provide the bug number for this problem? Can you get a response from the software engineers as to the believed source of the bug?

Nice.

[Stan Jones]

You mean as in "not" nice?

I have complimented the engineers when they occassionally have been very forthcoming with behind the scenes info - e.g. the reason it was not just a tweak to add smart rendering of mpeg files to Premiere. I really do believe that we would all rather do good than not, as long as it isn't too inconvenient! But transparency can be expensive.

I do wonder if we are at a point where they might give us a bit of data to save us the exercise in this thread. They may know it's futile....

And don't someone pipe in and remined me that we already know resistance is futile... I know, I know.

[Steven L. Gotz]

I suppose I should get around to testing the Panasonic DMC-GH3 to make sure that the clips work OK. I think it stops at 4GB and creates the next file starting at the end of the first one.

I will set it up right now and tape for a while just to see what happens. I have to change from shooting .mov to shooting .mp4 I guess. I'll give it a try.

Edit: I have a choice of mov, mp4 and what they call AVCHD so I set it to AVCHD and started it running. But at 28Mb/s it is going to have to run for a while befor it hits 4G. What, something like 20 minutes or so?

I'll be back.

[Steven L. Gotz]

OK, I'm back. The Panasonic has no AVCHD issues with Premiere Pro. Also, I did not have to copy over the entire directory. I just copied 00000.MTS and 00001.MTS over to my hard drive and put them on the timeline. The changeover from the first file, which was 20 minutes 25 seconds long at 4GB, to the second file which I stopped about ten minutes into it, or at just over 30 minutes total (no European 29:59 rules for my camera), was seamless. My wifes arm was moving at that time and I can see that there is not a single frame missing.

That should take the GH3 off of Steve's list of worries, although I doubted the problem would be in the GH3 if it was not in the GH2.

[joe_bloe_premiere]

Steven,

I appreciate your efforts, but since the bug seems to directly involve

the spanning of contiguous clips through utilization of the metadata,

shouldn't your test files be imported via Media Browser instead

of bypassing the metadata by copying the individual media files?

Thanks.

[Stan Jones]

shouldn't your test files ...

I had a related question yesterday: is there a protocol for demonstrating the bug? As illustrated by the thread Fuzzy pointed to, there are lots of false positives: some users were bypassing folders and getting problems which were eliminated by using the intact folders. But were those problems ever the spanning issue?

Any of us might create a protocol, but is there one already?

Steven's version raises the empirical question: does a failure demonstrated with the full folder structure not fail if using individual files? Can a failure be shown with only a single span?

[joe_bloe_premiere]

These are excellent questions that beg for clarification from Adobe.

Your final request in post #34 would be a great place to start:

Stan Jones wrote:

Kevin, can you provide the bug number for this problem?

Can you get a response from the software engineers as to

the believed source of the bug?

At this point, I think we must assume that the proper method of import

as described by documentation (copy folder structure intact from card,

use Media Browser for import) should be the benchmark for

judging whether or not a camera's media is affected.

[jasonvp]

Stan Jones wrote:

I had a related question yesterday: is there a protocol for demonstrating the bug?

I did this last year, but I removed all the sound from it because I was going to narrate it.  But I failed epically at attempting to create a narration track and just skipped it altogether.

Watch the program monitor jump around while I'm trying to play, scrub backwards, scrub forwards, reset to the beginning, etc.

jas

[joe_bloe_premiere]

Thanks for the post, Jason.

This is a great visual visual depiction of how

the bug is manifest in the Premiere Timeline.

I think the 'protocol for demonstrating the bug' Stan mentions has

to do with the method of import (full copy of entire folder structure to

HD, Media Browser or not, single spanned clip vs. more than two, etc.).

How did you import this media?

How many individual spanned clips does it utilize?

What camera / model is the source for this media?

What is the file type of this media?

[jasonvp]

joe bloe premiere wrote:

How did you import this media?

How many individual spanned clips does it utilize?

What camera / model is the source for this media?

What is the file type of this media?

Cameras: Sony Handycam HDRCX300 and HDRCX130.

I copied the entire SD card directory structure to the hard drive, then pointed Premiere Pro's Media Browser at that new directory for the import.

It includes 2 spanned clips (one from each Handycam), each having 3 individual .MTS files as part of them.

Does that clarify it?

jas

[joe_bloe_premiere]

Perfect, thanks!

[joe_bloe_premiere]

Sorry to trouble you Jason, but Stan's musing:

"...does a failure demonstrated with the full folder structure

not fail if using individual files?"

I don't suppose you still have the media files to test this hypothesis?

[jasonvp]

joe bloe premiere wrote:

I don't suppose you still have the media files to test this hypothesis?

I keep all of my media, so testing it shouldn't technically be difficult.  Finding the time to do so, on the other hand, is another story.

jas

[jasonvp]

joe bloe premiere wrote:

"...does a failure demonstrated with the full folder structure

not fail if using individual files?"

I don't suppose you still have the media files to test this hypothesis?

I can confirm that, with my source footage, the problem exists with individual MTS files or the entire directory structure.  It doesn't appear to matter.

jas

[joe_bloe_premiere]

I can confirm that, with my source footage, the problem exists with individual MTS files or the entire directory structure.  It doesn't appear to matter.

Hmmmm... Very interesting.

Thanks for the report.

[Jim_Simon]

I can confirm that, with my source footage, the problem exists with individual MTS files or the entire directory structure.  It doesn't appear to matter.

Are you sure they came in as individual clips?  When I tried it using the File>Import... method, PP still saw each one of the files as the full length clip, and brought each of them in not as 20 minute segments, but as full 50 minute clips.

[jasonvp]

Jim -

Jim Simon wrote:

Are you sure they came in as individual clips?  When I tried it using the File>Import... method, PP still saw each one of the files as the full length clip, and brought each of them in not as 20 minute segments, but as full 50 minute clips.

Ah HA!  You did the same thing I did on attempt #1: imported the .MTS files directly from within the AVHCD directory structure.  When you do that PPro knows what you're up to, grabs the appropriate metadata, and just adds the entire clip.

Attempt #2 is what I ultimately settled on: copying the .MTS files out of the AVCHD directory structure (into a tmp directory), and importing them from there.  When doing that, PPro had no idea what  belonged to what, and just imported them as individual files.

jas

[Steven L. Gotz]

OK. I'll bite.

What is wrong with that? Why not individual files if they are perfectly matched one frame to another? What is wrong with three clips instead of one if they look fine on the timeline together?

Damn. I am missing something important here, I know I am, and I can't figure out what it is. They are numbered. Just slap them on a timeline. After all, how many times do you shoot more than an hour of video at one time? That's three, maybe four files?

Aarrgghh!  Can someone help me get this through my thick head so I can understand the problem?

[Stan Jones]

What is wrong with that?

For the purpose of this thread, nothing, as long as that is the protocol for demonstrating the bug. Since there appears to be no single protocol, I see it also as a test as to whether the bug is really related to metadata or not. (I'm accepting the comment that the single file pulled without access to its metadata will not have it.)

[Steven L. Gotz]

What I am missing is the reason that so many people are upset with CS6. If the problem comes from importing the files a certain way, and importing them a different way works just fine, then what am I missing?

I get that they are VERY upset. I just don't get why and I pride myself on understanding Premiere Pro and this is a hole in my knowledge that I would love to fill. I understand the test procedure better now I think, but I don't understand what is so terrible about not having one large file instead of three or four, or even more, as long as they are perfectly sequential.

[jasonvp]

Steven L. Gotz wrote:

What I am missing is the reason that so many people are upset with CS6. If the problem comes from importing the files a certain way, and importing them a different way works just fine, then what am I missing?

The point: Premiere Pro is supposed to "just work" with AVCHD without the need of transcoding.  And, prior to CS6, it did that.  Several of us can use the same footage in CS5.5 without any issues whatsoever.  But when we edit it in CS6, we have all sorts of fun (fsvo: fun).

Does that make sense?

jas