Update to help native capturing into FCP?

[Burt Holland]

I was only able to attend NAB on the last day, but I was able to make it over to the JVC booth and of course asked one of the reps about the dropout when digitizing to FCP issue. To my surprise he said, "We just announced an update today that you can download that fixes that issue." He was in fact very happy about the ennouncement as he said he finally had an answer to give people asking that question.

I have been traveling since NAB and haven't had a chance to look, but now that I'm back I see nothing about it on the JVC site or here. Was I just dreaming? He had told me it would take a week or two to get the patch up, but I don't even see an announcement about it. Did anyone else hear about this?

Burt Holland
Encyclomedia
Atlanta, GA

[Tim Dashwood]

That's news to me...

I spoke directly to the Apple guy about it and as far as I can tell the only newsworthy announcement was 720P60 native support in a "future" version of FCP6. This means that the version to ship in FCS2 might not have the 720P60 support, and no one knows when that patch will be released.

There are a couple problems associated with ingesting natively in FCP successfully from tape in 720P24. The most obvious is the fragmented files FCP creates. The reason it affects FCP and not other NLEs is Apple's zero tolerance for any missing data in the mpeg2 stream. They seem to be demuxing the stream in ingest and re-writing the audio to 16bit 48Khz (as far as I can tell) and removing the pulldown frames. Any interruptions detected are immediately flagged as 'breaks' and the file is cut off and a new one begins (that's when the 5-6 second gap happens.) This process in itself suggests that very fast hard drives and fast processors will perform better (shorter gaps) than slower G4 or G5s. I use a Macbook Pro 2Ghz with 2BG of RAM without issue.
However, I become more and more convinced everyday that the faulty data streams are primarily caused by using non-recommended tape stock and not cleaning the heads on a regular basis. I ran into a few people at the booth who wanted to specifically tell me that they don't have the problem and capture long clips all the time. I asked these shooters what stock they used and I believe they all said the ProHD tape stock.
I spoke to about 9 or 10 people over the course of 4 days who do have problems capturing. Most of these people told me they use Panasonic MQ tape...some said Sony, but I'm not sure which flavour.

The rough statistics seem to support my theory.... but where's the proof?
I've decided I'm going to have to conduct a formal comparison. I just have to decide which camera to sacrifice and do it at a time when I have no critical work on the table.

In the meantime, I'd like you to ponder this:
Why would JVC go to the bother of creating a brand new tape stock specifically just for this camera system if it wasn't necessary? I truly believe they realized a very stable tape stock was going to be necessary for HDV, and went to the trouble for that reason.

I urge anyone with these problems who has been using Panasonic MQ to buy a head cleaning cassette, run a few cycles, and then start using JVC ProHD stock. Please let us know if it works any better on ingest.

[Jonathan Nelson]

I talked to [a JVC rep in the booth] and he told me there was a new update coming out in a few weeks that fixed the fcp problem.

I get really bad fragmented capture. I am not sure what causes it but when I recapture, the breaks happen at different places which leads me to believe that it is a much broader problem than using the wrong tape stock or dirty heads.

[Burt Holland]

Originally Posted by Tim Dashwood

There are a couple problems associated with capturing in FCP successfully from HDV tape in 720P24. The most obvious is the fragmented files FCP creates. The reason it affects FCP and not other NLEs is Apple's zero tolerance for any missing data in the mpeg2 stream. They seem to be demuxing the stream in ingest and re-writing the audio to 16bit 48Khz (as far as I can tell.) Any interuptions detects are immediately flagged as 'breaks' and the file is cut off and a new one begins (that's when the 5-6 second gap happens.)

This was essentially the explanation I was given by the rep. I had always heard that it had to do with the GOP structue. I am far limited in my technical understanding, however, so I don't know if these two are connected or not.

I might have some time this week to try the Panasonic MQ tapes. I've used the Sony HDV tapes in the past, but never really saw enough of a difference to warrant the expenditure for my applications. I've always used Panasonic DV.

Burt Holland

[Tim Dashwood]

If Apple were to lower the stream rejection tolerance then I could see it being a simple fix. But what would happen to audio sync on long clips? Are dropped frames or disrupted frames an acceptable compromise?
Maybe they figured out another way to deal with it? I don't know.

[Robert Castiglione]

Tim,

This is a very elusive problem. I thought I had the problem solved but it just came back completely randomly. I use JVC Pro HD tapes and it makes no apparent difference to the issue. My camera is pretty new and does not have a lot of hours on the head yet. How often would you recommend cleaning the heads? I am prepared to consider anything.

Lets hope that the JVC rep is on the money. I have been using the DVSCAP/MPEGstreamclip method and cant wait to just capture directly and trouble free to FCP - it seems a distant dream.


Rob

[Tim Dashwood]

Originally Posted by Robert Castiglione

I use JVC Pro HD tapes and it makes no apparent difference to the issue. My camera is pretty new and does not have a lot of hours on the head yet. How often would you recommend cleaning the heads?

I run the head cleaner before every major shoot, which averages to probably 8 - 10 tapes.

Have you ever used a different brand of tape on those heads, or just stuck to ProHD from day1?
Out of curiosity, do you have an Intel processor or G4/G5? How much RAM?
Have you tried the same tapes in a different camera?

I know it sounds like we are searching for the needle in the haystack, but any additional information would be helpful to try to solve this problem.

[Robert Castiglione]

Thanks Tim. Pertinent questions and here are the answers.

I am a bit surprised at using the head cleaner so often. I have shot about 40 hours and have not used it once yet.

As to tape I used to use Panasonic and then switched after about 20 tapes to JVC.

I use a G5 2 ghz (non intel) with 2 gigs of Ram.

What brand of head cleaner do you use? This sounds like the first thing that I should do.

Rob

[David Knaggs]

Originally Posted by Tim Dashwood

I run the head cleaner before every major shoot, which averages to probably 8 - 10 tapes.

1/ Have you ever used a different brand of tape on those heads, or just stuck to ProHD from day1?
2/ Out of curiosity, do you have an Intel processor or G4/G5? How much RAM?
3/ Have you tried the same tapes in a different camera?

I know it sounds like we are searching for the needle in the haystack, but any additional information would be helpful to try to solve this problem.

Hey, Tim. Let's find that needle!

Here's my info on this matter:

1/ When I first got my camera (July 05) the Pro HD tapes weren't available, so I used about 10-15 Panasonic MQ tapes (AY-DVM63MQ) then switched to Pro HD tapes once they arrived and have used them ever since. It's had a number of cleans (since the switch) with the Panasonic head cleaner. But I don't think it's had "a few cycles" at the one time as you suggest in your earlier post.

2/ 1.9 GHz G5 with 1.5 GB RAM. Also, I normally capture to external drives (La Cie and Seagate) with FW400 connections.

3/ No.

Extra info:
a) On possible variation within Pro HD tape stock itself, I shot an event on the weekend (2 tapes). The first tape was from my own Pro HD stock (grey-black casing) and the second Pro HD tape was provided by someone else (white casing). When I got home and captured, the entire white tape captured in one go. Then I put in the grey-black tape and, although I shot the event continuously, it broke it into about 14 different clips. It made me wish I'd shot the event with two white tapes! But it also made me wonder whether they use more than one facility to manufacture Pro HD tapes and whether standards (or the materials used) may vary from place to place.

b) Once there are problems, the longer you go, the shorter the clips get. It might start breaking the clips at 3-4 minutes each. But, as you continue, it can get down to about 15 seconds. I attribute this to either the camera heads getting too "warm" or the computer/hard drives and then turn the camera off and the computer to sleep for a while, come back later, reconnect and usually find it will capture the remaining tapes in much bigger chunks (like 12 minutes each). So this point tends to be consistent with your theory about the camera heads and/or the computer system.

[Daniel Weber]

I have just started to have problems with capturing and playback of Panasonic tapes.

I am ready to purchase a Firestore and skip tape all together.

Tim, you have a Firestore, what has been your experience with it?

Right now I really hate tape!!!

Dan Weber

[Drew Curran]

Originally Posted by Daniel Weber

Right now I really hate tape!!!

I'm with you there - and we're probably not alone...

I hope that solid state DTD recording arrives soon. I'm a bit afraid of spinning hard drives used for recording important footage. I know you can record to tape also.

I hope that someone creates an external device that records to express cards (cheap) and connects to the camera via firewire - giving a similar work flow to P2. That way we are not limited to the size of the internal hard drive like the firestore. We simply take the express card and plug it in to a card reader on the computer, copy it to a hard drive and begin to edit. No capturing required.

Why does this sound too simple?

Drew

[Dave Beaty]

Split Clips

I've been thinking it has something to do with the way the MPEG-2 is laid down on tape, thinking that since Dec. But when time I mentioned tape, I was told it's a FCPro problem.

But the one thing I notice, if you look very carefully at a monitor, you'll see a 1-2 frame visual stutter right at the break point. (Most times, not always) Like the stream has errors. I have tried 2-3 brands of tape both Sony HDV and Panasonic MQ the top of the line and it happens with both. Have not tried JVC. We don't head clean every shoot.

Seems like the further into the tape, the more breaks...so as mentioned above, it may have to do with heat on the heads/tape over time or tension on the tape.

Originally Posted by Tim Dashwood

However, I become more and more convinced everyday that the faulty data streams are primarily caused by using non-recommended tape stock and not cleaning the heads on a regular basis. I ran into a few people at the booth who wanted to specifically tell me that they don't have the problem and capture long clips all the time. I asked these shooters what stock they used and I believe they all said the ProHD tape stock.
I spoke to about 9 or 10 people over the course of 4 days who do have problems capturing. Most of these people told me they use Panasonic MQ tape...some said Sony, but I'm not sure which flavour.



I urge anyone with these problems who has been using Panasonic MQ to buy a head cleaning cassette, run a few cycles, and then start using JVC ProHD stock. Please let us know if it works any better on ingest.

[Dave Beaty]

Why was this thread moved to Area 51?

This topic is one of the most important issues of the JVC ProHD cameras.

IMHO

Dave Beaty

[Matthew Groff]

Just did some fact checking on the Apple website and the HD100 is NOT technically supported by any flavor of Final Cut (including the forthcoming FCS2) and as such are we really sure that Apple will ever fix this? The BR-50 is "supported" but only in FCP 5.0.X (not shocking in and of itself) AND then only in 720p30 mode.

I have to say I'm pretty bummed about this -- is there any official word that Apple is going to address this issue? If i'm going to have to DVHSCAP and transcode everything anyway, maybe Avid's DNXHD is a better workaround than DVCPRO HD.

BTW, as a workaround idea to maintain proper timecode, has anyone successfully used the header record function such that when capturing via DVHSCAP the first all black frame after the bars and tone can be reliably referenced as an exact timecode? I know this is easily modifiable in FCP (i.e. off the top of my head, my first black frame is always 23:58:30:10) so I can go to that frame and Modify > Timecode to set it correctly. Is there an equivalent modification function in Avid's Xpress?

[Tim Dashwood]

Originally Posted by Dave Beaty

Why was this thread moved to Area 51?

This topic is one of the most important issues of the JVC ProHD cameras.

IMHO

Dave Beaty

Unconfirmed rumours and speculation of software or firmware updates belong in Area 51. The thread will be moved back to the JVC forum if/when the magical fix is released.