AVCHD saga continues...

After hours of reading testimonials of AVCHD transcoding/transfer issues my excitement for the HMC is curbing drastically. Seems that some have few issues but most have significant issues. Seems that AVCHD is not a really stable codec so far...SO FAR. I have hopes that this will change. I'm an FCP user.

That being said here's something I ran across that may help someone out there. I don't understand what he means. Deleting a codec set from the library? This effects log and transfer? Not sure. Anyone?

HERE'S AN EXCERPT:
"I figured out what was wrong. It was a Quicktime Codec I had downloaded a couple of months ago. I deleted a few different codecs that I thought it might be, so I can't be 100% sure which one it was, but I have my suspicions that it was the DV1 codec. after I deleted it, Final Cut was able to transfer the footage without any problem."

The problems he described were motion artifacting and audio synch related.

HERE'S THE URL FOR THAT THREAD:
http://forums.creativecow.net/thread/8/1009038#1016557

Hope this helps. Can anyone explain why this would work? Not even sure I know what THIS is but maybe somebody more software savvy does.
 
The codec is fine and stable. The particular editing stations may not have implemented it properly yet. But your footage is good and will continue to be.

Are you running FCP 6.0.5? Make sure you've got that version; it was updated specifically for HMC150 compatibility.
 
The codec is fine and stable. The particular editing stations may not have implemented it properly yet. But your footage is good and will continue to be.

Are you running FCP 6.0.5? Make sure you've got that version; it was updated specifically for HMC150 compatibility.
Yes I will be using FCP 6.0.5. Should it show up as AVCHD in the CODEC list within say the sequence settings?

Hoping to borrow/rent a camera and do a test on my platform which is MacBook Pro dual 2.53 GHZ

Did that thread I quoted from creative cow make any sense to you?

Should clarify that I am only researching now so don't have a way to test things first hand so am only relying on forums and it is true that forums tend to display more about issues than successes.

I just like to know what I'm in for before I commit. Specially if dealing with deliverables to varied clients with different post platforms so am looking into all potential hang ups and solutions. Luckily most my clients are FCP 6.0.5.

My gut tells me that if FCP reports support for it and FCE reports support for it and Panasonic stands behind it that it has legs and will be fine but sometimes reading these convoluted testimonials of issues gets to my superstitious side.

Nice books by the way! Great tips. Glad they exist!
Thanks for the reply.
 
I bought my HMC150 about a month ago. And yes its true, I hav had a few workflow issues, such as long transcoding times, etc (I do have an older comp system, so that's largely to blame there). But I am sticking with it, because it offers a great combination of super clean CCD's with great low light sensitivity, extremly low noise, AND IT'S NOT HDV!

I just completed a project I shot for my daughters school--a holiday concert, shot in very poor lighting, and wow, is it clean--even in its downconverted SD version, it is one clean machine.
I am transcoding to DVCProHD using the transcoder from Panasonic's site and yeah, on my 2 yera old machine its very slow--but the results!!!
I am now planning on getting a new quad core Uber-machine maybe by February to solve that problem. This cam has alot of potential. Its initial workflow challenges will be worked out and it'll be the hot ticket.
 
I have had some issues with AVCHD. The transform to P2 doesn't always work. I am using Pannys P2 transcoder. In certaing files the program simply doesn't start to transform, it jams before the actual t-coding should start. Don't have a clue why. The file runs ok in camera and is also editable directly in EDIUS 5 (without transcoding). With P2 I have never had any problems. AVCHD comes from AG-HSC1UE and P2 from one of the first HVX200
 
FCP 6.0.5 imports HMC150 AVCHD as a transcode to pro res.

You can use Toast 9, or VideoPier HD to transcode to DVCProHD if you wish.

A professional editor will be cautious about downloading and installing esoteric codecs.

Not certain why an editor would wish to do this.
 
After hours of reading testimonials of AVCHD transcoding/transfer issues my excitement for the HMC is curbing drastically. Seems that some have few issues but most have significant issues. Seems that AVCHD is not a really stable codec so far...SO FAR. I have hopes that this will change. I'm an FCP user.

I'm a FC Express user, and while at the moment there is no 'native' Express support, I can log an transfer, with a transcoding to AIC, with no problem.

I currently have a Sony HDR-S11 which uses the AVCHD codec, and I've been transcoding that device for about 3 months now with 'no problem'...

I rented an HMC150 for a shoot, and also had no problem there except....

I shot in 1080 24p mode, so while Express transcoded to an AIC form which was 1080 24p, Express only supports sequence timelines which are PAL or NTSC compatible... that is either 25 or 29.97 fps. So, I selected the 25 PAL-esque option, and I believe I have a few clicks on the video due to slight timing mismatch... I recently got an upgrade to CS4 for my Produciton Bundle and will test on Primiere and the purported 'native' support for AVCHD... but not too soon...

The alternative was to use After Effects, but that is really abysmal for doing standard splice and dice on footage...
 
I'm a FC Express user, and while at the moment there is no 'native' Express support, I can log an transfer, with a transcoding to AIC, with no problem.

I currently have a Sony HDR-S11 which uses the AVCHD codec, and I've been transcoding that device for about 3 months now with 'no problem'...

I rented an HMC150 for a shoot, and also had no problem there except....

I shot in 1080 24p mode, so while Express transcoded to an AIC form which was 1080 24p, Express only supports sequence timelines which are PAL or NTSC compatible... that is either 25 or 29.97 fps. So, I selected the 25 PAL-esque option, and I believe I have a few clicks on the video due to slight timing mismatch... I recently got an upgrade to CS4 for my Produciton Bundle and will test on Primiere and the purported 'native' support for AVCHD... but not too soon...

The alternative was to use After Effects, but that is really abysmal for doing standard splice and dice on footage...


I have also tried to transcode from a Canon HF10, and that worked fine as well. Seems the transcoder to DVCPRO will work on more than just panasonic AVCHD files.
 
So it sounds like it depends a lot on the platform but seems that, at least from this thread so far, that transcoder software handles AVCHD much more consistently than NLE software making transcoding a potential necessary step to using the footage. Currently.

It seems reasonable to expect the bugs with native support for AVCHD to get worked out over time so that those who own the HMC or other AVCHD cameras will be able to benefit from one of the main advantages of solid state field acquisition - less time to load footage.

While transcoding is not as tedious as digitizing it is still another step that, on some systems, can be very time consuming for the machine.

It does sound that the codec itself (AVCHD) is stable within the confines of the recording camera and therefore a reliable acquisition codec. I have yet to read anything to suggest otherwise on this point. Which is comforting.
 
They provided the file converter to P2.

Yes, they did, or MainConcept did.

And they provided it to only part of its owner/purchaser/user group.

I'm not trying to be difficult, but this is one thing about the company that perplexes so many Mac users.
 
How fast does the provided P2 converter run? realtime? faster? much slower? And is this converter the best workflow from an image quality perspective if time of conversion is not of concern?
 
The best workflow as far as image quality perspective is to not convert at all. EDIUS, Vegas, Pinnacle and other editors just work with the footage.

For those who have no other alternative, the converter works at around realtime depending on how fast your system is; an 8-core is going to clearly deliver results tremendously faster than a 4-core will.
 
The best workflow as far as image quality perspective is to not convert at all.
but what about this?
http://cineform.com/products/NeoScene.htm

It converts AVCHD to Cineform AVI and ups the precision to 10bit and doubles the chroma to 4:2:2 and therefore claims the following...

"CineForm AVI files maintain much higher visual quality and higher editing performance than camera source files."

converting 1920x1080 8bit 4:2:0 24p to 1920x1080 10bit 4:2:2 24p seems like a no-brainer since most NLEs can handle Cineform AVI much easier. BUT this also seems to go against what I always thought...you cant increase color or resolution beyond what the camera records natively.

So although the footage would not gain resolution/color info- it would'nt loose anything (except it's small file size) :). So wouldn't it be better to transcode into this lossless codec that can handle post processing more effeciently like what Cineform claims in the above? I cant remember the last time I edited something that turned out to be cuts-only, there is always some transition, or color/gamma correction in post.
 
Last edited:
Cineform isn't lossless, although it's pretty good and holds up to recompression very well.

You can't increase resolution or color space beyond what's there already; I mean, you sort of can because you can use InstantHD but it'll be "guessing" what to fill in. However, where something like Cineform shines is when you do multiple generations of recompression. Working in a 4:2:2 color space with a mildly-compressed codec will pay dividends if you have to render down or export footage pieces to have them be worked on in other applications.

Cineform (or any intermediate/editing codec, like AIC or CanopusHQ or ProRes) ask you to make a compromise up front (the time and additional size to convert your original files to their editing codecs). Once you're past that initial hurdle and pay that initial price, you then get much speedier edit performance. But when computers are fast enough to handle the native files, then the appeal of going to an intermediate will dissipate some. Especially for those who don't do intermediate renders.

For someone who's going to be doing a ton of VFX work, a high-quality intermediate (or uncompressed) system is going to be better than working in the original footage. And that's what Cineform is claiming: they "maintain" higher visual quality (meaning that the footage doesn't break down nearly as much under generations of recompression). But think about it, how many of us transcode our DV footage into something else? How many of us transcode our DVCPRO-HD footage into something else? I dare say just about nobody. And while transcoding HDV was almost necessary in the early days, as computers picked up speed and editors became more efficient, how many HDV shooters are still transcoding their footage vs. how many are just using the native files? I don't know the answer, but I'd be willing to bet there's probably a ratio of at least 5:1 for those using the original files vs. those transcoding to an editing codec.

So is it "better" to transcode into an editing codec? Depends on what you're doing with the footage. If it's not going to look any better, and you're not going to take advantage of the recompression capabilities that the editing codecs provide, then why are you doubling/tripling/quadrupling your space? But if you are going to take advantage of it, then certainly there are benefits to be had.

All I'm saying is I expect AVCHD will get that way too. And for those who need immediate turnaround (wedding shooters, news, etc) I doubt they're going to be too anxious to pay that upfront conversion time penalty, vs. just cutting straight from the original footage.

There are certainly benefits to be had from the Cineform approach. If you simply cannot get the performance you need to edit the native footage in real time, but your computer *can* handle the CineForm AVI files, then that right there is reason enough for many folks to go with the intermediate codec approach. It just depends on how you work and what you're going to do with the footage.
 
But think about it, how many of us transcode our DV footage into something else? How many of us transcode our DVCPRO-HD footage into something else? I dare say just about nobody. And while transcoding HDV was almost necessary in the early days, as computers picked up speed and editors became more efficient, how many HDV shooters are still transcoding their footage vs. how many are just using the native files? I don't know the answer, but I'd be willing to bet there's probably a ratio of at least 5:1 for those using the original files vs. those transcoding to an editing codec.

OMG! Dude, remember Lumiere HD? This was one of the first dedicated programs to transcode this stuff. The site is actually still up, it's Lumierehd media now.

-Nate
 
Yesterday I shot ~50' of footage to the Panasonic 16G card; it was about 9.5Gb in all.

Importing the footage to the third internal HDD took eight minutes. This was a simple copy from card to disk, limited by USB2 speed.

Opening the Log and Transfer window; all clips appeared (it sees the last drive/folder you pointed it at).

I selected all clips, dragged them into the Queue window, and FCP went to work immediately. Transcoding to Pro Res 422 takes about 40% real time as far as I can see; I went to lunch. I loged all clips in the afternoon; all are perfect.

My take is this: the workflow for me is absolutely fine as it is. Not a dropped fame to be seen; the transcoding is automatic; the files once expanded are about four times the originals in size. All audio is 48KHz, and the images are excellent.

From my POV, there are no problems with the FCP native transcoding, and no downside. My workflow involves keeping the original AVCHD footage (as it is compressed, it is already in an excellent form for storage). If I do whoot anything that might be able to benefit from a no-transcoding workflow later, it will be there in its original form. The expanded Pro Res files sit on my fastest external drive, and FCP seems to be completely happy working with it. Personally, I don't miss tape at all, and the transcoding reminds me of batch processing in Adobe Camera Raw: an opportunity to get something to eat, or to go for a walk. Way faster than logging and capturing tape, and no dropouts (did I mention that?).

if you don'y use FCP, there is either Premiere, which sees the AVCHD files natively, or the Panny transcoder (transcdes to DVCProHD, from memory). Seems like a very workable solution to me.
 
Last edited:
Back
Top