I am a first-year Film & TV teacher teaching at a school in Australia.
The school I'm at is quite regional and is using fairly outdated/low budget equipment. They are using JVC Everio cameras (GZ-HM445BAA) and Adobe Premiere CS4. The computers my students are editing on are quite slow and tend to crash often.
Research has told me that editing raw AVCHD .MTS files natively from the cameras is not the best option, as these tend to be very processor intensive and this is the reason why Premiere is constantly crashing on these computers. Videos that are imported into premiere tend to look strange, even with the (I assume) correct sequence settings of AVCHD 1080P 25FPS. It also takes forever to render anything.
I'm wondering if there is any way that I can convert the files to a different file format (ie: MP4/H.264) in order to make the editing workflow smoother. I've tried converting the files to H.264 using the preset of HDTV 1080p 25 (which is what I assumed would be correct) but the coverted videos look choppy, as if something is wrong with the framerate.
I'm confused about what I need to convert my files to, as I've had trouble finding the proper specs of our cameras online. Do the converted files look weird because the cameras are shooting interlaced and I'm trying to convert to progressive?
Can anybody offer any assistance?
+ Reply to Thread
Results 1 to 24 of 24
When you say "slow" computers, in what way are they slow? Hard drives/StorageBuss? CPU? Video? My guess is "all of the above".
That means, you can't "rob Peter to pay Paul".
AVCHD is a consumer/prosumer format, derived from BD, that uses AVC (aka MPEG4 h.264) is its video compression. Your idea of converting your media to a more edit-friendly format right off the bat is a step in the right direction, but you suggest as an example "MP4/H.264"! Which is nearly the same thing as what you are already using! (Different container/wrapper, different bitrate, that's about it).
You also don't want to go the opposite direction and use fully-uncompressed files, as I doubt your HDD/Storage subsystem is able to handle the jump in data retrieval rates (and then THAT would become the bottleneck). No, you need something in-between. Something that has low CPU overhead (not complex decoding like h.264 is) and also low/medium datarates.
You could try losslessly-compressed files (HuffYUV, Lagarith, UTvideo, ffvideo) if your HDD/Storage subsystem can handle it without stuttering (this is also a good way to fully retain your quality until the very end when you make consumer distribution copies).
If that is still too much data, then drop your bitrate requirements down to the 100-200Mbps range and use a "visually lossless" lossy codec (DNxHD, Cineform, ProRes). It is somewhat more complex than Uncompressed or Lossless, but because it is still all I-frames, it should be smooth to edit with. Plus, although it is not strictly speaking a truly lossless format, you will be hard pressed to notice the generational loss unless you are doing serious processing or using many many generations-down material (re-used layers?). BTW, if this is happening (probably due to not being able to fully nest your effects), I would suggest using an alternate path for those functions such as AVISynth frameserving.
If that is still too high for you, I would suggest using 50-120Mbps DVCPro, MJPEG, or All I-frame MPEG2 formats. This would have quality loss that would begin to be noticeable, but the loss wouldn't build too much (until, say 7-9 generations down).
AFA Interlaced & Progressive goes, it depends on what you are outputting to/for. If your final medium is something which supports interlace (even if the display is progressive - where it will have its own I->P conversion), I would suggest LEAVING it Interlaced throughout your cycle. If that's not an option, then I would suggest you UP convert your rate (from 50i->50p, not DOWN convert it (from 50i->25p). Going down will add noticeable a "STUTTER" look to the motion. Of course, 50p has its own drawbacks (many media still don't handle 1080p50/60 correctly and in those cases, you'd want to create a copy specifically down rated for 25p).
I did a search on JVC's site and found your user manual: http://manual.jvc.co.jp/index.html/
Just narrow the search for Oceania/Asia, English, Model GZ-HM445 (the "BAA" is specific to optional sub-model specifics incl. colors and certain vendors like "Walmart").
This shows that yes, your cams do shoot AVCHD. IOW, h.264 in an MTS (MPEG-Transport stream) container, using the AVCHD compliant folder & naming structures. The max bitrate available to that cam is 24Mbps @ 1080. Doesn't say whether that is 1080i50 or 1080p25. It wouldn't be 1080p50 because that is not compliant with the original AVCHD spec, only the AVCHD 2.0 spec, which this cam doesn't support.
<edit>Checked the AVCHD site, and it says 25p is NOT supported for 1080 in v1.0, so you must evidently be shooting 1080i50</edit>
Hope that gets you further along,
Last edited by Cornucopia; 27th Mar 2014 at 02:34.
So what's happening is the footage format is too advanced for your computer, so you need to re-encode to an older codec like Mpeg2, like mentioned above by Scott.
But! Your current format is interlaced, which is used for broadcast. So you need to decide if you want to de-interlace it too. And that opens up another can of worms.
Use a lower resolution in your preview window to improve playback.
Last edited by budwzr; 29th Mar 2014 at 20:20.
I don't see any particular hardware in your profile nor in your post. You want to convert from the AVCHD (which is really a -delivery- format) to something with, preferably, no temporal compression. Any format that utilizes key frames and only changes between the key frames will be problematic. I'd stay clear of MPEG4 and even MPEG2 and, instead use something older like MJPEG. Converting to MJPEG will re-create each frame as -complete- so you can do frame-accurate editing. There are probably other good formats for Premiere that will do the job. Perhaps MPEG Streamclip will assist in the conversion.
@rumplestiltskin, AVCHD is both a delivery format AND a shooting format. It was designed as such. Unfortunately, it was done without much foresight, as even in this modern day, it requires a fast computer and still gives editing a problem (due to the med/long GOP length, and the complexity of h.264/AVC).
No, what needs to happen is to create a PROXY workflow setup. There are a couple of ways one can do this with PProCS4 (Update References, Folder or Filename swap, or use a 3rd party tool like reLink reTool.net). If the OP has CS4 on those boxes, he/she ought to be able to handle MPEG2 I-frame at ~50Mbps (I'm writing this on a slightly older box with CS3 and it works just fine). Could do SD rez, though I would try and see how HD works first (better composition). Personally, I would NOT use MPEGStreamclip for the job.
As I've said many times before: You start w/ Interlace, you stick with Interlace (unless it is otherwise very necessary). If you downrez to SD proxies, it will probably look crappy (inefficient interlaced-based resizing in PPro), but since they're proxies, it really doesn't matter in the end.
Assuming your students are learning to edit content, that this isn't primarily an exercise in encoding, I would just downconvert everything to DV and let the students work in that. It's robust, it's relatively compact, it edits easily, and though it's not as crisp as HD it can look pretty good especially when it comes from an HD source. (We used to do this for off lining broadcast material all the time when HDCam was in its infancy.).
When the show is finished they can either leave it as-is or relink to the HD material for a final output.
What Apple does, to disguise the shortcomings of their system, is force the users to always transcode to their own proprietary codec which is more akin to Mpeg2, but with more color depth. Then they pass it off as "Pro". Apple is very slick, and they don't want to pay any codec royalty, so they don't support AVCHD.
That's why, on face value, Mac appears superior. There's politics in all this too. Your students are better off learning how to work around these types of things if they want to be professional editors.
Editing is an ART, but it's also a SCIENCE.
Last edited by budwzr; 30th Mar 2014 at 13:05.
For the nth time, with the possible exception of certain Intel processors, I don't have/ever owned anything whose name starts with "i".
@smrpx, I agree about DV. That's why I mentioned it above.
@budwzr, not sure which Apple codec you are talking about there (ProRes, AIC or something else), but ProRes certainly IS a high-quality, pro codec, regardless of whatever bias towards or against Apple that you or I may have.
And I don't get your assertion that Apple doesn't support AVCHD - of COURSE it does. QT has had it's own AVC/h.264 support since the format came out. Heck, it DEFAULTS to that codec! (even though their implementation is not the best quality nor the fastest)
It is just the FCP/FCS/FCE/FCX and iMovie apps that have trouble with it. And even then it's not the codec, it's the container (TS) and the fact that both of those app lines have a certain assumption:
1. It is assumed that FCP-type users are either semi-pros or pros who are used to the idea of transcoding to DI (digital intermediate) formats to have improved edit response and quality in their workflows. Also understood is the preparatory logging process, and so this is a good ingest point at which to transcode if necessary (and it is usually justified). ProRes is the default DI codec of choice, but you can use AIC (not such a good idea) or a few others.
2. It is assume that iMovie-type users are very simple consumers who don't want to figure out the vagaries and extra options, exemptions & considerations involved with a plethora of codec types - they just want to load it in and edit it and load it out & play it. So iMovie transcodes behind the scene (defaulting to raw DV streams or AIC QT MOVs)
Of course, turk690 is correct, there is nothing saying the OP's student's PCs are specifically or solely Macs. So that whole business was a red herring. The OP mentioned Premiere Pro CS4, which is cross-platform.
Transcoding IS a good idea. Which is why I recommended it in the first place. It should of course be geared toward the system that one has available. Which is why I suggested trying out a few test encodes in the various tiers of bitrates & complexities to find the best fit. And, ProRes, Cineform and DNxHD as well as DV, MPEG1, MPEG2, MPEG4 and AVC can all be used on both platforms. Problem only comes with choice of cross-platform lossless codec, as Mac doesn't have nearly as many options (BitJazz/Sheer, Ensharpen, UTvideo, QT Animation, QT PNG, maybe something else through ffmpegX).
....well, it's been a number of days and the OP hasn't even responded once...hmmm. I think I'll wait before contributing further.
Depends on what you mean by "support". Like I said, it expects to either use a 3rd party plugin (to handle the additional housework/overhead) or it expects to transcode to a more appropriate & edit-friendly DI codec. It was never meant to be a "we can handle anything you throw at us" app, because that has certain assumptions also (usually: hobbyist getting multiple conflicting types from various sources, including high processor-intensive & long GOP material), and it doesn't want to follow that path, just like other pro & semi-pro apps. They already know the common Producer's workflow and accommodate only those. If it doesn't fall under that more standardized & rigorous expectation, either use 3rd party plugins or transcode to a friendly format.
Not sure what you mean by "MP4". Do you mean the container? or the Codec? And remember, sometimes "incompatibility" is just about a difference in some header metadata (e.g. fourcc of "AVC1" instead of "h264" confoundiing understanding).
I have tried FCPX and while I think it has a few interesting features, otherwise I can't stand it and won't use it without external necessity.
I just tried a trial of the $399 Magix "Pro" NLE. Hahaha, that's a friggin joke. I didn't know how good VegasPro really is compared to that.
Also, It's already been brought up but I'll bring it up again - Cineform is a fantastic, free solution for full editing of AVCHD. I use it all the time with great results.
There are a LOT of professionals who use Vegas Pro professionally. But a lot depends upon the situation & workflow setup. If you are expected to work with groups at workstations or work via lots of transfer back-n-forth between other parties, Vegas wouldn't be the #1 choice. For "one-man-band, do-it-all" projects, Vegas is very good. And it is certainly capable.
I would also say (as it is my target workflow of choice) Vegas Pro is the BEST choice for small budget, quick turnaround, start-to-finish high quality stereoscopic material.
They're all tools in the toolkit, not all swiss army knives. Best to use the right one(s) for the job.
I do agree about Cineform, though.
Vegas is an honest product, and you get your money's worth.
What I like about it the most is the object-oriented approach vs. linear. Vegas has nine ways to skin a cat right on the porch vs. long walks in the woods.
Some of these NLE's, if you make a blunder, you get punished by having to rejigger. I don't like the AE layout where all the keyframing is stacked up on a huge grid, and one errant mouse click can throw the baby out of the bathwater.
Last edited by budwzr; 1st Apr 2014 at 11:43.
@budwzr, why are you still making blunders?!
i know this person may just be teaching students how to edit video, but i figure he/she needs to return to this thread to respond to some of the questions being asked before offering all these "alternate" options.
however, if the institute that he/she teaches at needs to understand that if they are serious about what they are doing, and are offering the students cameras that record in full HD avchd format, they need to get serious about giving this teacher, and the students the best possible software and computers for these students to learn with, otherwise using avchd camera's to shoot HD video is a waste of time if they cannot utilize this format and edit the video as it really is.
avchd is easy to edit, and it can be done with some degree of patience using slightly lower end computers, but you should not have to down convert or change the format of this full HD avchd video just to teach the students how to edit, to me that is a rediculous scenario.
i shoot all my wedding videos in either 1080/50p avchd @ 28Mbps (m2ts) with my AG-AC90 camera, which i edit (very easily) and then i use smart rendering to output the file back to m2ts and i output another copy into an mp4 container for better playback support, both of which do not get re-encoded.
sometimes (at extra cost) i will i plug my ninja2 recording box into my camera using hdmi, and i record onto a 1tb laptop hdd in 1080/25p ProRes format @ 200Mbps, which i then color grade and edit, then i convert it to high bitrate h264 in both m2ts and mp4 containers for my clients to play direct from the usb powered portable hdd onto their HD tv.
if these students are not able to edit avchd video using what they currently have, then they are not getting the education that they really deseve given this is 2014 and not back in the dark ages of standard def video format.
my son is studying film & television at uni (2nd year) and they are using anything from DSLR camera's, full production television broadcast camera's, and even BlackMagic cinema cameras that shoot video in 220Mbps ProRes HQ 422 and Raw CinemaDNG, and if the university was still offering these student old and outdated computers and software (they use Mac's) to deal with the types of files these camera's shoot, then the student are being robbed.
Last edited by glenpinn; 4th Apr 2014 at 20:23.
While I agree with your argument about taking the high road and not compromising in ones professional and educational expectaions, glenpinn, it may not be under the OP's power or jurisdiction, or possibly even in his career best interest, to take a stand on this.
I've been in similar situations, where the agency/institution you work for doesn't have (and possibly never will) the budget that would be proper/fitting for your department. But that's the kind of challenge you find in the real world. And that is what I was trying to address.
Btw, I am getting a Ninja2, Gopro Hero 3+ Black, and a few other new tools to add to my Indie/Freelance dslr arsenal, in about 2 weeks, so I know where you're coming from.
editing, but more so converting full HD avchd video even to dvd compliant mpeg2 files even requires a decent pc or laptop, and that was one of the reasons why i said what i did in my previous post, i dont see the point in down scaling the video because maybe the old computers wont even cope with re-encoding to mpeg2 format anyway.
do you know or even understand about the exact format that the students are shooting in with these JVC camera's (resolution, framerate, bitrate, audio etc)
also, if the computers are crashing all the time using CS4 even when you are converting to the formats that you already mentioned, there is no way they will perform any better trying to convert to other formats to edit in, because those computers are still going to crash doing the conversions, even to standard def dvd format.
Last edited by glenpinn; 5th Apr 2014 at 02:27.
The problem in the school systems is they don't actually know any of this stuff, and rely on Apple too much. They want everything to be turnkey, and the kids don't get to learn workarounds like in the real world.
Like anything else, you get the real knowledge when you get into the industry itself.
@glenpinn, the problem with "Do it right or don't do it at all" is that, based on modern-day realities of fiscal conservativism - particularly WRT education, then a good percentage of eligible learners would be totally WITHOUT that avenue of learning. I would think that a shame. If, and I believe it to be true, budwzr is right and most knowledge is gained in the industry, or "in the doing", then even a half-assed setup might be just the leg-up necessary to get a student into a chance of an industry slot. That means something.
Also, I dispute your contention that "PCs that cannot do AVCHD probably can't even do downrezzing to SD/MPEG2/etc". Again, not "all or nothing at all", but there are incremental steps that one can take, depending upon their horsepower. Hell, my old workstation can do MPEG2 downrezzing HD like nobody's business, but would literally come to a standstill on editing (particularly multiple tracks) of longGOP AVCHD. Same with my older laptop. As we know, SD is easier to deal with than HD, which is easier than 4k. DV & MPEG1 are less computationally intensive than MPEG2 which is less than MP4-ASP, which is less than AVC, which is less than HEVC. I-frame is less complex than Short-GOP, which is less complex than Long-GOP (and ClosedGOP less than OpenGOP). There incremental changes in capabilities there that allows for "wiggle room" to find the right combination.
But think about yourself: do you ALREADY have that brand spanking new BMD 4k Production CinemaCam or that Red Epic? No, you are working your way up the ladder, making do with shortcuts, as you can afford. So these students might have to start a little early with the same kind of compromises we have to deal with, big deal. It is a dose of reality.
The advice I gave was based on years in the industry and on sound scientific principles. Plus, I seem to have been backed up by others in this thread. So where is this "totally wrong" stuff you mention? Prove to me with objective, scientific concrete examples and I'll change my tune...
Last edited by Cornucopia; 1st May 2014 at 17:36.