AVCHD - professional or what?

Question

Here goes nothing....
I am / was  considering buying a new Panasonic AG- AF100 for some of my projects.
What scares me is ... AVCHD as a format  when it comes to be edited.
I will be undertaking extensive tests (shoot and workflow) before I purchase but...my limited experience with AVCHD is not good.
NB: I have never edited any AVCHD that I shot myself yet.
As recent as today I have been editing some AVCHD footage supplied by a client from his Am Cam Canon.  It was horrendous  to work with and confirms my suspicions that PPRO does not seem to "enjoy" monitoring it. (Source / Program) 
Note: the ******* bought and uses a NTSC camera in PAL Land and I had 29.xx frame rates.  Us Pal landers aint used to that crap!  I did edit in a matched sequence ...BTW ...in case you wondered.
Unlike DVCPRO HD ..the small amount of AVCHD  I have edited has proven  to be a less than desireable editing experience especially once FX or CC are applied.  (MB Looks did not want to know about it for some reason)
MB Misfires were o.k but took an eternity to render.
I do not want a workflow that includes transcoding.  I want the silky smooth "butter" like clips  I edit and export  currently . (DVCHD Pro / avi / mov/  Pro res)
Any pro editors here  working with AVCHD  and what are your experiences in a pro workflow. (eg Broadcast, FX, CC, Compositing, After Effects etc..)
Note:. I am not talking YouTube or home movie production here.  I am seriously looking for the possible gotchas with this format and they tend to display when you get to a certain level where the quality is monitored and specced..

Answer 1

I would consider myself a professional, and I do use AVCHD to some extent (I've got a b-camera that records in AVCHD).
As several other posters have said, if you're just doing a simple cut up of the footage and editing something together, AVCHD works fine. It's when you try to start doing other things that you run into problems, and here's why:
There are basically two types of encoding algorithms: one is called inter-frame, and the other is called intra-frame. In intra-frame encoding (the type used in DVCPRO-HD, JVCPRO-HD, and XDCAM-HD, among others), each frame of material is encoded separately, meaning that as you work with the footage, each frame can be dealt with on its own. The file sizes are typically bigger, but the quality is much higher. Inter-frame encoding (the type used in AVCHD and others) uses what is called a Group of Pictures (or GOP) where, typically ever six frames, you have one frame that is encoded at its full resolution. The intermediate frames aren interpreted based on the full frame (or I-frame) which falls every six frames (basically the same thing as tweening). What this means, is that if your cut happens to fall in the middle of a six-frame GOP, all six frames have to be decoded and analyzed by your NLE before anything can be done, adding to the slowness of your system. It also means that in order for any FX to be applied, each GOP has to be broken down into individual frames, and finally re-encoded back into the GOPs. Interframe encoding is a horrible spec, slows down every NLE (even the mightiest of computers), and was really meant for the home-movie enthusiast who simply records, copies to the computer, does some basic edits (maybe) and then burns to a DVD. It wasn't really designed for use in professional cameras. However, the small file size has become so popular, that professionals are attempting to use it to save on space. Read the Wiki article on inter frame encoding for some more information and diagrams of how this all works.
There's a little bit of the tech speak on it, here are some of the real-world results I see on a day-to-day basis:
Color correction sucks. It's really really hard to make my AVCHD b-cam footage match anything else--HDV, DVCPRO-HD, RED, Canon 7D... anything. I've found that it typically has a distinctive yellow tint to it (may just be my camera, but there is no way to manually adjust the K white balance) which usually cannot be removed in post.
Fast pans or recording of any sort of action (especially if recording in any of the interlaced modes) really sucks. You can see some pretty bad interlacing artifacts in your footage, and you get what has been termed the "jello-effect" of the GOP as it "tweens" between every six frames and so you get a leaning effect of vertical objects in a pan (buildings, trees, people, etc.)
Encode times are typically three to four times longer than footage that's compressed using and Intraframe codec.
I'm guessing this is also due to the GOP thing and the frequency of the light spectrum), but under most lighting conditions (basically anything except the sun or incandescent around 3200K), I get a really bad "pulsing" effect to my image which sometimes disappears after encoding to a final format, and sometimes it doesn't.
Using the motion effect to increase the size of the image at all ("zoom-in in post") ends up giving me really crappy, blocky-looking footage.
For that form-factor, you could pick up JVC's GY-HM100U for half the price of the Panny. I'm about to pull the trigger on it's big brother, the GY-HM700U, which can be had for $2k more than the Panny. Both of those cameras use the much better Intra-frame encoding algorithms. And keep in mind, the Panny doesn't come with lenses, either, so unless you've already got lenses to fit that camera, you've got more cost that's gonna eat up your budget. If you want interchangeable lenses, the GY-HM700U can do any 1/3" bayonet lens, and can be adapted for any 1/2" or 2/3" bayonet lens.
Good luck!