Vitaliy Kiselev
Carbonite Member
I suggest you to look at video produced by new FZ compacts, as if they don't have proper codec (B frames, CABAC, improved motion estimation) it is low probability that GH2 will have it.
GH1 discontinued; GH2 in September/October
- Thread starter Emanuel
- Start date
- Status
maranfilms
Veteran
I wouldn't be suprised if Panisonic offered tester13 a job. Seeing is he does more with the software than their own programers. lol
Vitaliy Kiselev
Carbonite Member
I can not agree here.
Panasonic programmers are very good.
You just must understand two things:
1) They use codebase that is many years old. So, they reinvent things many times.
2) They are guided by marketing department. If this guys said that 17Mbit is enought, no one have right to say that it is not.
JoeJITSU
Veteran
Pietro Impagliazzo
Well-known member
Hmmm so we just need you employed in the marketing department.
:thumbsup:
Bloody retrograde companies...
I can provide additional technical perspective on this debate, having some industry experience in coding and optimizing commercial MPEG-2 and AC-3 codecs. While it's correct that the GH1's AVCHD codec is limited to encoding I and P frames, that does not necessarily make it "crippled".
Codecs designed for professional applications, such as encoding a feature film on a Blu-Ray disc, have huge advantages over embedded codecs designed for portable devices. The primary advantage is that a pro encoder isn't constrained by real time - it can make multiple analysis passes over the raw input stream and determine its frame-by-frame encoding strategy in advance. Some codecs go so far as to create multiple speculative encodings of a frame, and simply keep the one that turns out best.
The reason the GH1's AVCHD codec does not produce B frames is because to do so, the encoder would need to make use of the next key frame in the future, as well as the just passed key frame. With the 15-frame GOP used in FHD mode, the GH1 would need to continuously store over 2/3rds of a second's worth of raw, unencoded video data in an internal streaming buffer until both key frames became available. To do this efficiently in real-time would take far more hardware resources than Panasonic could afford to build into the GH1.
While this claim is also correct, it glosses over the amount of sheer horsepower Panasonic has made available to each brand of codec. The primary advantage that AVC encoders have over MPEG-2 encoders is the ability to dynamically scale down the macroblocks used to dice up the contents of each frame. What gives the AVCCAM codec an edge is that it is equipped with enough processing bandwidth and memory resources to do a thorough analysis of each frame in advance of encoding its contents. This (along with CABAC efficiency) enables it to minimize the mud using fewer megabits in the limited amount of time available to encode each frame.
Given enough bitrate, as we now have in the GH13, an AVCHD encoder can potentially produce comparable image quality as an AVCCAM encoder, with the sole drawback of requiring more encoded megabits to do it. But if the AVCHD encoder has to operate with constrained resources, it will be forced to shortcut its macroblock analysis in order to complete each frame in real time, resulting in reduced image quality.
In addition, we've discovered that the GH1's AVCHD encoder was not engineered robustly enough to gracefully shortcut its encoding processes in a reliably error-free manner when confronted with excessively detailed images. It's a consumer encoder, it was evidently not subjected to the rigorous testing required of a professional codec. From Panasonic's perspective, it would not have been cost-effective to test its reliability beyond the requirements of its 17Mbps and 24Mbps AVCHD specifications.
Vitaliy Kiselev
Carbonite Member
Next thing to remember is that in high detailed scene with movement even B frames can be ineffective. If motion prediction is unable to find matches, almost all P and B frames become similar to I frame, and mud as result. Still B frame is better (as you can find more matches).
Plus you do not need to store 2/3s of raw data. As you need to store only ref frames + 1 raw images. In case of simple I-P-B encoder ref frames = 2 only. So, encoder uses, for example one I frame and next P frame to encode B frame.
Plus you do not need to store 2/3s of raw data. As you need to store only ref frames + 1 raw images. In case of simple I-P-B encoder ref frames = 2 only. So, encoder uses, for example one I frame and next P frame to encode B frame.
Barry_Green
Moderator
Maybe so, but rumors don't know anything, so it's counterproductive to spread them.
As others have already pointed out, just counting mbps is pointless. It's HOW the mbps are used. And it is very likely that 24mbps of "properly implemented" AVCHD would blow away the hacked 30-50mbps of inefficient codec.
And it's statements like this, with no basis in fact, that show exactly why we frown on passing along unsubstantiated rumors. You know nothing about any proposed GH2 at all yet, because Panasonic (the ONLY people who know anything about it) haven't said anything yet, but you're already ready to consign it to the garbage dump.
Talking about rumors can be fun, as long as people keep in mind that the only people spreading rumors are ones who don't actually have any idea what they're talking about. Because if they DID know what they were talking about, they'd be under a nondisclosure agreement and unable to talk about it!
Barry_Green
Moderator
Exactly my point -- there is nothing to be known, so why talk about a complete unknown.
See, this is what's ridiculous. How can you proclaim a nonexistent product a "joke" or a "crippled stills camera"? It doesn't exist yet. Nobody knows anything about it yet. This is -- well, it's basically slander of an unannounced product's reputation. How does that help anything?
Barry_Green
Moderator
Um, DVCPRO *was* great. Stood the news market on its head. Prior to the introduction of DVCPRO, the news market was about 100% Sony/Betacam. After DVCPRO was introduced, news went to 70%+ DVCPRO. They didn't do it because of illusions, they did it because, as a product, it was pretty darn good.
Correct. AVCCAM is a brand, to differentiate their AVCHD products away from consumer products. AVCCAM means a few things though -- it means a 3-year warranty, it means it's optimized by the broadcast division, it means it has P2-style metadata, it means that it has PH mode which provides for multi-slice technology and also has higher quality audio than any of the other AVCHD modes (that I know of). AVCCAM PH mode is 384kbps AC-3, whereas the other modes are 256kbps or less (GH1 is 128kbps, I think).
So it's not an entirely different format from AVCHD, no. But it is "professionalized."
For some people. I think there are a lot of other things that matter too, such as how usable it is to get that look.
Barry_Green
Moderator
Can you name any pro camera of any type that records to any recording format, and is "close to Red"? Probably not.
But I can name you some vastly more expensive pro cameras that have a worse codec than the 24mbps AVCHD. Those would include the $20,000 Sony PMW-350, the $8300 Sony EX3, the $6,000 Sony EX1, and the entire XDCAM-HD lineup, as well as the $7,000 JVC HM700.
None of those cameras have a codec that's as good as what you can get on a $1800 HMC40, which uses 21mbps of AVCHD.
So don't just talk about megabits unless you know what KIND of megabits you're talking about. They're not created equally.
17, but it has lousy megabits. The native GH1 codec is one of the worst ever released, imo. That doesn't mean it can't produce fantastic work (look at our "Shady, Texas" and "Depth of Phil" or Fat Monster's "Dispatch" to see examples of just how amazing the stock GH1 can look). But there's vast room for improvement, as Vitaliy has shown.
Groundless speculation. You don't know this at all, so don't say it until we know.
How do you know this? How can you declare this?
Homunculus
Carbonite Member
Barry you have many good points so I'm curious to hear your feedback regarding this:
let's say hypothetically GH1 (original) is a 3 on a visual quality/raw power/codec etc scale. Let's say the AF100 is designed to be a 6 so that it is decently ahead of the GH1 enough to warrant a person to pay the extra thousands for the cost of the AF100. So let's say to fill in that gap Panasonic had planned the GH2 to be a 4 or 5. That's all well and good. It fits perfectly in that filler niche role. You have gh1, then if you want to spend a little more for better quality you have gh2, then you want to spend a little more for even more quality you have AF100.
BUT, in comes Vitaly throwing the monkey wrench into the system. He took the score of 3 GH1 and turned it into a 4 or 5, which is only 1 away from the AF100 and which is exactly the niche spot in the middle that Panny had long planned for the GH2. Obviously GH2 can't be less powerful than GH1 according to panny, but obviously it can't be MORE powerful than AF100. It has to be in the middle to make sense. Yet GH13 is in the middle. So THIS is my reasoning for why I find it hard to believe that GH2 will be more powerful than GH13, or at least not significantly more.
let's say hypothetically GH1 (original) is a 3 on a visual quality/raw power/codec etc scale. Let's say the AF100 is designed to be a 6 so that it is decently ahead of the GH1 enough to warrant a person to pay the extra thousands for the cost of the AF100. So let's say to fill in that gap Panasonic had planned the GH2 to be a 4 or 5. That's all well and good. It fits perfectly in that filler niche role. You have gh1, then if you want to spend a little more for better quality you have gh2, then you want to spend a little more for even more quality you have AF100.
BUT, in comes Vitaly throwing the monkey wrench into the system. He took the score of 3 GH1 and turned it into a 4 or 5, which is only 1 away from the AF100 and which is exactly the niche spot in the middle that Panny had long planned for the GH2. Obviously GH2 can't be less powerful than GH1 according to panny, but obviously it can't be MORE powerful than AF100. It has to be in the middle to make sense. Yet GH13 is in the middle. So THIS is my reasoning for why I find it hard to believe that GH2 will be more powerful than GH13, or at least not significantly more.
Jason Ramsey
Admin
well, you also have to keep in mind all the other things an AF100 type camera will have that dslr's do not... that being a video form factor, with proper video functionality and features, inputs, outputs, etc...
So, the only difference won't just be in the codec, better image quality, but in the form factor, etc as well... af100 will have prosumer video cam prices with prosumer video cam form factor and feature set, etc... the dslr's are still... at the end of the day... dslr's... and they lack a lot of the functionality that was par for the course before the dslr revolution.... People are obviously waiting for somethig with at least dslr image quality and sensor size, but with proper video form factor, etc, etc.... at the price of a dslr... we aren't quite there yet...
So, the only difference won't just be in the codec, better image quality, but in the form factor, etc as well... af100 will have prosumer video cam prices with prosumer video cam form factor and feature set, etc... the dslr's are still... at the end of the day... dslr's... and they lack a lot of the functionality that was par for the course before the dslr revolution.... People are obviously waiting for somethig with at least dslr image quality and sensor size, but with proper video form factor, etc, etc.... at the price of a dslr... we aren't quite there yet...
Barry_Green
Moderator
Your premise is fundamentally flawed. The AF100 is made by an entirely different company than the GH1 or (rumored) GH2. Panasonic Broadcast and Television Systems does not care one bit about what impact their AF100 will have on any GH1 or GH2 or whatever. The GH1/GH2 are made by Panasonic Consumer Electronics. And Panasonic Consumer doesn't know or care what impact their GH1, GH2, whatever - will have on any Panasonic Broadcast product. They are different companies. Same overall corporate umbrella, yes, but the two companies are basically as much competitors as Nikon is against Canon. Or, maybe more appropriately, as Chevrolet's LUV truck was a competitor against the Hummer H2. Different brands entirely, different customer bases, different dealers, and products sold to different customers -- even if they did share a corporate parent.
Secondly, the GH1(3) will never compare against an AF100 for dozens of reasons, not the least of which is: the AF100 won't be subject to aliasing on wide shots (or, if it is, it will be at a greatly reduced rate, as they've already said they're properly engineering the system to deliver proper 1080p video).
So what would a GH2 be? Who knows, as such a product hasn't been announced, but a reasonable person might assume that if there were to be such a GH2, it would likely be somewhat similar to a GH1, with commensurate improvements over the GH1 in the same way and same degree as the G2 has over the G1. Which means, it's still going to be a stills camera with a video feature. And that means it'll have nearly nothing in common with the AF100.
The AF100 isn't a stills camera with a video feature. It's a video camera that's been designed by the broadcast division of one of the largest video camera manufacturers in the world, to serve as a professional video and digital cinema camcorder. That is something the GH1 or GH2 could never be. Even if, in a side-by-side comparison, the video looks similar.
Barry_Green
Moderator
The product cycles in major Japanese corporations are not affected by whether one hacker made a firmware change or not. These products have certain lifespans. The GH1 has been out for about 18 months now. That's probably its lifespan, and if there's a GH2, it's got to have been in the works for a lot longer than the firmware hack has been out!so-
you can say that one man (Mr Kiselev) stopped by himself (+testers help) the manufacturing of the GH1..
roei z
Veteran
- Status