canon xl2 8 bit or 12 bit

[vamshi]

Hi guys,good morning.

Well ,After revealing the image quality, canon xl1s is 12 bit it is rich and colourful. But canon xl2 , initially they stated it as 8 bit ,later reffered as 12 bit., with only few known people in the forums. After seeing some footages posted in internet, i still has a doubt weather it is 8 bit or 12 bit?
My questions is,Can anybody clarify how can a customer can test weather 1t is 8 bit or 12 bit.

any suggestions greatly appreciated.

vamshi

[Policar]

All DV cameras are 8 bit.

[Policar]

Although the xl2 has a 12 bit dsp (the xl1s I always thought was 9 bit but I could be wrong...). You can't really test to confirm it, but it just means that colors will be rendered more truthfully.

[PFP_VIDEO]

Maybe Canons lying. They changed there story and said its 12 bit. This has me a little suspicious, especially after seeing the more muted colors. Perhaps somebody can figure out a way to see if its 12 bit. Any xl2 owners willing to crack open there camera and peek inside ?

[Barry_Green]

I'm almost certain they wouldn't be lying. I'm sure the "changing-their-story" thing is simply due to the sheer massive size of an organization like Canon. In software development we'd be asked to proofread the manuals to the games before shipping, and you'd be *amazed* at some of the stuff that the manual writers had put in there. I have no idea where they got their information, whether it was from the producers or the designers or the testers, but us down in programming were responsible for making it do what it does, and there were many times when the manual was completely wrong about what the functions really were.

(of course, usually they'd say "well, fix the program then"... sigh...)

Anyway, I'm sure that the Canon is using the latest up-to-date technology, and any errors or omissions are just due to the logistics of trying to get such a massive development team all communicating together.

[nullphonic]

Good point Barry. I’m in software dev also and on several occasions doc writers would be asked in on design meetings for notes, suddenly some of the info would be mysteriously included in release docs and then marketing/sales would relay some info to clients. We (the geeks) have been horrified during several demos. And it works both ways (as is the case here possibly) where sometimes the actual capabilities are undercut badly or overblown drastically (which means release scramble or 85 hour work weeks to catch up to the marketing material :-/)

[vamshi]

hi everybody.

I am back....from this week hurricane distruction in jamaica.

Well I appreciate for the answers. Alright i accept that ,you guys mean to say its a printing mistake, or an immature task by an doc. writer. Well ill ask you one question. the question is we knew this thing before the manual is been written. And its an hot topic , and it was an initial argument raised on canon xl2. What do you mean to say about this.? Some guys who attended the expo. in the usa when asked the canon personnel l , they come up with saying " canon xl2 is 8 bit" How come this ? They are well trained. got it? Answer please.

vamshi

[Shaw]

Well the Canon *is* 8bit (just like every other MiniDV cam). It does have 12bit DSP which is quite different.

The canon seems to reproduce colors very, very acurately IMO. Perhaps slightly desaturated (perhaps, haven't tested any of the footage on a high quality monitor) but the _color rendition_ is very good.

[Barry_Green]

There are two "bit"-related subjects here, leading to the potential for confusion. As Shaw points out, the Canon (and all other DV cameras) record DV at 8-bit quantization. ALL DV is 8-bit. Use a $25,000 SDX900, record in DV mode, it'll record 8-bit.

The other thing, as Shaw pointed out, is the DSP (Digital Signal Processor), the computer that does all the digital manipulation of the video signal (like gain, gamma, matrix, etc). The DSP on the XL2 was originally reported to be 8-bit, but the documentation was later corrected to say 12-bit.

I would strongly suspect that it is actually 12-bit. It would be amazingly silly of Canon to release a $5,000 camera with an 8-bit DSP in 2004. I think some JVC cameras are using 14-bit DSP's now, so I'm sure (with no hard evidence) that Canon is using 12-bit just as they say they are.

But really, what difference does it make? All that matters is that you like the way the footage looks. If the Canon was able to make gorgeous filmlike footage with an 8-bit DSP, don't think for a second that that would stop me from buying it. And if it makes lousy flat washed-out video even though it has a 12-bit DSP, I don't think anyone will say "well, sure it looks bad, but at least I can "geek out" and brag that I have a bigger DSP than someone else does."

[PFP_VIDEO]

Does the extra 4 bits do anything extra when It comes to video manipulation and proccessing, although only 8 bits are recorded? Is it primarily who makes the DSP and the differences in the DSP'S different manufactures use?

[Barry_Green]

The extra bits are used as basically fractional values, to make the calculations more accurate. In an 8-bit system you have basically integer values from 0 to 255, so your precision is 0, 1, 2, etc. In 10-bit, you have the same 0 to 255, but now you have four levels between each number (i.e., it's pretty much 0, 0.25, 0.50, 0.75, 1, 1.25, 1.5, etc). In 12-bit you have sixteen levels of precision between each number. With this extra accuracy you can get cleaner effects, smoother gradients, etc.