Alan Roberts (BBC) tests the XF300 - 305

cowpunk52

cowpunk52

Veteran
Some quotes from Alan Roberts on a "different site"

"Chaps, I've completed tests on a 305, and it's rather lovely. It preforms dramatically better than a 1"/3 camera should do. The test was for an independent production (not European) and Canon, with the BBC taking a back seat for a change. So the settings doc does not carry the BBC's imprint. Nevertheless, the tests and reporting procedure are exactly as for any other test."

"I think this camera raises the issue of whether 1"/3 is ok for broadcast, certainly it performs as well as, if not better than at least 2 1"/2 cameras that are approved by the BBC, and doesn't need external recording to achieve that performance. And as for 13 user buttons, well ....."

"the 305 appears to out-perform the EX in all respects, if only marginally, which is entirely the wrong way round, it shouldn't be so,. Clearly Canon have been doing some very clever things behind the scenes for a while, some of which I still don't understand."

"the XF appears to be a little more sensitive than the EX, and to have slightly better noise performance, and the range of usable aperture values is about the same."

"this camera goes to F/1.6 and is good nearly to F/8, which is significantly better than it should be, the lens is much better than I'd expected."

"this camera should force a change of broadcasters' rules. I think that may already have happened in the BBC; although the tests were independent of the BBC this time, they have the report just like anyone else."
 
You can find his 15 page report in pdf format by googling around a little. It's a great read if you're seriously considering this camera.

I just ordered mine, expect it in my hands by the end of next week - completely biased review soon to follow :cheesy:
 
Interesting
I thought it wouldnt have better lowlight... Based on what ive read so far.
Would love to see some indepth review from barry ;)
 
clarification by Alan Roberts on sensitivity of the xf300/305 vs EX:

"Incidentally, I've slightly mislead on sensitivity: technically, the 305 is almost a stop slower (F/7 for 2000 lux) than the EX (F./10 for 2000 lux), I've had to do some digging about to find this. But, the 305 is about 2.5 dB less noisy than the EX1 on test. So Canon haven't changed the laws of physics, only better interpreted them. What I'd lodged in memory is that, given the 2.5dB better noise, and the larger maximum aperture of the Canon, the effective sensitivity is about the same (i.e. light level for full amplitude video from a 90% reflectance, with lenses wide open and gains set to produce the same noise level). I didn't deliberately set out to mislead, these things happen sometimes. The 305's noise level is also something you can control, the gain levels for the gain switch are settable in steps of 0.5dB, and the noise reduction is fine tweakable, the Auto setting gives 2.5dB and is quite conservative, no effect on resolution. In full manual control, at the extreme, it drops noise a lot more, and actas as a very clean prefilter for downconversion to 720p or SD. A really nice feature."
 
Bassman2003 said:
This is a real big fact. f7 @ 2000 lux is pretty slow and takes some of the lustre off this camera.
Click to expand...

Not to pick a fight, but that is just one stop from the EX. And with better noise performance on the Canon, it looks like low-light performance is a non-issue between the cams.
 
On the other side of the coin, low-light is just one issue to consider if you're trying to decide between the an XF or a EX. This is a side by side comparison of an EX1 and a pre-production XF300. Shot @ f/5.6

100% crops

1278259234.jpg
 
The Canon looks really nice in those shots. I wonder if the sun was behind a cloud for the Canon shot because all of the reflections are not nearly as bright in its shot. The EX-1 seems to have more dynamic range than that image shows as well - no highlight detail and crushed blacks.

Some certified reviews of this camera will clear everything up. I just think that if the f7 @ 2000 lux is the true number, that is slow for a modern pro video camera. The EX-1 is fine for light gathering but I would not call it a low light monster. For my use, f10 @ 2000 lux is the bare minimum I would want to have.
 
cowpunk52 said:
On the other side of the coin, low-light is just one issue to consider if you're trying to decide between the an XF or a EX. This is a side by side comparison of an EX1 and a pre-production XF300. Shot @ f/5.6

100% crops

1278259234.jpg
Click to expand...
If these images are accurate, the last picture has much better sharpness and latitude. The first one has aberrations as well.
 
Measured S/N is not very useful. The EX and almost all CMOS cameras use heavy noise reduction. They operate far beyond the capabilities of their sensor. You need a series of shots side by side and a subject with chroma detail and detailed texture in order to see how noise reduction affects image quality. And I mean real texture, not a test chart, medium and low contrast texture on human skin, fabric, gravel etc. In some gain settings the EX appears to have low noise, but dynamic range is low, there is virtually no texture and the overall look is very consumer.
 
dcloud said:
Interesting
Would love to see some indepth review from barry ;)
Click to expand...
On the contrary I find his opinions quite biased. Any opinions contradicting his statements are quickly removed from this forum as well (like my questions regarding his article about XDCAM vs AVCHD test).
 
No, Robert, that has nothing to do with the forum, anyone is free to contradict me at any time, as long as they're prepared to back it up with facts. Baseless squabbling is, of course, discouraged.

I'm just tired of your particular way of going about it. You don't show any obvious desire to advance the base of knowledge, you just want to attack. Your "questions" regarding the XDCAM vs AVCHD test started with "this test is bogus". Sorry, but that really, really isn't the way to go about opening a discussion based on facts and reason.
 
robertrogoz said:
On the contrary I find his opinions quite biased. Any opinions contradicting his statements are quickly removed from this forum as well (like my questions regarding his article about XDCAM vs AVCHD test).
Click to expand...

robert, what are the cameras that you use cause im just curious at what is it that you do.

Ive seen things barry has done and Im not his fan, but i do believe his tests are scientific and well founded.

Robertrogoz however, all ive seen is whine rant and complain while not really providing any kind of absolute information. "yeah your test is wrong you should do this and do that" "avchd is gonna look like **** on the big screen" Ive read your question that was deleted. it was definitely rude and trolling
 
IMHO, This Canon cam is at least three years late to the market. Canon should bank on their HDSLR technology instead.
 
LuckyStudio 13 said:
IMHO, This Canon cam is at least three years late to the market. Canon should bank on their HDSLR technology instead.
Click to expand...

I can totally understand how someone would feel this way - however, what I think you're not considering is that Canon's XF cameras are aimed at a different demographic from what it's video DSLR's are aimed at. They really are a completely different class of camera from each other, and can complement quite well. What Canon has done with the XF line is release a class-leading camcorder, regardless of whether you're interested in that class anymore or not.
 
Barry_Green said:
No, Robert, that has nothing to do with the forum, anyone is free to contradict me at any time, as long as they're prepared to back it up with facts.
Click to expand...
Barry, my brother, it does seem Robert is not used to read myself under the glass POV LOL ;-) Did I mention my beloved 5D Mark II...? :laugh:

Peace bros,

E. :-)


<PS:> BTW

We are here 'cause we are all friends or go to somewhere else.
We are here for sharing knowledge, trading our experience and POVs, making 'allies' aka friends, having fun and nothing else.
 
What market segment/s you think Canon is targeting with these two Cameras ?


cowpunk52 said:
I can totally understand how someone would feel this way - however, what I think you're not considering is that Canon's XF cameras are aimed at a different demographic from what it's video DSLR's are aimed at. They really are a completely different class of camera from each other, and can complement quite well. What Canon has done with the XF line is release a class-leading camcorder, regardless of whether you're interested in that class anymore or not.
Click to expand...
 
Barry_Green said:
No, Robert, that has nothing to do with the forum, anyone is free to contradict me at any time, as long as they're prepared to back it up with facts. Baseless squabbling is, of course, discouraged.

I'm just tired of your particular way of going about it. You don't show any obvious desire to advance the base of knowledge, you just want to attack. Your "questions" regarding the XDCAM vs AVCHD test started with "this test is bogus". Sorry, but that really, really isn't the way to go about opening a discussion based on facts and reason.
Click to expand...
On the contrary I asked 2 very specific questions in regard to your article, and based upon your statement:
Here I quote: "Mathematically I think it should do pretty well; the oft-stated efficiency factor is that H.264 (also known as AVC) is two to 2.5 times as efficient as MPEG-2. Meaning that AVC should be able to get equivalent quality using half to 40% of the bitrate as compared to MPEG-2".
So mathematically you should be comparing HA or HG modes of the AG-HMR10, not the PH, as XDCAM EX is the lowest of the XDCAM by Sony.
My other question, which was also unanswered is what frame grabs did you use: I, P or B? AS Mike Schell from CD writes: "I suspect you are looking at the I or P frames when you identify the slightly higher quality frames in the GOP. We saw similar results with the Video Clarity tests. The I-Frame-Only quality was very smooth, while the Long-GOP quality varied from frame to frame (I-Frame was best, followed by the P-Frame and then the B-Frame). There was not a huge difference from frame to frame, but it was noticeable."
 
And had you asked it like that, instead of starting out with "this test is bogus", you might have gotten an answer, instead of getting your post deleted.

And the answer is: I compared one long-GoP, 8-bit, 4:2:0 codec to another. It makes no sense to compare a 4:2:2 to a 4:2:0; comparing AVCCAM to XDCAM-MPEG422 is a totally different classification of products. I compared the two products that are most marketed against each other. In fact, Sony seems to think XDCAM-EX is of higher value than their own NXCAM, so it was a doubly relevant way to compare.

The lowest-cost XDCAM-MPEG422 camera is about $30,000. The highest-cost AVCCAM camera is $3,000. Why should they be compared against each other?

And why should the AVCCAM be unnecessarily crippled just for the point of comparison? I had a recorder in hand. I wanted to see how it performed. I put the best the Sony could do, against the best the AVCCAM recorder could do. In a year and a half of using AVCCAM products, I have never, ever used any mode other than PH, and I can pretty much guarantee you that if you polled the AVCCAM users here on the forum, you'll find that about 98% of them are the same way. PH is the only mode to use, unless you have specific need for extended-length recording (such as a 3-hour continuous-take concert or something). I think if you did a similar poll of EX1/EX3 users, you'd find that 98% of them have never put it in 25mbps mode. Why would anyone use anything less than the best that's available to them?

As for which particular frame is being compared -- trying to dissect whether a frame is an I, P, or B-frame out of a long-GoP codec is just silly. You're going to have to deal with all the frames, at all the times. The I-frame represents only 6.7% of the frames in any video sequence. I made no effort to distinguish between I-frames and P-frames and B-frames, I looked for examples where the action was identical between the two, and over the course of the many pictures I compared, it's entirely likely that you'll see examples of each type of frame, from each codec.

But separating them out? To what purpose? You don't get to choose to only use I-frames when you're recording a long-GoP codec. If you want only I-frames, you should use an I-frame-only codec, such as HDCAM-SR, DigiBeta, or any of the Panasonic P2 codecs.

If you read the article, I laid out the purpose of my testing quite explicitly. I was in the unique position of being able to isolate the recording format out, and use a camera head that was capable of generating images that would tax each codec, so I compared the two codecs and found that in most circumstances they performed very similarly, but with AVCCAM taking up about 40% less space. And when the going got rough, AVCCAM was substantially superior. As it should be, since h.264 technology is quite advanced over MPEG-2.
 
Last edited:
You must log in or register to reply here.
Back
Top