Has P2 been completely murdered by KxS?

Status
Not open for further replies.
adamr316 said:
http://adamreuter.com/bagpiper.jpg
Click to expand...
It terms of sharpness this grab looks very good.
But there's some ugly clipping. I guess, this was shot using a std-gamma. I highly recommend using cine-gammas instead of std-gammas (if you don't know how to use knee properly), because they always use the latitude of the sensors optimally and have very nice soft-clipping. With cine1/3/4 the very important near-to-peak-highlight-information is located above 100IRE, so it's not initially visible. Retrieving those highlights is a bit tricky.
So I highly recommend for you using cine2. Cine2 is the same as cine1, but it doesn't produce above-100IRE-information. Try that and you'll be surprised how great the highlights will look. :)
adamr316 said:
http://adamreuter.com/mattpless.jpg
Click to expand...
The same problem as in the first pic, but additionally there's a lot of noise in the darks because of gain. You have to live with that.
 
Last edited:
due to the look most people get from the panny p2 cameras. i'm sure some would shoot on a 32 gig cARD even if it only gave us 30 minute recording time. With that in mind. that should anwer your question. In order to play, you must pay. unfortunate but true.
 
mrjazz said:
I'm sorry, but you err. DVCPRO-HD doesn't support 1920x1080. The HPX3000 uses the AVC-codec for full-hd-recording.
Click to expand...

Not in the sense that you are used to with square pixels.

DVCProHD doesnt record square pixels. So 1280x1080 non square pixels results in the same image size as 1920x1080 square pixels.

The 3000 records 1920x1080 square pixels. DVCProHD lays the image down as non square as opposed to AVCIntra. So weather the image has non square pixels or square pixels the image size is the same.

The AVCIntra codec has nothing to do with recording FULL HD. As I stated 1280x1080 non square is FULL HD. The AVCIntra codecs winning smile is its ability to record 10bit to card.

But, in a sense my terminology was wrong to begin with. I should have said same image size instead of the reference to 1920x1080.
 
Noel Evans said:
Not in the sense that you are used to with square pixels.

DVCProHD doesnt record square pixels. So 1280x1080 non square pixels results in the same image size as 1920x1080 square pixels.

The 3000 records 1920x1080 square pixels. DVCProHD lays the image down as non square as opposed to AVCIntra. So weather the image has non square pixels or square pixels the image size is the same.

The AVCIntra codec has nothing to do with recording FULL HD. As I stated 1280x1080 non square is FULL HD. The AVCIntra codecs winning smile is its ability to record 10bit to card.

But, in a sense my terminology was wrong to begin with. I should have said same image size instead of the reference to 1920x1080.
Click to expand...
You agree, that 1280x1080 is less rendering-work than 1920x1080? And so it will take less time? But it will be less sharp?
 
Last edited:
mrjazz said:
You agree, that 1280x1080 is less rendering-work than 1920x1080? And so it will take less time? But it will be less sharp?
Click to expand...

Not when its acquired from the 3000's 1920x1080. Not at all.

My 500 is as sharp as the EX from about 20% down the zoom, full wide its not, but negligible. Thats a different thing all together.

And you somewhat missed my point. A 1280x1080 image from DVCProHD image is the same size as a 1920x1080 image from the EX1. Render time is based on frame size not pixels. But to be clear. Lets say the EX1 image at 1920x1080 is (just made up figures for reference) 10inches x 8 inches - DVCProHD at 1280x1080, again using non square pixels, is also 10 inches x 8 inches.

And to get back to the point of this thread - the additional render time is not because of resolution.

To get this in perspective I converted the 5 second DVCPro footage into Apples prores 4:2:2 1920x1080 (an i frame codec). Then ran the same filters as before. The results were almost exactly the same.

If you dont believe me - take a 5 second clip, convert it into apple prores 4:2:2, then take the original in a XDCAMEX timeline, the prores into a a prores timeline apply a couple of heavy filters or do some timeline composite, exactly the same for both and then check the render times for yourself.

I like the EX - on a wide shot at a function event etc or doing some wide exteriors the image is outstanding. It doesnt do what I need beyond that, but neither does the 200a, the 170 or the 150 for the most part based on chip size and lens.

And to reiterate the OP's topic - he asked is Kxs a p2 killer and again based on the points I made, I said no. This has nothing to do with the EX1 or HVX.
 
Last edited:
Noel Evans said:
Not when its acquired from the 3000's 1920x1080. Not at all.
Click to expand...
You can't acquire 1920x1080 from hpx3000 with DVCPRO-HD, because DVCPRO-HD doesn't support 1920x1080. Period.
hpx3000s 1920-signal will be downsampled to 1280 or 1440 before DVCPRO-HD-encoding.
Noel Evans said:
Render time is based on frame size not pixels.
Click to expand...
Render time is based on pixels not on something imaginary like frame-size. Please, please, please understand that! :crybaby:
Your NLEs rendering-engine doesn't know anything about imaginary aspect-ratios or frame-sizes. It just knows that there are a lot pixels it has to process, and the more pixels the more work to do.

For example if you're using a plugin which inverts your pictures, then its code would look similar to this:

for (int y = 0; y < pixelCountY; y++) {
for (int x = 0; x < pixelCountX; x++) {
newPixel[y][x] = 1.0 - oldPixel[y][x];
}
}

The more pixels the more loop-iterations the more time needed.
If my monitor would be made of rubber and I would stretch it, then the frame-size/aspect-ratio would change, but the count of pixels wouldn't and so the render-time wouldn't, too.
 
mrjazz is correct. Noel you don't grasp the concept of digital data. It's irrelevant what aspect or the shape of the pixel on screen, each pixel is a discreet unit of data to be processed. More pixels takes more time to process.
 
Well, yes and no. Yes, in that there's less source data, but no because that's not really relevant unless you're working in a 1280x1080 timeline. If you're working in a 1920x1080 timeline then it would actually take longer to work with the DVCPRO-HD data because it would have to be unstretched to 1920x1080, a process that the EX footage wouldn't.

I think Neil's original point is getting lost anyway, because adding all those filters skews the time difference between the two for editing efficiency. I mean, forget the filters and just use the raw footage; I can easily edit five streams of DVCPRO-HD on my laptop because the intraframe nature of the footage means no long-GOP frame searching or reconstitution. You couldn't do that with EX footage. That's all he was really getting at.
 
The aspect of a file is nothing but a signature of how whatever reads the file is to display it. You technically are looking at "stretched pixels."
 
Barry_Green said:
unless you're working in a 1280x1080 timeline.
Click to expand...
Noel explicitly talked about editing DVCPRO-HD on a DVCPRO-HD timelime, so I implied that.
Barry_Green said:
I can easily edit five streams of DVCPRO-HD on my laptop because the intraframe nature of the footage means no long-GOP frame searching or reconstitution. You couldn't do that with EX footage. That's all he was really getting at.
Click to expand...
But that pixel-independent advantage of DVCPRO-HD just doesn't exists. I just tested it:
I've got a Mac-Pro, and if I play back DVCPRO-HD with 1440x1080 pixels it takes about 48% of one core. XDCAM-EX with 1920x1080 takes about 53%. So the performance/pixel-ratio of XDCAM-EX is actually higher.
Skiping or playing back backwards is of course more exhausting for XDCAM-EX because of long-gop.
 
Mr. Perrone Ford , can you tell me more about this sdi out convergence box , i have a new hvx205a so it has the sd slot on the side . to answer the question for all of you why not just buy both ?
 
lawriejaffa said:
Oh what a mind numbing discussion, mr. jazz are you on commission or something? ;)
Click to expand...

Not for me, interesting discussion on GOP versus a discrete frame codec, done at the level a lit major like meself comprehends.

I'm surprised that the defacto position is to deal with EX footage in its native form instead of an IC, with HDV I always use an IC, but then again, my computers are all lame a**.
 
Noel Evans said:
I like the EX - on a wide shot at a function event etc or doing some wide exteriors the image is outstanding. It doesnt do what I need beyond that, but neither does the 200a, the 170 or the 150 for the most part based on chip size and lens.
Click to expand...

Strange,
so why is the EX, HVX200a, and HPX170 only good for your wide shots?
What is it you need it to do when not as wide?
 
Price for the recorder is $4995 + $995 more for the optional thingymagig input...

I just picked up another used hvx (no P2 card) for $2700 today...

For $6k I can freakin shoot 35mm film (and develope it) for the 54 minutes I need it.
 
So what I have gathered here so far is yes P2 has been murdered by KxS (price wise)...

Not a lot of cheap options for us...
 
Status
Not open for further replies.
Back
Top