Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Red One Measured Resolution

reality said:
Wow. All I can say is that's quite irresponsible for a review to be that lazy especially since Adam is quite well respected. What's the point of a long detailed review if it's going to be incompetent. At least take the time to learn the f23 and red.
Click to expand...

Especially, since he's in line to purchase a couple Reds. I can see this review being spread all over by Red's detractors, now.:waaa: Me thinks Adam might want to do a little revising to his review, once he gets his own camera(s) - He did it before, with the HVX.
 
Kenn Christenson said:
Me thinks Adam might want to do a little revising to his review, once he gets his own camera(s) - He did it before, with the HVX.
Click to expand...
He also did a follow-up to his EX1 review, and then a further follow-up to his HVX vs. EX1 article. He's very good about that.

He also put as many disclaimers as possible, all over it. People do tend to see pictures first and draw their own conclusions instead of reading; that's part and parcel of dealing with a short-attention-span society. But he did clarify throughout, and acknowledged that his test was totally unfair and that he didn't know how to operate the F23 and the Red. I have no doubt a follow-up will ensue.
 
Even with disclaimers, he's making conclusions that people will read and it's can easily mislead people even with disclaimers. Why even bother reading the review if chances are most of it is inaccurate? It's a shame that he's being so sloppy. He will lose his hard earned reputation.
 
Adam writes: "...we had two people, three cameras, four setups, and six hours, so please forgive the rushed nature of the tests."

A "quick look" is all you can really do with that situation in that time. I'd think of it as a starting point for doing your own tests- certainly not as the one final conclusive test (which I don't think can ever be done- too many variables).
 
considering his overall impression of the RED was positive, what is everyone worried about? The only issue I would have is this appears to be little more than filling space on the new blog. What's the point of a pointless test?
 
Graeme Nattress said:
That could explain a few things then....... I'll look into that. Thanks Miltos.

For 2k, I'd expect about 1.6k which is significantly more than 720p or 1080i.

Graeme
Click to expand...

Graeme,

I must be missing something.

78% of 4096x2304 is 3195x1797 (about 3.2k)
78% of center cut 2048x2045 is 1597x899

1597x899 is not significantly more than 1920x1080. Am I wrong?
 
I looked at the Adam’s tests and one thing is missing, the information on how much of sharpening was applied to the material from EX1 and F23 cameras.
Also the scaling from 4K to 1080p is not just “I scaled it down to...) thing.

You can screw up all good in scaling and sharpening.
I can sharpen and scale picture to the point that every pixel in the RAW will be visible as a separate line, and what good it will produce?

I see perfectly smooth artifact and aliasing free picture samples compared to over-sharpened noisy pictures trying to estimate the vertical resolution of lp/ph

In terms of evaluating the pictures in this aspect we have to push both pictures to the point where artifacts are equal and then read the lp/ph out of it.

Can we get un-sharpen RAW from EX1 and F23 or it is already sharpen before RAW is produced?
 
Terry, you're not wrong, but neither am I, and:

a) 1080i, being interlaced may start out with a pixel resolution of 1920x1080, but will measure about 70% of that vertically, about 750 or so. That's the way interlace works.

b) Although nominally 1920 horizontally, actually measuring that is not always the case, but more importantly, most 1080i has been 1440 horizontally in most 1080i recordings.

c) Even taking the recording to be 1920 horizontally, that cannot be achieved without significant aliasing from chips with 1920 pixels horizontally. The "uncorrupted" resolution will be less.

d) I should have made myself more clear what I was talking about....

Graeme
 
Graeme Nattress said:
Terry, you're not wrong, but neither am I, and:

a) 1080i, being interlaced may start out with a pixel resolution of 1920x1080, but will measure about 70% of that vertically, about 750 or so. That's the way interlace works.
Click to expand...

Respectfully disagree, Graeme. 1080i has EXACTLY same vertical resolution as 1080p and that is exactly 1080 pixels when captured with 3xCCD.

You are right on the other points, though :biggrin:

Cheers,

Pawel
www.achtel.com
 
Pawel Achtel said:
Respectfully disagree, Graeme. 1080i has EXACTLY same vertical resolution as 1080p and that is exactly 1080 pixels when captured with 3xCCD.
Click to expand...

It would be logical to think that, and I used to think so also, but it is not true :-). As far as I am aware, all interlaced video cameras produce the interlaced fields by averaging two vertical scanlines, first 1&2, 3&4, etc and then 2&3, 4&5 etc. If you do not do this two-line averaging, you get severe "line twitter" when anything detailed and in focus is shown on an interlaced display.

This also produces the 1-stop change in sensitivity you have probably noticed when switching dual-mode cameras between interlaced and progressive scan modes; in interlaced mode you bin two pixels vertically together into one, hence twice the signal (1 stop exposure change).
 
jbeale said:
It would be logical to think that, and I used to think so also, but it is not true :-). As far as I am aware, all interlaced video cameras produce the interlaced fields by averaging two vertical scanlines, first 1&2, 3&4, etc and then 2&3, 4&5 etc. If you do not do this two-line averaging, you get severe "line twitter" when anything detailed and in focus is shown on an interlaced display.
Click to expand...

Yes, but then, there are two fields to a frame making up for this....unless...., you use something like S-EVS, which scans all the lines in every field improving vertical resolution and alising problems. Then you end up with 1080i that visually looks sharper than 1080p because of higher temporal resolution. The price you pay is 1 stop of light loss.

I have done several tests of 1080p and 1080i and the later shot with S-EVS wins the resolution contest hands down. You need a LOT of lighting, though.

Cheers,

Pawel
www.achtel.com
 
Graeme Nattress said:
Terry, you're not wrong, but neither am I, and:

a) 1080i, being interlaced may start out with a pixel resolution of 1920x1080, but will measure about 70% of that vertically, about 750 or so. That's the way interlace works.
Graeme
Click to expand...


Ok I understand I missed the "i" in 1080i :)

This discussion answers a question that I have had for awhile about the resolution of 2k. In all the 2K footage I have played with I could never get a sharp image at 1080p. Now I know it is not possible, and I understand why.

This might not be the most positive thing to discuss, but this is great information to understand.
 
Vertical resolution

Vertical resolution

Ah the great broadcast debate:

Interlaced (displays) have been assumed to provide 70% of an equivalent resolution progressive (display). So if 1080p is perfect, it suggests about 750 lines for 1080i. You can certainly drive an interlaced camera - especially when looking at a non-moving TV test chart - up near max resolution, but if you do so, it will then alias.

Hence the minority preference for 720p from Fox, ABC etc, as its considered to have a similar effective vertical resolution, and slow motion that doesn't twitter field to field.
 
2K multiburst chart test

2K multiburst chart test

Sorry I missed this thread. I had the opportunity to test a prototype 2K multiburst chart from DSC with Red X98 last month. Here's the specs:

February 28, 2008
Red ONE 4K format
Cooke S4 50mm@T4 1/3
Focus 4'8"
fps 23.98 1/48 shutter ISO 320
Light Source Tungsten
White Balance set to 3180K

Red ONE 2K format
Cooke S4 50mm@T4 1/3
Focus 9'
fps 23.98 1/48 shutter ISO 320
Light Source Tungsten
White Balance set to 3180K

4K looks to be somewhere between 1500 to 1750 lpph and 2K looks to be somewhere between 750 to 1000 lpph.
 
David, that's an interesting question (and as a film brain guy, I'm always having to re-adjust to this) so this is less of an answer than a further query: the film enlargement ratio was/is more pronounced/effective because all things were equal (more or less)- the film emulsion/grain and processing of 35 mm equaled that of 8x10, so the larger area obviously produced greater results at any given output size. I say "all things being equal, more or less" because 8x10 large format optics generally tended to be inferior to 35mm's- I think this was due to the lens manufacturer's conscious decision- just try cutting out a 35mm portion of an 8x10 neg and compare it with a 35mm frame, ouch. In this regard, the digital "frame" is the pixel and not, perhaps, the frame size- i.e 35mm vs 65mm and so on. If one were to create two digital frames with the same pixel size then the enlargement (quality) factor would obviously apply (all things being equal). However, if, as you suggest, one was to create larger "photosites" on a 65mm frame and compare this with smaller sites squeezed into a 35mm frame, then we should expect little gain since the enlargement ratio benefits in the 65mm frame is negated by the fact that the sites have already been (effectively) enlarged. A film analogy might be to compare a very fine grain 35mm film stock with a very course grain stock of an 8x10 format; would the 8x10 in this case be more effective at resolving detail? I think this is why we see old optics on larger film formats outperforming new optics on smaller formats; the new film stocks haven't outperformed the older optics.
Or am I out to lunch on this?????
 
Yes, I got 3K+ on 16mm Master Prime across 4K since it is less sensitive to the focus settings, but then I discovered that I was 2 inches to close to the chart and I didn’t have a chance to repeat the test since then.
Very easy to make this mistake.

I forgot that if the chart scale is one to ten 1:10 we do not measure the distance from the chart to the camera sensor to be 10 times further then focal length. We have to set the distance 11 times further because we should actually count from the focal node of the lenses.

I wonder what distance he was using.
Well, maybe he just knows better how to setup the test. I am still learning.

Just found his chart picture!

805_1210806245.jpg


I love this smooth creamy look of surfaces, this is Graeme look I call it here.
wavelt - actually realy velvet (compression)
Don’t see it on the film, don’t see it on the blu-ray it is so characteristic to RED that if someone will show me a high detail picture I will recognize it from the one mile.
 
You must log in or register to reply here.
Back
Top