Well, the problem does exist, and I would say, why not shoot a test at a higher quality capture to see if it's improved? The answer is yes, it was improved, but not that much. There are learning curves to everything, and this is part of the process for me, learning about the camera and it's settings and such.
AF with Atomos Samurai test screen grabs
- Thread starter pulpwoody
- Start date
Well, the problem does exist, and I would say, why not shoot a test at a higher quality capture to see if it's improved? The answer is yes, it was improved, but not that much. There are learning curves to everything, and this is part of the process for me, learning about the camera and it's settings and such.
John Andersson
Well-known member
I had a Samurai and used it on my AF101. When I compared on a per pixel level what the difference was it became clear that an external recorder couldn't improve the footage in a noticeable way. Workflow became better but that was about it. IMHO AVCHD on the AF100 is as good as it gets.
Preston Davis
Member
My tests with the Samurai + AF100 and prores codec were similar. I did however see noticeable improvements in noise smoothing and highlight detail when using the dnxhd codec compared to the same footage in avchd.
Shooter
Veteran
Correct...The problem definitely exists if one stresses the sensor with over exposure. Capturing the "problem" with a higher quality codec only shows the "problem" at better quality. Its not a fix.
Learning any camera and settings via tests is a commendable and valuable activity to undertake.
My issue with a banding test such as yours is it sets off those that believe the camera has issues. eg endless threads about Highlights / Banding.
It is well proven that those that have learnt to operate the camera properly...do not have these issue. Therefore the "issues" are O.E (and inexperience)
Went through the other gamma settings, and I found SD norm has the least amount of banding, and noise was pretty even across all of them. Not too bad on the noise, nothing the client will ever notice.
Couldn't agree more Shooter. I'm just now starting to get into the settings, and I'm glad I am. I'm also learning more about the camera and how it performs, which is quite well for the price and the features it has.
Shooter
Veteran
Can I recommend that you test the camera in Default Settings (1-6) at some time early on so as you have a baseline knowledge before you apply Scene recipes.
Starting with presets designed by others... in conditions that you know nothing of... is a way of getting confused real quick!
Testing against the defaults shows one exactly what is happening when a setting is changed.
Starting with presets designed by others... in conditions that you know nothing of... is a way of getting confused real quick!
Testing against the defaults shows one exactly what is happening when a setting is changed.
I don't know about anybody else, as mileages may vary. But I definitely see a difference between the internal AVCHD and DNxHD codec recorded with the Hyperdeck Shuttle 2. I don't even need to pixel peep or zoom 400% to see it. First, the colors from the DNxHD footage are visibly richer. Second, noise is much more distributed and smaller instead of macroblocking. Third, it keys better and also grades better. Fourth it's much easier to work with DNxHD in post than AVCHD. These reasons alone are enough for me to use an external recorder.
Preston Davis
Member
Thanks! Finally someone gets the same positive results with the DNxHD codec, though mine is with the Samurai. I had someone on a different post explain why technically those improvements were highly unlikely.
I just did some color correcting with Colorista 2 and MOJO and those improvements were magnified.
I am going to continue to trust my eyes and leave the technical assessments to others.
I just did some color correcting with Colorista 2 and MOJO and those improvements were magnified.
I am going to continue to trust my eyes and leave the technical assessments to others.
Shooter
Veteran
My experience with the test I did AVCHD (cam ) vs ProRes HQ (samurai ).
I placed one clip on top of the other in 2 Premiere CS6 video layers so I was able to switch between the 2 easily for instant and direct comparison of smae frame.
I was barely able to perceive a visual difference in any of the 3 monitors. Maybe a limitation of the monitors but the primary QC Monitor was a Panasonic HD broadcast monitor and the NLE monitors were fairly high end and calibrated to Bars and Pluge
What was noticeable was a difference in the scopes. ie the data.
That data makes a considerable difference in the post production - CC/ Grade, Compositing etc.
We have tested both codecs but it is difficult to "quantify" in any practical terms apart from its "easier", more "elastic" and "smoother" to work with.
The AVCHD is definitely not a slug to work with..it should be pointed out as well...unless there is ann issue such as under/over exposure.
I placed one clip on top of the other in 2 Premiere CS6 video layers so I was able to switch between the 2 easily for instant and direct comparison of smae frame.
I was barely able to perceive a visual difference in any of the 3 monitors. Maybe a limitation of the monitors but the primary QC Monitor was a Panasonic HD broadcast monitor and the NLE monitors were fairly high end and calibrated to Bars and Pluge
What was noticeable was a difference in the scopes. ie the data.
That data makes a considerable difference in the post production - CC/ Grade, Compositing etc.
We have tested both codecs but it is difficult to "quantify" in any practical terms apart from its "easier", more "elastic" and "smoother" to work with.
The AVCHD is definitely not a slug to work with..it should be pointed out as well...unless there is ann issue such as under/over exposure.
Barry_Green
Moderator
To back up what shooter's saying... there are reasons I get a little antsy whenever someone says "I'll just trust my eyes..."
http://web.mit.edu/persci/people/adelson/checkershadow_illusion.html
Barry_Green
Moderator
(for those who haven't seen this before; squares "A" and "B" are EXACTLY the same shade of grey... don't believe me? Sample it in PhotoShop and see).
I don't know. But I really never seen a case of 4:1:1 or 4:2:0 looking as good as 4:2:2 color-wise. To me, the color difference is always there, just by looking at it. When I bring it into color grade then things jump even more on your face. Besides, it's not a case of trusting your eyes only. It's the eyes, the numbers, the data.
Color resolution is by default color "information". The point is the color is just better.
Shooter
Veteran
Image Data is very interesting and the time I notice how much the eye can be tricked (by monitors) is when I legalise a shot for broadcast.
Our cameras generally produce an image that is non legal broadcast. Chroma and Luminance.
When soft clipping images to bring Chroma Levels and Luminance into line..it is amazing how much one can clip them and there is no apparent difference visually (in the monitors).
TheWFM and Vectorscopes are showing a massive shift but the images do not change at all.
Our cameras generally produce an image that is non legal broadcast. Chroma and Luminance.
When soft clipping images to bring Chroma Levels and Luminance into line..it is amazing how much one can clip them and there is no apparent difference visually (in the monitors).
TheWFM and Vectorscopes are showing a massive shift but the images do not change at all.
That's interesting that you say DNxHD is useless on PC because Media Composer runs on PC and DNxHD is tailor made for it. So any of the Avid family editors will work with it. If you want a specialized grading software, best in the business-industry standard DaVinci Resolve works fine with DNxHD.
No, you're talking in a way that suggests the colours just pop when captured in 422 over 420 as if there is more colour information being captured or that the scene file we see on screen is somehow skewed to looking different when capture via SDI and looks different to the AVCHD and that's not the case. 422 is double the colour resolution of 420, thats it, hence why it's easier to key but we are gaining next to no extra colour information inside the 8 bit output. Lower compression via prores or dnxhd is reducing/removing compression artefacts and giving us I frame encoding but that's it, it doesn't suddenly bloom with colour.
Coming from a user of the AF100 with external recording via the hyperdeck.
By the way we're still waiting to see some of your work Adam, warts an all
Coming from a user of the AF100 with external recording via the hyperdeck.
By the way we're still waiting to see some of your work Adam, warts an all
Last edited:
