I am guessing this post will be very controversial. I'm sure it will unleash a few opinions.

In this test I am going to be switching from A to B, and your job at the end of the video is to decide if you can tell which one A or B was sharpened in camera or done in post processing.

For this test I will be using the 720p resolution, I will explain why later. Each comparison will use the same settings except one of them will have the sharpness in camera turned to 3 or the sharpness turned down to 0 and sharpening added in post to match. For all these tests I shot this in the neutral picture style with the contrast at 1, I have not added any post processing except for sharpening for the appropriate clips.

Frame Captures at 400% Zoom

"A" In camera sharpening at +3 400% zoom
"A" In camera sharpening at +3 400% zoom

 

"B" In camera at 0 with post sharpening added 400% zoom
"B" In camera at 0 with post sharpening added 400% zoom

I watch all the top people in the industry from Philip Bloom to Vincent LaForet and most seem to say that in camera processing is the devil or evil or something like that and they advise to turn it all the way down. Since I like to test everything I can, I wanted to put this to the test as well.

From what I gather from these top people is that the small chip inside the camera that applies the sharpening is not as good as what your computer can do in post.

Now the reason I am shooting at 720p is because of aliasing or moire patterns. Since the camera sensor can shoot at a much higher resolution for stills than for movies it has to throw out horizontal lines for 1080p and even more for 720p. So from my previous tests I have noticed that 720p will show aliasing more than 1080 so that is why I choose to shoot in 720p in this test to show the difference of aliasing in the worse case scenario. But this test is not about aliasing it is about which one looks better.

Many people believe that one of the tools you have to combat the ugly aliasing is to lower the sharpness.

So you want to know which one is A and which one is B – watch the video? Even if you got it wrong I think it would be cool to see how many got it right so in the comments below let me know if you got it right or wrong.

Creative commons music by LoveShadow.

 

32 comments

  • Nice,

    First off, sharp doesn’t = bad (necessarily).

    For the first half I couldn’t tell, but for the last several shots B looked much better.
    But what most people don’t think about is trying to un-sharpen footage. It usually gets mushy and looses quality. Thats why I recommend sharpening in post. Better to add than need to take away.

  • Great test I personally don’t use sharpening at all and barley ever add it in post sometime ill use the unsharp mask but very rarely.

  • I though it was pretty apparent right off, but that’s probably because i had an idea of what to be looking for.

    The problem with sharpening in post is that the artifacts from compression end up getting amplified as well as any other details. Sharping in camera is being done before any compression happens. so it’s only sharpening actual details that the camera is seeing then it crunches the image down.

    But the problem with sharpening in camera is the aliasing and moire, as you stated in the video…

    I think it’s one of those settings that you should be aware of how it will affect the final image and tweak as needed.

    Ok… there’s a lot of action, colors, and movement. this might not compress well. And there are no potential dangers of aliasing and moire… Maybe shoot it with a little extra in camera sharpening.

    Ok… in this shot it’s pretty simple movements and a small color palette, And that brick wall behind the guy in the pin stripes looks like it may be a problem… maybe shoot witht he sharpness cranked down and ramp it back up in post.

    Personally i would prefer to capture as much of the final look in camera but leave room to tweak in post, rather than create a bunch of work correcting things later.

    Sharpening in post reminds me of spending all kinds of time getting an audio mix to sound perfect then mixing it down to a 32kbs MP3 to send to get mastered. Then asking the engineer if he can pump up the highs and lows with out getting that “magic chimes” sound on the cymbals.

  • I felt B looked much better right off the bat as well. Some of the high contrast lines are much smoother and less aliased. I also agree with Caleb, you can always add sharpening later, however taking it away is much more difficult to pull off without stepping all over your footage.

    Thanks for the great test Dave!

  • Great work Dave, thanks for looking into this. I preferred the B shots, so I’ll stay with that approach for now. Keep up the fantastic work you do.

  • More test = good! I dont even have the time for doing test so very grateful that you guys do them. And my verdict on this one is.. to close to bother about it. I always shoot with sharpening turned down since I dont even like too sharp pictures in the first place but, I guess I could add it if I need it. If time is an important factor, in camera sharpening looks great.

  • For me, I was quite hard to tell the difference between all of the clips except the last one. That was because the girl’s skin was better in B. And so I thought that since you can control the sharpening in post, there will be less artifacts (or so I think).

  • @O’Ryan I like your audio analogy, well stated.

    My opinion is that aliasing is the devil and not in camera sharpening. But again I don’t have the eye that Vincent and Philip have.

  • hummm “A” looked better, Ii would look at one point say a twig or a rock and it seemed to me each time from first 3 shots I could tell a every so slight change. I didn’t really see enough change to worry about, at least at my level of skill. Either way Dave, thanks so much it is so good to learn with someone and not have a learned expert tell me to do this without me having a clue to what he is talking about. Keep questioning the experts.

  • Interesting, I liked the contrast of A better, it seemed to pop more while B seemed to dull the image a bit.

  • I was able to see which one was which and preferred ‘b’. Enlarged version of ‘A’ seemed better for some reason.

    Maybe you could include one without any sharpening (in camera or post) as a third sample if you are to test this further.

    The difference between the two methods in my opinion will show, when the post processing is heavy.

    In any case, I would place the sharpening effect last in the effect chain. My personal preference is not to use any sharpening at all. – where did you place sharpening in the chain?

    ps. with the ‘unsharp mask’ one is able to eliminate the sharpening of the compression artefacts by adjusting the thereshold value

    pps. As a side note I was wondering about the compression project you were talking about some time back. How is it going? And have you looked into x.264 codec and MeGUI the ‘doom9’ guys made as a freeware.

  • @puujalka I have not started on my compression project I am currently working on revamping my ‘Getting Starting Guide’.

  • I have found that adding some sharpness in camera is better than using the unsharp mask in post. Seems since the last time we talked I’ve been setting in camera sharpness to the second notch from the left. Seems to be the sweet spot and definitely looks better than adding in post, IMO.

  • I agree with Caleb… sharpening sometimes looks just plane BAD… there is a time and place to sharpen… the bokeh is what we should use to soften from our lenses and thats what DSLR’s are very capable of doing… using a good quality lens should be enough to give us the sharpeness that we require… its when we oversharpen that it looks really BAD!!!

  • Thanks Dave for publishing your test. I’m new at this too but here’s one thought: I think part of the reason for using picture styles with no sharpening is that sharpening in camera significantly limits your ability to gracefully and effectively modify color and contrast in post. So it isn’t that in-camera or post sharpening is better than the other, but that if you apply the sharpening before color grading and tweaking your contrast, brightness, etc., you simply won’t have as much latitude to push or tweak.

    That may be part of the reason your A/B tests here look so similar–to me I couldn’t tell the difference without the close-up still image. I’m assuming you didn’t have to tweak the color very much in the A clips? If you did, I think that’s where we’d really start to see a difference.

  • It appears we are the same sharpness quest. I started to put some stills and video tests together for You Tube. I know that for digital stills that sharpening makes a visible difference. The real question is what settings yield the best results?

    The goal that is worth chasing is to get a camera set up that is consistent. My Canon T3i is producing exceptional pictures, but it’s not quite dialed in yet.

  • Hey Dave

    Great job. I personally chose B myself, although A seemed to be too perfect almost hyper real as my son would put it. I guess depending upon what you are going for it wouldn’t be a bad thing!

  • I’ve done a few tests myself, and honestly think that camera sharpening give me more natural better result. I use +2.

  • @Richard Allen Crook @Rob I have been using +1 but I might be going to +2 now that I have done this test.

  • Great exercise! I personally preferred A in most of the shots. I am in agreement with Richard and Rob and think sharpness at +2 on my T3i works best. I noticed the use of an ND filter in your credits…. was it a fader ND? If so, any comments on the brand, quality, or distortion?

  • Hi Dave,

    I have a 18 – 55 standard Canon Zoom and a 1.8 50ml Prime, the biggest problem for me is getting lots of DOF focus and sharpness for landscapes, I would love to know how you achieve the best results? For maximum DOF, what is highest F number you will go?

    Kind Regards
    Michael Nind
    Melb Australia

  • I tried som shots with different sharp settings aswell..
    I did shot with 1920×1080, 60fps 22f and sharpness 7.
    But of the shot, nothing got sharp… :S

  • I’m agree with Caleb. I always turn my sharpness settings all the way down. as well as contrast and saturation. I think it’s easier to make all the process in post prod. It’s also easier to have the best result by making your process into a computer screen than on a small dslr’s screen.

    That’s what I think

  • I’m 75 years old and have just taken up photography. So my old eyes have difficulty seeing the differences. So I’ll go with the concensus so people with better eyes won’t notice. (I hope)
    Anyway compliments on your site. I always enjoy it and have it on my favourites list.
    You call yourself as an amateur but it always looks professional to me.
    Cheers and keep it up. Most enjoyable
    (I just bought a T2i about two months ago. I sure have a lot to learn for an old geezer.)

  • No real difference. Had this been a movie there are no issues. Use the one you like best. Thank you Dave for felling another “they said” mantra.

  • ~A~ had more depth as well as well as a higher degree of sharpness. It makes sense to sharpen the picture before recording them, to me as well. I found about 3 of the B videos to be better then the A’s.
    I have my sharpness turned up half way, and like it, I don’t always do post processing so having it look good from the start is important to me.
    Great Article/test,Thanks Dave.