Personal View site logo
Make sure to join PV on Telegram or Facebook! Perfect to keep up with community on your smartphone.
Noise when shooting VIDEO of landscape with aperture of 22
  • Very bright sunny day standing on a hill looking across to other large hills far away.

    14-140mm kits lens, 50 shutter speed, ISO 160, aperture closed down for correct exposure to 22 (no ND used).

    Once I got back and checked footage on large screen I noticed this video had quite a lot of noise.

    The shot that I took with the Voiglander 25mm f0.95 didn't seem to have much noise. I think this was set at f2.8 and then I used a lightcraft variable ND to get the correct exposure.

    So my question is... Does closing down the aperture fully (22) create this horrid noise?

    Does anyone have any advise on what would be a good aperture to use when shooting landscapes? Is there a recommended aperture not to go passed, like 16 or something?

  • 29 Replies sorted by
  • Are you sure the ISO was locked at 160? Did you convert/edit the video with any software? Can you post a short clip? If the provided image isn't bright enough, it may add noise by "intelligent" increase of iso value?

  • 100% that the iso was 160... and everything else manual as described... Nothing edited... this is the raw MTS file. Not sure how it would increase the iso intelligently as iISO is always off. I'll get a clip uploaded asap. Thanks holmes6 for you're help.

  • can you provide a screen grab? at f22 you maybe seeing dust marks on sensor or the lens its self

  • Guy's anyone know a good site where i can upload the 98mb mts file... Unfortunately i don't have my ftp details here with me for my server.

  • R.I.P megaupload :(

  • What was the picture profile used? Patch settings?

  • @bwhitz I was using smooth with -2, +1, 0, -2. And the spanmybitchup hack.

    MTS files almost uploaded, one sec.

  • 14-140mm kits lens, 50 shutter speed, ISO 160, aperture 22. http://dl.dropbox.com/u/59273165/00026.MTS

    Same as above but zoomed into 140mm http://dl.dropbox.com/u/59273165/00036.MTS

  • Vimeo will work as well.

  • Shooting at 24p. Smooth -2, +1, 0, -2. Is it because I have NR at -2? This is something I have overlooked and if this is the case I didn't realise setting NR to -2 would cause this much noise. I have previously compared +2 NR with -2 NR in my back garden and didn't notice much difference, but now with this landscape shot could this be why I am now getting noise?

  • @Jrb 140mm zoom looks to me like you were using EX TeleConv.

  • Due to diffraction it's generally not good to go over f8-9 (if you want the sharpest possible image), however that should not create noise. I don't know the reason in this case, other than that the camera produces noise in the shadows, all over the iso spectrum. For minimal noise it is better to overexpose (make sure to protect whites).

  • Smooth also has a bit more noise...

  • @Tommyboy yes i think your right... ignore clip 36 then.

    I'll go out today and make some more tests... this time noting down all settings

  • @RRRR Many many thanks for you advice here... I did some more tests and it appears that there is noise in the shadows at aperture 22.... when bringing this down to an aperture of 8 and using some ND for correct exposure I have hardly any noise. So this is now resolved and I will never go over an aperture of 8 when doing a landscape shot. Many thanks.

  • @Jrb No problem! I guess what that must mean from a theoretical perspective is that an unsharp image introduces false detail = noise (in a highly detailed image). Or, it can be that the ND filter helps to flatten the image and keep exposure range within better limits.. Could you check your histogram to see whichever it is? If the histogram is pretty much the same then it means that diffraction does introduce noise in detailed scenes.

  • I couldn't get the 8 aperture to play back on this computer, but from looking at the 22 ap - it definately seems like the noisyness could be due to diffraction unsharpness and a lack of real detail in darker areas.

  • jrb, this is true on all cameras... Stay away from closing down that far or else suffer from lens diffraction and noise levels.

    My old JVC HD100 was horrible at f16 - real soft due to diffraction and quite the noise bath.

  • Thanks everyone for your help... I just went out and shot the same scene and found that aperture 4 with some ND on gave me the best results on the 14-140 lens.

    But this time I also compared 3 different firmwares: The SpanMyBitchUp hack, the mpgxsvcd 40mbps hack and the original. To my surprise the original firmware looks to have better detail. Any thoughts on why this is happening???

    I will upload the 3 mts files now.

  • Here are the firmware comparisons, all shot at 50 shutter speed, 160 ISO, 4 aperture and smooth -2+1+0-2:

    The original firmware looks better that the higher bitrate hacks... Any ideas why?

    SpanMyBitchUp hack: http://dl.dropbox.com/u/59273165/00002%20SMBU%20FM%204ap.MTS

    Original Firmware http://dl.dropbox.com/u/59273165/00005%20ORIG%20FM%204ap.MTS

    mpgxsvcd 40mbps hack http://dl.dropbox.com/u/59273165/00010%20mpgxsvcd%20FM%204ap.MTS

  • Anyone have any idea why I am experiencing a better image on high detail landscape scenes with the original firmware compared to the SpanMyBitchUp 170Mpbs hack? All settings shot 100% same.

  • Hang on, I'll check it out. I can say straight off that the Original Firmware does pretty well in well lit static scenes, because the information it has to handle is less complex. For gop1 you have a constant bitrate - the same amount of information for each frame.

    Also, for a green scenery like the one you have shot you've got most information in the green channel, which has to most information out of the box. (In other words, although it should not perhaps perform better, it should perform very well).

  • Watching the SpanMyBitchUp file it does not look like it should.. It might have gone in to fall-back mode due to the high detail, which is a lower quality "escape" setting for when the encoder cannot keep up.

    (perhaps someone with streamparser can reject or confirm this)

    and also - how long did you record for? and what sdhc card did you use?