Forum Replies Created

Page 30 of 317
  • did capture noise print noise reduction work better for you than automatic noise reduction? just curious.
    a really nice trick for noise reduction is choose different FFT sizes, listen for output only in multiple passes.

  • it just adds a little noise, in convert sample type, use dither gaussian and then automatic noise removal. default settings. will sound great, well, at least for 8 bit lol. if you can keep the 22khz, that’s good too.

  • Chris Wright

    February 24, 2019 at 6:00 pm in reply to: Dialogue Rescue Needed

    since it is clipped so harshly(about half the waveform is gone), you can only declip so much. So after that, you have to focus more on reducing the overal harshness. Our ears are sensitive to transients, dynamics, and noise.
    in RX, i did in this order, declip, dialogue isolate, voice denoise. Then in, audition, studio reverb 100hz. classical and finally multiband compression.
    https://f1.creativecow.net/13150/dialogue-clipping?uploaded=file

  • Chris Wright

    February 20, 2019 at 4:39 am in reply to: All Projects Content Disappeared?

    was it an external drive? can you attempt data recovery?

  • Chris Wright

    February 19, 2019 at 5:03 am in reply to: Exposing for late afternoon or evening shots

    another IRE. I should have made that more clear.

  • Chris Wright

    February 18, 2019 at 5:39 pm in reply to: Exposing for late afternoon or evening shots

    log is just a word. its how the bits are spread out over a logarithmic curve. in terms of precision, the highest quality point on the log graph is where the curve is exactly 45 degrees(pure linear). this is also where you get the most precision with white balancing. (this is why native white balancing rec. 709 used to be 90 but log is 50, also known as white balance compression)

    a good rule of thumb for vlog :Exposure for VlogL is 55 IRE for Caucasian skin tones, and 80 IRE for highlights.
    whenever you deviate from this, you are losing bits, in the shadows or in the highlights, you have to decide which is more important. always try to get it exposed right in the camera, then grade in post.

    you can light some scenes a certain way, but that is the way they will end up, with less wiggle room. back in the days of film, they did this right the first time as film was very expensive.

    one modern trick with CMOS sensors is that they love 5000k over 2000k lights for less noise, even the RED camera does.

  • Chris Wright

    February 17, 2019 at 6:00 am in reply to: Converting Prores RAW without loosing quality

    I understand what you’re saying. you could go Apple ProRes 4444 XQ 12 bit 495 mbits/sec or 2,237 mbit/sec full uncompressed. prores can go multiple generations with low PSNR values. https://www.apple.com/final-cut-pro/docs/Apple_ProRes_White_Paper.pdf
    or you could go EXR half float for perfect math copy but your hard drive would fill up too fast.

    even the prores 422hq 10 bit is good for multiple generations so I’d do a few tests to see what you could get away with.

    I usually match up the bits to determine my codec as that determines precision with colors. If your Atomos Shogun
    only shoots 10 bits, 4444XQ is probably overkill.

  • Chris Wright

    February 12, 2019 at 4:41 am in reply to: Imitating early TV flattened out color

    it might be hard to tell unless your eyes are trained for it, but its about 50% set design, 10% wardrobe, 20% film stock, 10% lenses, and 10% post grade.

    here’s a rec. 709 pure color no luma modification that would get you closer. I fiddled around a while ago after researching this on the HSL and YUV scopes.

    Essentially in HSL, you raise saturation, but lower lumanence so its practically zero, then rotate skin to almost pure red.

    “Douglas Monce The look of the original Star Trek has a great deal to do with the way Gerald Finerman Perry photographed the show. The lighting style was very much influenced by Classic Hollywood films, of the 40’s including film noir, with an emphasis on unusual colored lighting. Perry would Splash walls with purple or green lighting in the background. Even using colored lighting on the hair lights of the actors. He used hard key lights with fill, very rarely if ever using diffusion on the lights. He frequently used soft filters on the lenses for close-ups of both men and women.”

    The 5254 kodak film stock has very little luma per chroma and very creamy whites. D55 white point at least.(D55 can be emulated by lowering blue whitepoint in color mode)

    To get this, luma has to be crushed before even adding chroma curve with saturation compression. the skin tone is in between skin line and red on vectorscope. putting it all together below:

    64 iridas lut
    https://drive.google.com/open?id=1hjFy8sD2o3cIjwR6r7jjM2KQmv_rTmAv

    also, quite a lot of well known films were shot on it.
    https://shotonwhat.com/film-negative-stock/eastman-color-negative-100t-52547254?fbclid=IwAR0IVIAKq44PJk88OLeSu8ZkOkqV70k60PblTBOKjGppguoCNaTxfJ6wxoU

  • Chris Wright

    January 13, 2019 at 5:12 am in reply to: Matching C100 and T6i

    shoot both with largest dynamic range as possible and no major contrast burned in. then use a color chart to match them with resolve’s auto chart match. one might have a larger dynamic range so you’d have to force clip the better camera if you wanted to match them on a sunny day.

  • Chris Wright

    January 4, 2019 at 5:09 pm in reply to: Footage getting color dimmed

    i 2nd ffmpeg. you’re trying to re-compress h.264 twice. you’ll get a lot of artifacts if not using a crazy high bitrate.
    if you re-import that rendered file back into AE, it will probably look exactly the same as 3rd party players and web browsers use various gamma and white point.

Page 30 of 317

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy