VideoHelp Forum

+ Reply to Thread
Results 1 to 24 of 24
Thread
  1. Herllo guys!

    Here is the short lossless original captured 4K video file (5GB)

    https://sendgb.com/7fZNs4iED29

    I recorded it in lossless NVENC_HEVC from the HDMI out with a laptop via Camlink 4K.

    HEre is the longer 2min version, which was later transcoded into a LOSSSY Hevc

    https://sendgb.com/O1nD0hIWqfv (4.2GB)

    Pay attention to the veining of the leaves and every little detail!

    What is your opinion about the video quality?
    Quote Quote  
  2. Video quality is good. It's a good sensor for a consumer camera in well lit conditions

    The re-encode removes some fine details and noise as expected, but it's still good considering the compression ratio is ~ 1:5 to 1:6

    If you posted the default onboard compression recording, I suspect the foliage detail would be very mushy, especially in the shadow areas such as underneath the trees. That would show the difference between default, and if a user should spend extra on external recording

    There is evidence that the camlink recording was quantized, as if some adjustments were made in 8bit - these don't negatively impact the image on this sample - but you can see gaps in a waveform. Maybe there are some settings you can adjust on the camlink or drivers
    Quote Quote  
  3. Originally Posted by poisondeathray View Post
    Video quality is good. It's a good sensor for a consumer camera in well lit conditions

    The re-encode removes some fine details and noise as expected, but it's still good considering the compression ratio is ~ 1:5 to 1:6

    If you posted the default onboard compression recording, I suspect the foliage detail would be very mushy, especially in the shadow areas such as underneath the trees. That would show the difference between default, and if a user should spend extra on external recording

    There is evidence that the camlink recording was quantized, as if some adjustments were made in 8bit - these don't negatively impact the image on this sample - but you can see gaps in a waveform. Maybe there are some settings you can adjust on the camlink or drivers
    Sorry Posondeathray, I made a mistake, I compressed the original video with 8bit NVENC HEVC, here is the 10bit lossy compressed version:

    https://sendgb.com/PENbJ9sRqpH

    It preserved more details than the 8bit NVENC_HEVC
    Last edited by Truthler; 30th Aug 2021 at 10:30.
    Quote Quote  
  4. Originally Posted by Truthler View Post

    Sorry Posondeathray, I made a mistake, I compressed the original video with 8bit NVENC HEVC, here is the 10bit lossy compressed version:

    https://sendgb.com/GRcTA3WgsrC

    It preserved more details than the 8bit NVENC_HEVC

    I don't see a mistake - The 2nd link in the 1st post was already 10bit420 HEVC - "LOSSY HEVC.mkv"

    Or did you mean original video, was the wrong original video ? ie . original recording was 8bit instead of 10bit ?

    I bet either re-encode looks better than the default onboard recording
    Quote Quote  
  5. Originally Posted by poisondeathray View Post
    Originally Posted by Truthler View Post

    Sorry Posondeathray, I made a mistake, I compressed the original video with 8bit NVENC HEVC, here is the 10bit lossy compressed version:

    https://sendgb.com/GRcTA3WgsrC

    It preserved more details than the 8bit NVENC_HEVC

    I don't see a mistake - The 2nd link in the 1st post was already 10bit420 HEVC - "LOSSY HEVC.mkv"

    Or did you mean original video, was the wrong original video ?

    I bet either re-encode looks better than the default onboard recording
    This second upload will be better, because there is no color shift.


    The moment when I felt that I can touch the leaflets of the trees on my Monitor: Watch this photo in full screen!

    http://users.atw.hu/educator/picture1.png
    Quote Quote  
  6. Yes, external recorder makes big difference. You should post example of onboard compression, to highlight differences

    Things like waving trees/leaves/foliage are very hard to compress and will turn to "mush" when using default onboard compression
    Quote Quote  
  7. Originally Posted by poisondeathray View Post
    Yes, external recorder makes big difference. You should post example of onboard compression, to highlight differences

    Things like waving trees/leaves/foliage are very hard to compress and will turn to "mush" when using default onboard compression

    I think it is also the 1" sensor, the XMOR RS version.

    Here is the development of XMOR sensors: https://www.framos.com/en/news/what-is-sony-s-exmor-technology-anyway

    Other consumer camcorders have smaller sensors, with more noise and lower real (visible) resolution, even the newsest Canon & Panasonic models have smaller sensors, because they manufacture large sensors only for their big broadcast and cinema videocameras, Sony always gives bigger sensor even for high-end consumer models... That makes the difference!
    Last edited by Truthler; 28th Aug 2021 at 16:43.
    Quote Quote  
  8. Originally Posted by poisondeathray View Post
    Yes, external recorder makes big difference. You should post example of onboard compression, to highlight differences

    Things like waving trees/leaves/foliage are very hard to compress and will turn to "mush" when using default onboard compression
    Did you play the files with Potplayer? Because the video was recorded in HLG format, and MPC and VLC play these videos with over saturization and over contrasting...

    Does it mean that these players (VLC and MPC-HC) have serious deficiencies?
    Last edited by Truthler; 30th Aug 2021 at 13:13.
    Quote Quote  
  9. If you recorded with specific settings, you have to use a LUT or tell the player what to use.

    There is no metadata in the NVENC file that tells player what to do

    e.g. if you record in slog2 or some other profile there is no way for player to know that so you get wrong playback
    Quote Quote  
  10. Originally Posted by poisondeathray View Post
    If you recorded with specific settings, you have to use a LUT or tell the player what to use.

    There is no metadata in the NVENC file that tells player what to do

    e.g. if you record in slog2 or some other profile there is no way for player to know that so you get wrong playback
    However Potplayer plays it correctly.... Other players can not handle the contrast and colors (over saturated) Is Potplayer more intelligent than other players?
    Quote Quote  
  11. Do you think the screenshot you posted in post#5 is accurate ? Is that what you "see" in potplayer ? How about when you were there (if you can remember) ?

    The main difference is just full vs. limited range. It's not actually using HLG to 709 mapping. So it's not really "smart" - it gets many camera footage wrong, because it simply cannot know what profile you shot with. You can se other player to use full range YCbCr to RGB , and it will look the same

    In order to display camera footage correctly, you must use correct display LUT
    Quote Quote  
  12. Originally Posted by poisondeathray View Post
    Do you think the screenshot you posted in post#5 is accurate ? Is that what you "see" in potplayer ? How about when you were there (if you can remember) ?

    The main difference is just full vs. limited range. It's not actually using HLG to 709 mapping. So it's not really "smart" - it gets many camera footage wrong, because it simply cannot know what profile you shot with. You can se other player to use full range YCbCr to RGB , and it will look the same

    In order to display camera footage correctly, you must use correct display LUT

    Hello Poisondeathray!

    What about this? (5GB lossless video)

    https://sendgb.com/slKxfzNgZq7

    Here is the snapshot: http://users.atw.hu/adatok1/bricks.png

    Visible individual brick pieces from 200-250m distance!
    Brutal, isn't it?
    Quote Quote  
  13. Snapshots of video frames. Enlarge it to full screen, and press F11.

    Do you see the tiny spokes of a motorcycle? Or the many small letters?



    Can you see the individual bricks on that huge disctance on the building?




    Can you see the small 1mm holes on the white sunshades of the windows? From such distance.



    Quote Quote  
  14. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    What I see in those last few images is lack of contrast. Not stupid consumer ideas of "contrast", cranking the software/TV knob to 11. But optical contrast of lenses. That's to be expected of lower-end consumer camcorders. Colors here are not just blah, but downright odd in some places, and it's not HDR. This will always negate any possible advantages by better colorspaces. Your optics are first for quality.

    The image is also completely flat, but that's a DOF issue (or choice?)
    Quote Quote  
  15. Originally Posted by lordsmurf View Post
    What I see in those last few images is lack of contrast. Not stupid consumer ideas of "contrast", cranking the software/TV knob to 11. But optical contrast of lenses. That's to be expected of lower-end consumer camcorders. Colors here are not just blah, but downright odd in some places, and it's not HDR. This will always negate any possible advantages by better colorspaces. Your optics are first for quality.

    The image is also completely flat, but that's a DOF issue (or choice?)
    Why don't you use HDR Display? I optimalized the settings for HDR displays. You can not optimalize a video for both HDR and SDR. IT is a question of choice.

    This represents my own taste, it is fully under manual controll, except the focus, which was auto.
    Last edited by Truthler; 4th Nov 2021 at 16:11.
    Quote Quote  
  16. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    Are those straight from camera? Or processed. Because I'm still seeing some oddness that can't be attributed to HDR, but can to optical contrast. I guess if the camera shot HDR, and it's optimized only for an HDR with post process, maybe.

    Is the choice due to limitations of the camera, either bokeh or aperture? Or just desire to shoot flat, sort of like a real-life cartoon?

    Anyway, good luck with it all. It's mostly about matching settings, basic video.

    I'm still acquiring my shoot setup,which can do 4Kp60, but I'll probably just stick to 1080p24/60 most times, no 4K TV, nor the desire to flood my HDDs with large video shoot files. I just cannot justify it yet.
    Quote Quote  
  17. Originally Posted by lordsmurf View Post
    Are those straight from camera? Or processed. Because I'm still seeing some oddness that can't be attributed to HDR, but can to optical contrast. I guess if the camera shot HDR, and it's optimized only for an HDR with post process, maybe.

    Is the choice due to limitations of the camera, either bokeh or aperture? Or just desire to shoot flat, sort of like a real-life cartoon?

    Anyway, good luck with it all. It's mostly about matching settings, basic video.

    I'm still acquiring my shoot setup,which can do 4Kp60, but I'll probably just stick to 1080p24/60 most times, no 4K TV, nor the desire to flood my HDDs with large video shoot files. I just cannot justify it yet.
    Things changed greatly from the late 2010s. The problem with consumer level DSLRs is that even top category mobile phones (from Iphone Samsung) can shot equal or better videos. Only DSLRs over 3000 Dollar category are better than them. The latest Samsung mobile shot in 8K. It is not as good as 8K professionals, but easily beat any 4K DSLRs. So you must buy professional camera to make better quality video than top m.phones.


    You watched the photos in SDR. When I turn on the HDR in Windows, the colors are too vivid and have too high contrast for HDR video. I had to adjust it to even flatter (by SDR standards). You don't need computer color grading, if you have experience with the internal settings of AX700.
    Quote Quote  
  18. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    Originally Posted by Truthler View Post
    Things changed greatly from the late 2010s. The problem with consumer level DSLRs is that even top category mobile phones (from Iphone Samsung) can shot equal or better videos.
    Not really. Price and age doesn't matter, the sensor does. You cannot compare a tiny cell phone sensor to a large dSLR sensor, even the cropped sub-35mm sensors (Nikon DX, etc). The SLR sensor is still 4x+ larger. As sensor size decreases, you lose DR and SNR. The crappy phone optics are super inferior in every way, and even a cheap SLR kit lens is better than the phone optics. Even pre-DSLR film lenses for SLRs (less fine resolve, not for mega MP era) are generally better.

    The latest Samsung mobile shot in 8K.
    You confuse the resolution palette with the actual resolve, which is itself separate from the other aspects of the image.

    It is not as good as 8K professionals, but easily beat any 4K DSLRs. So you must buy professional camera to make better quality video than top m.phones.
    My Z6 II will beat the crap out of a phone camera, hands down. And with a cheap kit lens. Put on some of my actual glass, and it's such a no-contest that it's LOL ridiculous. You cannot tell me a phone is going to shoot better than a good f/2.8 lens (or comparable f/4 or f/5.6 on much longer lenses). The sharpness alone does not exist on camera phone optics.

    You watched the photos in SDR. When I turn on the HDR in Windows, the colors are too vivid and have too high contrast for HDR video. I had to adjust it to even flatter (by SDR standards). You don't need computer color grading, if you have experience
    That's ... huh? You have HDR and SDR. If you have to tone down saturation of the HDR, then you have colorspace or shooting issues. Or it's a pre-processed image. You're not allowed to just randomly monkey with a monitor, in order to get color "just right". Never badly correct video to a non-calibrated viewing device. "Experience" absolutely does not preclude the need to calibrate, and in fact should reinforce that fact.

    I think it's safe to assume most computers are not using HDR monitors. Niche.
    Last edited by lordsmurf; 10th Nov 2021 at 17:44.
    Quote Quote  
  19. Member Cornucopia's Avatar
    Join Date
    Oct 2001
    Location
    Deep in the Heart of Texas
    Search PM
    @Truthler, you uploaded SDR images (as all the allowed picture filetypes- jpg,png,gif - are SDR-only). Why would you expect ANYONE to evaluate them using HDR criteria?
    If you expect to have HDR images/videos reviewed, you will need to upload as std. data file and link to it. E.g. OpenEXR.
    or
    You can tonemap (hopefully not just brute force "tone down") your HDR files to SDR and post them, understanding that they will be reviewed as SDR images at that point. And when doing so, it is better for tranparency of argument to divulge how you tonemapped.

    BTW, Your rant about "smartphone=good/new, mirrorless=good/new, dslr=bad/old" is as ridiculous as it is tiresome. Not just opinionated, but counter to long accepted scientific facts. Why don't you just stop at "I like the look of smartphones better" and "I like the mirrorless camera that I bought" and be done with it? I don't think anyone would have a problem with that, as that is totally personally subjective.
    If you truly want to do equipment shootouts, there are plenty of other gear sites that do that (though take care, as most of them are filled with pros & gearheads who know volumes about the tech, and they do employ those scientific methods of peer review that I've previously talked about).


    Scott
    Quote Quote  
  20. Originally Posted by lordsmurf View Post
    Originally Posted by Truthler View Post
    Things changed greatly from the late 2010s. The problem with consumer level DSLRs is that even top category mobile phones (from Iphone Samsung) can shot equal or better videos.
    Not really. Price and age doesn't matter, the sensor does. You cannot compare a tiny cell phone sensor to a large dSLR sensor, even the cropped sub-35mm sensors (Nikon DX, etc). The SLR sensor is still 4x+ larger. As sensor size decreases, you lose DR and SNR. The crappy phone optics are super inferior in every way, and even a cheap SLR kit lens is better than the phone optics. Even pre-DSLR film lenses for SLRs (less fine resolve, not for mega MP era) are generally better.

    The latest Samsung mobile shot in 8K.
    You confuse the resolution palette with the actual resolve, which is itself separate from the other aspects of the image.

    It is not as good as 8K professionals, but easily beat any 4K DSLRs. So you must buy professional camera to make better quality video than top m.phones.
    My Z6 II will beat the crap out of a phone camera, hands down. And with a cheap kit lens. Put on some of my actual glass, and it's such a no-contest that it's LOL ridiculous. You cannot tell me a phone is going to shoot better than a good f/2.8 lens (or comparable f/4 or f/5.6 on much longer lenses). The sharpness alone does not exist on camera phone optics.

    You watched the photos in SDR. When I turn on the HDR in Windows, the colors are too vivid and have too high contrast for HDR video. I had to adjust it to even flatter (by SDR standards). You don't need computer color grading, if you have experience
    That's ... huh? You have HDR and SDR. If you have to tone down saturation of the HDR, then you have colorspace or shooting issues. Or it's a pre-processed image. You're not allowed to just randomly monkey with a monitor, in order to get color "just right". Never badly correct video to a non-calibrated viewing device. "Experience" absolutely does not preclude the need to calibrate, and in fact should reinforce that fact.

    I think it's safe to assume most computers are not using HDR monitors. Niche.
    Large DSLR sensors in consumer category? Are you kidding?
    The Non Zeiss manufactured lenses in consumer level DSLR or camcorders can not give good quality picture either. Lens torsion values of other lens manufacturers are worse than Zeiss. No wonder that Zeiss produces the best lenses sins the establish of the firm.

    Think about the square grid booklet, take it, and shot some photos with consumer grade Canon JVC Panasonic etc camcorders or cameras...

    Image
    [Attachment 61861 - Click to enlarge]


    After the test shots, look the squares around the edges and corners of the picture. The pictures made by the Zeiss lens with Sony cameras looks like if somebody scanned them, they remain straight like if it would scanned by a traditional flat-bed scanner.... However other cameras and lenses produces more curved lines around the edges and corners...



    Think about the fancy and horror priced lenses of the NASA space telescopes, it is always Zeiss who can fullfil the brutal strict requirements for space telescopes.

    Second, Sony is the biggest innovator of imaging devices let it be CMOS CCD or anything since the 1970s, and he spend the most money for the CMOS/CCD innovation, and Sony bought up a lot of new licenses of other developers around the world.

    The net true photo-sensitive area of other sensors are smaller than the Sony sensors of the same "size".

    Neither of you can shot such quality videos with their consumer level DSLRs as I can with my two AX700 camcorers.

    Proof: 4min 9.8 Gb video
    http://sendanywhe.re/ZFICD1YZ

    If it is necessary turn on the limited range value.
    Last edited by Truthler; 17th Nov 2021 at 16:10.
    Quote Quote  
  21. Originally Posted by Cornucopia View Post
    @Truthler, you uploaded SDR images (as all the allowed picture filetypes- jpg,png,gif - are SDR-only). Why would you expect ANYONE to evaluate them using HDR criteria?
    If you expect to have HDR images/videos reviewed, you will need to upload as std. data file and link to it. E.g. OpenEXR.
    or
    You can tonemap (hopefully not just brute force "tone down") your HDR files to SDR and post them, understanding that they will be reviewed as SDR images at that point. And when doing so, it is better for tranparency of argument to divulge how you tonemapped.

    BTW, Your rant about "smartphone=good/new, mirrorless=good/new, dslr=bad/old" is as ridiculous as it is tiresome. Not just opinionated, but counter to long accepted scientific facts. Why don't you just stop at "I like the look of smartphones better" and "I like the mirrorless camera that I bought" and be done with it? I don't think anyone would have a problem with that, as that is totally personally subjective.
    If you truly want to do equipment shootouts, there are plenty of other gear sites that do that (though take care, as most of them are filled with pros & gearheads who know volumes about the tech, and they do employ those scientific methods of peer review that I've previously talked about).


    Scott
    With all of my respect:
    Theoretical concepts are good and smart things...but practical reality can be sometimes different.


    A good example:

    iPhone 12 vs Canon R5 vs RED

    https://youtu.be/GQA1UpamBiM
    Last edited by Truthler; 17th Nov 2021 at 16:09.
    Quote Quote  
  22. Member Cornucopia's Avatar
    Join Date
    Oct 2001
    Location
    Deep in the Heart of Texas
    Search PM
    NOT a good example.

    NOT a "shootout" (no direct A/B/C/X comparisons, no reference scenes or shooting methodologies).

    But it is a good example of how OVERLY ABUNDANT outdoor light can give enough exposure to hide noise issues, to hide framerate issues, to flatten contrast curve ends, to run super-small shutter angles to hide rolling shutter issues. Of how slo-mo & fast-mo can hide issues with framerate consistency. Of how predominance of long shots can avoid comparisons between depth-of-field.

    Scott
    Quote Quote  
  23. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    Originally Posted by Truthler View Post
    Zeiss manufactured lenses
    Zeiss is a brand.
    Which exact lens is it that gives you these Zeiss-gasms?

    in consumer level DSLR or camcorders can not give good quality picture either.
    Eh, they work fine. The main issue is performance in low light, features like adding external mic. For example, I have a D3500, but it's limited. It is far better than the phone in my pocket, but I won't shoot epic movies with it.

    Lens torsion values
    Term does not exist. What are you referring to here?

    . No wonder that Zeiss produces the best lenses
    Zeiss makes both good and bad lenses, including tiny crappy plastic optics on cheap consumer items.

    Think about the square grid booklet, take it, and shot some photos
    To what extent? What is the point here?

    After the test shots, look the squares around the edges and corners of the picture. The pictures made by the Zeiss lens with Sony cameras looks like if somebody scanned them, they remain straight like if it would scanned by a traditional flat-bed scanner.... However other cameras and lenses produces more curved lines around the edges and corners...
    That's ridiculous. Part of this is determined by the angle of the glass. A good prime 85mm won't do that, be it Nikkor or Canon or whoever.

    Think about the fancy and horror priced lenses of the NASA space telescopes, it is always Zeiss who can fullfil the brutal strict requirements for space telescopes.
    Zeiss hasn't made a NASA lens in 50+ years, and it wasn't for a telescope. The current work-in-progress JWST isn't even finished, delayed for years. And the JWST doesn't even use the Zeiss as primary optics, merely instruments on board. JWST has taken so long to R&D that it's years obsolete image-wise. I still remember when it was called the Hubble 2 (then 2.0 -- yeah, it was in R&D before the term "2.0" was even in use).

    Second, Sony is the biggest innovator of imaging devices
    This part is very true. Or was. Nikon and Canon give Sony a run for its money these days. That new Z9 just handed Sony crow to eat.

    The net true photo-sensitive area of other sensors are smaller than the Sony sensors of the same "size".
    False. It fully depends on factors.

    Neither of you can shot such quality videos with their consumer level DSLRs as I can with my two AX700 camcorers.
    The AX700 is a "rich consumer" camcorder (not a camera) in the sub-$2k range. In terms of overall camera+mic+lens, that's essentially a mid-range camera at best. I have glass that costs more than that entire camera.
    Quote Quote  
  24. Originally Posted by lordsmurf View Post
    Originally Posted by Truthler View Post
    Zeiss manufactured lenses
    Zeiss is a brand.
    Which exact lens is it that gives you these Zeiss-gasms?

    in consumer level DSLR or camcorders can not give good quality picture either.
    Eh, they work fine. The main issue is performance in low light, features like adding external mic. For example, I have a D3500, but it's limited. It is far better than the phone in my pocket, but I won't shoot epic movies with it.

    Lens torsion values
    Term does not exist. What are you referring to here?

    . No wonder that Zeiss produces the best lenses
    Zeiss makes both good and bad lenses, including tiny crappy plastic optics on cheap consumer items.

    Think about the square grid booklet, take it, and shot some photos
    To what extent? What is the point here?

    After the test shots, look the squares around the edges and corners of the picture. The pictures made by the Zeiss lens with Sony cameras looks like if somebody scanned them, they remain straight like if it would scanned by a traditional flat-bed scanner.... However other cameras and lenses produces more curved lines around the edges and corners...
    That's ridiculous. Part of this is determined by the angle of the glass. A good prime 85mm won't do that, be it Nikkor or Canon or whoever.

    Think about the fancy and horror priced lenses of the NASA space telescopes, it is always Zeiss who can fullfil the brutal strict requirements for space telescopes.
    Zeiss hasn't made a NASA lens in 50+ years, and it wasn't for a telescope. The current work-in-progress JWST isn't even finished, delayed for years. And the JWST doesn't even use the Zeiss as primary optics, merely instruments on board. JWST has taken so long to R&D that it's years obsolete image-wise. I still remember when it was called the Hubble 2 (then 2.0 -- yeah, it was in R&D before the term "2.0" was even in use).

    Second, Sony is the biggest innovator of imaging devices
    This part is very true. Or was. Nikon and Canon give Sony a run for its money these days. That new Z9 just handed Sony crow to eat.

    The net true photo-sensitive area of other sensors are smaller than the Sony sensors of the same "size".
    False. It fully depends on factors.

    Neither of you can shot such quality videos with their consumer level DSLRs as I can with my two AX700 camcorers.
    The AX700 is a "rich consumer" camcorder (not a camera) in the sub-$2k range. In terms of overall camera+mic+lens, that's essentially a mid-range camera at best. I have glass that costs more than that entire camera.
    You selectively quoted some of my lines and your answer which was refuted countless times. Your repetition of your texts does not make your claims, which have long since been refuted, true.

    Maybe you checked recent prices of late 2021.
    When AX700 was released it was between 3000 and 3500$. I purchesed two Ax700 camcorders in 2018. You need prof dead-cat microphones for camcorders, because the built in mics are not WINDPROOF.

    My AX700 can beat your D 3500 even in an overcast rainy day.

    Proof video: https://sendgb.com/W5PkGlAYvoU
    Last edited by Truthler; 5th Dec 2021 at 14:11.
    Quote Quote