4K will require hardware update
- Thread starter evilthought
- Start date
Even if the FS700 were to get the bare minimum broadcast friendly bump like the C300 (4:2:2: 50mbps 8-bit) it would still be far inferior to the F5's internal 4:4:4 10bit 440 mbps codec so not sure what Sony would be worried about.
@ freiheit, SKYLIKE
This test is closer to a keylight test than a chart. That could give us a lot of information. What I can see from this test is that the detail in highlights and the schadows is still a lot better with the S-Log. Especially the room in the highlights are very promising. No more burned out hotspots!
A keylight test is really simple if you want to try, and the result is that you can tell the exact latitude of your camera. And if you add different colors as well, you can also tell if the camera reacts differently to certain colors when under and overexposed. This test was used in the old days mainly to know the latitude of filmnegatives with certain lights, make-up, costumes... Basically you light a subject (e.g. a person) with a key light (like you did on the statue) and in a perfect world the darkest side would be properly exposed and the brightest side would be 2 stops overexposed. If you only use one light it will be a least 4 stops I guess, so you can use some fill light from a reflective screen to bounce some light back from the key light. If you don't have a light meter, just make sure that the things that should draw the most attentions are your normally exposed things in the picture.
Once you have that, you do all the camera settings manually and start to change the exposure. Preferably you keep the same iso and aperture and change the shutterspeed. A perfect base exposure would have a shutterspeed around 1/64th of a second. That can mean that you have to put your light closer or further away from you subject or you have to dim the light. That should give you enough room to under and overexpose. You can add a number in the image to make sure you don't make mistakes afterwards and keep a list with all your settings. O would be the number which represents proper exposure. -1 would be 1 stop underexposed by dubbling the shutterspeed to 1/125th of a second. Next -2 would be 2 stops under doubling the shutterspeed again to 1/250th of a second. Next one -3 for 3 stops under and so on, all the way to -7 if you can pull that of. For overexposure we start with +1 for 1 stop overexposure by halving the base shutterspeed to 1/32nd of a second. Next +2 for 2 stops over halving the the shutterspeed again to 1/16th of a second. Next +3 for 3 stops over and so on all the way up to +7. So normally the end result would be a series of 15 images (-7,-6,-5,-4,-3,-2,-1,0,+1,+2,+3,+4,+5,+6,+7) which would show you the exact latitude (dynamic range of the camera) for all of the colors you tested. All the images where the properly exposed area is still visible and not completely white or black, count as a good shot. The ones just before it turns black or white determine the limits. Just to add: the biggest numbers can be hard to do if your shutterspeed can't go that far, then you can change the aperture if possible.
A test like this would be really helpfull and give a lot of information, even when it is basic, at least if you find that time and still have the camera. But I'm happy with what you've been doing so far ;-)
Cheers,
David
This test is closer to a keylight test than a chart. That could give us a lot of information. What I can see from this test is that the detail in highlights and the schadows is still a lot better with the S-Log. Especially the room in the highlights are very promising. No more burned out hotspots!
A keylight test is really simple if you want to try, and the result is that you can tell the exact latitude of your camera. And if you add different colors as well, you can also tell if the camera reacts differently to certain colors when under and overexposed. This test was used in the old days mainly to know the latitude of filmnegatives with certain lights, make-up, costumes... Basically you light a subject (e.g. a person) with a key light (like you did on the statue) and in a perfect world the darkest side would be properly exposed and the brightest side would be 2 stops overexposed. If you only use one light it will be a least 4 stops I guess, so you can use some fill light from a reflective screen to bounce some light back from the key light. If you don't have a light meter, just make sure that the things that should draw the most attentions are your normally exposed things in the picture.
Once you have that, you do all the camera settings manually and start to change the exposure. Preferably you keep the same iso and aperture and change the shutterspeed. A perfect base exposure would have a shutterspeed around 1/64th of a second. That can mean that you have to put your light closer or further away from you subject or you have to dim the light. That should give you enough room to under and overexpose. You can add a number in the image to make sure you don't make mistakes afterwards and keep a list with all your settings. O would be the number which represents proper exposure. -1 would be 1 stop underexposed by dubbling the shutterspeed to 1/125th of a second. Next -2 would be 2 stops under doubling the shutterspeed again to 1/250th of a second. Next one -3 for 3 stops under and so on, all the way to -7 if you can pull that of. For overexposure we start with +1 for 1 stop overexposure by halving the base shutterspeed to 1/32nd of a second. Next +2 for 2 stops over halving the the shutterspeed again to 1/16th of a second. Next +3 for 3 stops over and so on all the way up to +7. So normally the end result would be a series of 15 images (-7,-6,-5,-4,-3,-2,-1,0,+1,+2,+3,+4,+5,+6,+7) which would show you the exact latitude (dynamic range of the camera) for all of the colors you tested. All the images where the properly exposed area is still visible and not completely white or black, count as a good shot. The ones just before it turns black or white determine the limits. Just to add: the biggest numbers can be hard to do if your shutterspeed can't go that far, then you can change the aperture if possible.
A test like this would be really helpfull and give a lot of information, even when it is basic, at least if you find that time and still have the camera. But I'm happy with what you've been doing so far ;-)
Cheers,
David
Last edited:
Kraut69
Well-known member
On 4/18/13, I notified the sales department at B&H Photovideo by email that the advertising copy for the FS-700 was no longer accurate: "Hello, you are currently advertising the Sony FS-700 camcorder to be upgradeable to work with a Sony 4K Recorder via a firmware upgrade only. This is not true, as Sony announced at NAB 2013 that a hardware upgrade is also required for the Sony FS-700 to work with the Sony 4K recorder in addition to the firmware upgrade. When do you plan on correcting your advertising? This is not a trivial issue." They promptly replied on 4/21/13 , "Thank you for contacting the Sales Department at B&H Photo Video and Pro Audio.
Also thank you for making us aware of this issue. I shall alert our Webmaster of this issue.
In the future, you may avail yourself of the B&H web feature that allows you submit any corrections directly to the Webmaster.
Again thank you for pointing out this issue to us.
Please let us know if there is anything else we can assist you with.
Thank you, we appreciate your business. "
Today, I see the advertising copy has changed: "The 3G/HD-SDI port will get expanded capability when a future firmware and hardware upgrade is available (not included). This will allow 4K output via the BNC as a data stream, which can be recorded to an optional Sony 4K recorder."
Also thank you for making us aware of this issue. I shall alert our Webmaster of this issue.
In the future, you may avail yourself of the B&H web feature that allows you submit any corrections directly to the Webmaster.
Again thank you for pointing out this issue to us.
Please let us know if there is anything else we can assist you with.
Thank you, we appreciate your business. "
Today, I see the advertising copy has changed: "The 3G/HD-SDI port will get expanded capability when a future firmware and hardware upgrade is available (not included). This will allow 4K output via the BNC as a data stream, which can be recorded to an optional Sony 4K recorder."
paulcurtis
Veteran
freiheit,
Thank you for doing this, how did you work out exposure for the comparisons? I see the iso's are different so what did you use to ensure that they were exposed in a way to make the comparison work?
If you could do it again i think the scene is perfect but i would have added a black trap (something totally solid black in there) and perhaps exposed that so it was 0% and then let the highlights fall where they do to compare. Then perhaps do the opposite and pin exposure just before clipping to see how the shadows are.
I set up a lightbox with a transmission step wedge in from stouffer for my tests a while back, that's quite a cost effective way of doing it as the transmission wedges aren't so expensive. You'd need to backlight them but they show around 11stops in 1/3stop increments.
Thanks for taking the time, how did you find the camera?
cheers
Paul
Thank you for doing this, how did you work out exposure for the comparisons? I see the iso's are different so what did you use to ensure that they were exposed in a way to make the comparison work?
If you could do it again i think the scene is perfect but i would have added a black trap (something totally solid black in there) and perhaps exposed that so it was 0% and then let the highlights fall where they do to compare. Then perhaps do the opposite and pin exposure just before clipping to see how the shadows are.
I set up a lightbox with a transmission step wedge in from stouffer for my tests a while back, that's quite a cost effective way of doing it as the transmission wedges aren't so expensive. You'd need to backlight them but they show around 11stops in 1/3stop increments.
Thanks for taking the time, how did you find the camera?
cheers
Paul
@paul, david,
thank you very much for letting me know about the accurate test for dynamic range. as i mentioned before this firmware is still the beta test version, i found couple of minor bug and error...i bet they will fix these by the time we get the final version so i rather have a fun time with this baby now.
however, i brought her to the park and found nice result about highlight aliasing. the famous artifact, my major pain in ass.
TA-DA! there was no highlight aliasing with S-Log2! coooool!!!
*UPDATE*
i don't know but i couldn't upload the original size of jpag, 1920x1080, so please check the footage i will post tomorrow on the vimeo!
this is the capture from my small test, shot with ROKINON CINE 35mm T1.5. the camera pointed straight to the sun in the woods and the aperture opened T1.5 for totally overexposure. SINE4 and AB_NEXLOG had typical artifact we've seen but not with S-Log2! nice!
i will post the footage on vimeo tomorrow so please download the original ProRes file and tweak with your PC!
-----
freiheit
thank you very much for letting me know about the accurate test for dynamic range. as i mentioned before this firmware is still the beta test version, i found couple of minor bug and error...i bet they will fix these by the time we get the final version so i rather have a fun time with this baby now.
however, i brought her to the park and found nice result about highlight aliasing. the famous artifact, my major pain in ass.
TA-DA! there was no highlight aliasing with S-Log2! coooool!!!
*UPDATE*
i don't know but i couldn't upload the original size of jpag, 1920x1080, so please check the footage i will post tomorrow on the vimeo!
this is the capture from my small test, shot with ROKINON CINE 35mm T1.5. the camera pointed straight to the sun in the woods and the aperture opened T1.5 for totally overexposure. SINE4 and AB_NEXLOG had typical artifact we've seen but not with S-Log2! nice!
i will post the footage on vimeo tomorrow so please download the original ProRes file and tweak with your PC!
-----
freiheit
stefanhuberfilms
Veteran
thanks for sharing freiheit, glad that slog2 also adresses the highlight issue, looking forward to the footage!
paulcurtis
Veteran
ProRes isn't AFAIK a supported codec yet on the Odyssey. Dan might step in to clarify though.
My belief is that the Odyssey will record full colour depth uncompressed DPX files as standard. There is no official clarification on what this colour depth is, from previous literature and info it was supposed to be 12 bit but with the hardware upgrade this may change.
I'd hardly called 12 bit RAW crippled though
Perhaps 8 bit yes....freiheit,
That's really odd! The blooming is a sensor artefact, not something generated by the AD converter. I suspect the same data is there but just obscured by better highlight handling. Very interesting find.
thanks
Paul
Last edited:
Dan Keaton
Well-known member
ProRes isn't AFAIK a supported codec yet on the Odyssey. Dan might step in to clarify though.
My belief is that the Odyssey will record full colour depth uncompressed DPX files as standard. There is no official clarification on what this colour depth is, from previous literature and info it was supposed to be 12 bit but with the hardware upgrade this may change.
I'd hardly called 12 bit RAW crippled though
Paul
Dear Friends,
Our first codec for the Odyssey7 and Odyssey7Q will be Avid DNxHD. Then we will work on, and then attempt to get certification for another popular codec.
Our work on Avid DNxHD will help us greatly in our work on the other codec.
Yes, our Gemini 4:4:4 records the full bit-depth, in DPX file format (currently) that the camera gives us.
On some cameras this is 10-Bits, on others it is 10-bits and 12-Bits.
Please note, that the bit-depth, at least on some cameras, can change depending on the mode that the camera is in.
For example, the Canon C500 can output 12-Bits in HD and 2K. In 4K Raw data is output in 10-Bits.
Please understand that at 4K, at 60 fps, at 10-Bits, the datarate is 666 MBps (that is Megabytes per second, two-thirds of a Gigabyte per second, so please understand that these decisions to output 10-Bits or 12-Bits are not taken lightly by camera manufacturers, there are a lot of considerations.
Respectfully,
Dan Keaton
Convergent Design
rogar-Shock(DBG): "The number of the beast"i just uploaded the another sample from FS700+S-Log2 stuff.
FS700+S-Log2 : SUNNY DAY IN THE PARK
https://vimeo.com/65408916
please download the ungraded data and check the difference between CINE4, AB_NEXLOG, and S-Log2.
@paul,
yes, you're absolutely right! i believe that highlight artifacts causes by sensor too. i have meeting with Sony people next week so i will add this thing to my question list!
FS700+S-Log2 : SUNNY DAY IN THE PARK
https://vimeo.com/65408916
please download the ungraded data and check the difference between CINE4, AB_NEXLOG, and S-Log2.
@paul,
yes, you're absolutely right! i believe that highlight artifacts causes by sensor too. i have meeting with Sony people next week so i will add this thing to my question list!
guy i just put one image about the neon sign comparison i did last night:
http://www.freiheit.jp/FS700/SLOG2_TEST/AKB_UEN/NEON.html
i will post this as the 3rd test very soon but see the difference between these 3 setting. it's still kinda mystery about that "highlight artifacts" issue however i clearly see the difference between them and i totally am addicted to use S-Log2!
hope we will get even better one as the final version next month. i can't wait!
cheers,
freiheit
http://www.freiheit.jp/FS700/SLOG2_TEST/AKB_UEN/NEON.html
i will post this as the 3rd test very soon but see the difference between these 3 setting. it's still kinda mystery about that "highlight artifacts" issue however i clearly see the difference between them and i totally am addicted to use S-Log2!
hope we will get even better one as the final version next month. i can't wait!
cheers,
freiheit
paulcurtis
Veteran
Great results from the highlights - especially on the neon, what a huge difference. The tone jump is more an issue of the fact it's 8 bit only. Just imagine 10 bit slog out - how good would that be?
What would be very useful would be a timeline for the firmware and hardware upgrades if you get to talk to people 'in the know'. My 700 is back to sony and ideally i'd like to combine both jobs.
cheers
Paul
S.Blanchardon
Well-known member
This is getting exciting ! :thumbsup:
And thank you so much for your tests freiheit !
どうもありがとうございました !
And thank you so much for your tests freiheit !
どうもありがとうございました !