Odyssey Firmware 2015.5

And while I'm laying into Catalyst Browse (sorry Sony, but if anybody there is reading this, please take it as constructive criticism) I notice that it supports external SDI monitoring but the output range is inconsistent with rendered movies.

I have Catalyst Browse outputting via SDI on a BMD Mini Monitor to my Odyssey. Set to S-Gamut/S-Log2 with no look profile, the Catalyst waveform clips at 97% and the Odyssey one at 106%. This is at least consistent with rendered ProRes files. With the LC_709TypeA look profile applied, the Catalyst waveform clips at 98%, but the Odyssey one clips at 107%. This is not consistent with rendered ProRes files, which also clip at 98%.

Sony, please at least be internally consistent with your inconsistency!
 
Last edited:
This is starting to sound like the old 7.5 IRE pedestal conundrum of NTSC days....never to be resolved until REC2020 becomes de riguer.
 
cuervo said:
This is starting to sound like the old 7.5 IRE pedestal conundrum of NTSC days....never to be resolved until REC2020 becomes de riguer.
Click to expand...
It's very similar. Just like 7.5 IRE setup, the existence of super-whites is a throw-back to analog video, where it was necessary to allow for some over-shoot. It was preserved in the standards for digital video, because that was originally captured from analog sources (yes, I know a camera sensor is an analog device, but let's ignore that) so needed to allow for preserving analog over-shoot. In my opinion, Sony's decision to continue using the super-white range in their log formats causes confusion, and often requires work-arounds to avoid clamping it off. Just like the complications of incorrect digital handling of the analog 7.5 IRE setup.
 
this whole thread has made for some very informative reading. I have a question: I find I get the best results with the FS700 and Odyssey exposing S-Log2 about a stop and a third to a stop and a half over. With zero experience in creating a LUT in Resolve or Premiere, how difficult would it be change a preexisting LUT to reflect this exposure difference?
 
digger said:
this whole thread has made for some very informative reading. I have a question: I find I get the best results with the FS700 and Odyssey exposing S-Log2 about a stop and a third to a stop and a half over. With zero experience in creating a LUT in Resolve or Premiere, how difficult would it be change a preexisting LUT to reflect this exposure difference?
Click to expand...
Two of the LUTs I created for Convergent Design (available for download in the Additional LUTs package) are LC709A LUTs for S-Log2 with one and two stop exposure compensation.
 
digger said:
These are the ones I've been using so far, and they look great. Do you think you'll do one in between? I feel like the 2 stop is cutting it a little too close to highlight clipping (I do a lot of doc work), and 1 stop isn't quite enough.
Click to expand...

I've NEVER been a fan of using an electronic LCD viewfinder to set exposure. It's frought with the potential for WAY too many errors. IMHO, it's much more accurate to expose an UNLUTTED image so that zebra comes on for 90% white at 90 IRE. if you want a backup, set the second zebra for 18% gray at 40 IRE. These exposure values work very well for me. If you want to look at a pretty image, or want to send a pretty image out to your director or client, go ahead and use the LC709A LUT, or whatever LUT you like that makes a pleasing picture. Just make sure your zebra, WFM and histogram are displaying for unlutted images.
 
digger said:
These are the ones I've been using so far, and they look great. Do you think you'll do one in between? I feel like the 2 stop is cutting it a little too close to highlight clipping (I do a lot of doc work), and 1 stop isn't quite enough.
Click to expand...
In Resolve, if you load in a clip which is 1.5 stops over, you can apply my -1 LUT and then reduce the exposure a bit more with the offset control until it looks right. Then right-click the thumbnail and "Export 3D LUT".
 
cuervo said:
I've NEVER been a fan of using an electronic LCD viewfinder to set exposure. It's frought with the potential for WAY too many errors. IMHO, it's much more accurate to expose an UNLUTTED image so that zebra comes on for 90% white at 90 IRE. if you want a backup, set the second zebra for 18% gray at 40 IRE. These exposure values work very well for me. If you want to look at a pretty image, or want to send a pretty image out to your director or client, go ahead and use the LC709A LUT, or whatever LUT you like that makes a pleasing picture. Just make sure your zebra, WFM and histogram are displaying for unlutted images.
Click to expand...
In general I agree with you, but I will also state that I am not one to tell others how they should work. I talk about how I am more comfortable working. I would also state that the OLED panel on the Odyssey is extremely accurate as are the waveform, false color and other tools. Of course no monitor is all that accurate under difficult viewing conditions, so one really has to decide how he/she is comfortable using the available tools.
 
Mitch Gross said:
In general I agree with you, but I will also state that I am not one to tell others how they should work. I talk about how I am more comfortable working. I would also state that the OLED panel on the Odyssey is extremely accurate as are the waveform, false color and other tools. Of course no monitor is all that accurate under difficult viewing conditions, so one really has to decide how he/she is comfortable using the available tools.
Click to expand...

Agreed, Mitch! My apologies for sounding professorial. While this describes my own practice, I am no one to tell someone else how to work. Again, my apologies for sounding as if i was.

I beg your indulgence for a small rant. It seems to me this whole LUT thing is running a wee bit amuck. This thread, alone, shows the complexities of their use. Not the least of which is having to understand different exposure threshholds for each and every different LUT. It's just so much simpler to learn the exposure threshhold for a camera's Log, one time. For whatever reason, even high end cameras, like my Sony F5, change the zebra, wfm/histogram with an MLUT. Rather counterproductive, I think.<End of Rant>
 
Last edited:
What happened to the false colors? I tapped it on today just checking stuff after the firmware upgrade and now everything is red with occasional green is certain areas. (non of the other colors are there any more.) was that by default now? I know I can change all the settings to show all the different ranges but was it setup that way (the 2 colors) by default?
 
Chris Johnston said:
What happened to the false colors? I tapped it on today just checking stuff after the firmware upgrade and now everything is red with occasional green is certain areas. (non of the other colors are there any more.) was that by default now? I know I can change all the settings to show all the different ranges but was it setup that way (the 2 colors) by default?
Click to expand...

The bottom value for all your ranges is probably set at 0%. Bring them up so the ranges don't overlap and you'll be good.
 
When recording 100/120/200/240fps into PRORES HD/2K, does the image stutter/seem choppy on the live view of the odyssey 7q+? Playback is fine though.
On the other hand when recording in RAW, live view is perfectly smooth.

Is this normal?
 
cuervo said:
I've NEVER been a fan of using an electronic LCD viewfinder to set exposure. It's frought with the potential for WAY too many errors. IMHO, it's much more accurate to expose an UNLUTTED image so that zebra comes on for 90% white at 90 IRE. if you want a backup, set the second zebra for 18% gray at 40 IRE. These exposure values work very well for me. If you want to look at a pretty image, or want to send a pretty image out to your director or client, go ahead and use the LC709A LUT, or whatever LUT you like that makes a pleasing picture. Just make sure your zebra, WFM and histogram are displaying for unlutted images.
Click to expand...
I agree that zebras and waveforms are generally better for setting exposure, but LUTs are useful for previewing what the end result will look like. If you are deliberately over-exposing by a stop to reduce shadow noise, then a preview of what the image will look like when "printed down" is useful to see.

I would also add a few other comments. If you are using S-Log2 then 40 IRE does not represent 18% grey, and 90 IRE certainly is never 90% white. If you look at the table below, taken from Sony's S-Log documentation, you will see that in S-Log2 18% grey is represented by 32 IRE. If you set grey at 40 IRE you are actually over-exposing by about 2/3 of a stop. Nothing to say you can't do this if you want to, of course.

90% white is represented by 59 IRE in S-Log2. If you want the numbers, 90 IRE represents nearly 500% which is way up in the specular highlight region, being about 4.75 stops above 18% grey.

S-log_values.png

Note: I have cleaned up the table slightly from Sony's original to remove some potentially confusing information
 
Last edited:
Nick,
Thanx for the info. This Sony data is certainly relevant for Slog, 2,3, as published by Sony.This thread is about the FS700, yes? I've had this discussion with Mitch, the "Slog2" encoded files that C-D writes to the O7Q from the FS700, as you know, do not reflect this exposure guidance.( I suppose i should make note of the fact that the RAW data out of the FS700 is non-compliant with this Sony exposure guidance) If I expose for this guideline, I get horribly noisy and apparently underexposed footage. As C-D has recommended(and you know) overexposing by 1-2 stops, and pulling exposure back in post, is much more appropriate. At least, this has been my experience. Exposing 90% white at 90 IRE is consistent with a 2 stop overexposure. And I've found this to be my practice, rather than try using a 2 stop underexposure LUT, and exposing to these Sony values, as you have suggested. I suppose either methodology works, however, I prefer not to get distracted by relying on a LUT. Too many times, 3rd party LUTS are published which just don't seem to work as described by the publisher. Not sayin' this is true for your LUTS, it's just another factor that I would need to prove out. SOSO. Pick your poison. I think the most important practice is to pick one methodology/workflow and stick to it. Trying to use a variety of different methodologies is just confusing and adds to the possibility of making a mistake. Having said that, I routinely use a PMW-F5 on one day and an FS700 on another say. trying to keep straight, in my weak brain, which LUT I should be using and what exposure values I need to apply is just too error prone. At least for my aging brain cells.

Agreed. Having an image in my VF that represents the REC709 color space is much preferable to one in the Slog space. Choosing a good VF MLUT that could be used to get a fairly representative image for judging contrast and balance is important, I don't disagree.But, then, the implication here is that the same LUT be applied in post, at least as a starting point. I wonder how many users in these fora, do that. Accepting a 3rd party LUT without understanding what went into the creation of that LUT, or how to use it, is just problemmatic. For example, getting my head around your distinction between External and Legal labels for setting video range. If I hadn't read this post, and I started using your LUTS without understanding them, I could, potentially, be clipping some data. Case in point.

As a small side note, i really wish we could all move on from this REC709 space. Like 601, REC709 is kinda passe; and, has a bunch of problems associated with trying to fit 14 stop DR cameras into a 5 stop DR design. yes, I know TV's and monitors are still REC709, but, this too will change. I hope that by the time REC2020 becomes de riguer, these DR, superblack, superwhite, and log exposure, issues go away.

edit: excellent excellent article by Art Adams- http://www.provideocoalition.com/the-art-adams-zone-system-for-hd/page-2
 
Last edited:
I would request that CD find a way to avoid having to build the playlist on every boot. I was shooting the other day and shot about 200 clips and it was taking probably 20-30sec just building the playlist every time I booted up, and I didn't even use any playback which I assume is what the playlist is for.
 
cuervo said:
As C-D has recommended(and you know) overexposing by 1-2 stops, and pulling exposure back in post, is much more appropriate. At least, this has been my experience. Exposing 90% white at 90 IRE is consistent with a 2 stop overexposure. And I've found this to be my practice, rather than try using a 2 stop underexposure LUT, and exposing to these Sony values, as you have suggested. I suppose either methodology works, however, I prefer not to get distracted by relying on a LUT.
Click to expand...
I am not advocating applying a LUT and judging exposure from the LUTted waveform/zebras. That is the beauty of the Odyssey LUT system – you can view a LUTted image, and have your exposure tools work off the log.

I was showing Sony's table as a reference for what their "normal" IRE values for different exposures are. I was not suggesting you should use those values if you want to over-expose. My LUTs include exposure compensation AND the LC709A LUT concatenated, so looking for the Sony values downstream of the LUT would be inappropriate anyway.

For reference, here are the IRE values in S-Log2 and S-Log3 for 1 and 2 stops of over-exposure:

S-Log2
18%90%
032 IRE59 IRE
+143 IRE72 IRE
+255 IRE84 IRE



S-Log3
18%90%
041 IRE61 IRE
+149 IRE70 IRE
+258 IRE79 IRE

But these are just the mathematically calculated values from Sony's formulae. Whatever works for you to get the result you want is the right approach for you.
 
ahhhh...thanx, Nick. I'll play with your numbers. My numbers are rough, so, maybe I can find a tweak, for the better, with yours.
Oh, and I had misunderstood you, earlier. Yes, in fact I use LC709A or sometimes REC709(800%) as my display LUT, while monitoring zebra/wfm/histo with the unlutted signal. I'm really happy with this approach and C-D has done a bang-up job making this approach workeable. It's beyond my understanding why Sony doesn't do it this way. The F5 is a bit of a beast because the LUTS are sometimes available, not others, the histogram is sometimes there, sometimes not, and the zebras don't go low enough to use. I have to always pause to see what I'm sending to the monitor and what isn't there. The C-D O7Q makes life so so much simpler. I love it as just a monitor and exposure tool.
 
Last edited:
cuervo said:
The C-D O7Q makes life so so much simpler. I love it as just a monitor and exposure tool.
Click to expand...

This is the main reason why i have not sold the FS700+7Q for a FS7 yet.

With the image quality from the FS7Q0 i cant justify FS7 on its own, and an FS7 without the excellent 7Q monitor really isnt feasible either. And FS7+7Q feels like a waste, since i dont shoot raw and have paid for the FS7** raw option.
 
You must log in or register to reply here.
Back
Top