SpeedLooks for Speedgrade, Resolve, Scracth, Lustre, FilmLight

[jimhare]

I'll definitely do that Jake! My biggest problem is I rarely do anything that's under 5 minutes, and usually work on 1-2 hour programs. These tend to be difficult to do anything interesting with in a short time.

But I'm really interested to see what can be done so next I have something short, I'll bring it to you and let's see what happens!

Actually Jim, you'd be surprised how much can be achieved for $500. I do these kind of jobs around the world all the time. Just last week I graded a full commercial in Belgrade, Serbia for close to that amount. Anyone with full version of Resolve can hire me for two or three hours, especially for a commercial type of a job. Very often you'd get better results for your clients, than trying to do it yourself. And let's face it, your time, spent fussing with grading is worth something. Do not hesitate contacting me, if you need help. It may be cheaper, that you think
Your time can better be spent interfacing with your clients, while offering better quality product. And in these strange times, where everyone offers pretty much the same equipment, quality what will separate you from the rest.

 

[jake blackstone]

I'll definitely do that Jake! My biggest problem is I rarely do anything that's under 5 minutes, and usually work on 1-2 hour programs. These tend to be difficult to do anything interesting with in a short time.

But I'm really interested to see what can be done so next I have something short, I'll bring it to you and let's see what happens!

Sure, even if you select a few main shots from the long program, in a short time you can get the "look" set up for that particular project. No generic prebuilt LUT can ever do that.

 

[jimhare]

PMed you Jake.

 

[Jeff Whitehurst]

That's fine if you want to attack me and my work. I don't mind showing my work (that's why it is prominently displayed in my signature) and I realize, that not everyone will like it. That's a part of having an individual taste. And that's exactly why I question this approach of reducing everything to the lowest common denominator of button pushing.
And with that said, I will let everyone decide for themselves whose work they prefer, your's or mine

No attack was made, and I didn't say I don't like your work. It just kind of surprised me when I saw your images feature a lot of crushed blacks and unnatural skin tones, but for some reason that's your main criticism here on this thread. Based upon your comments, you seem to be interested in forcing this to be a conversation about easy button presets and Matrix looks. I'm definitely not making any claims like that. All I'm saying is I might rather start from a nice looking LUT than from LOG, and in some instances this could be helpful. If that's not for you, great. However, making antagonistic comments about how this LUT looks ungraded is about as relevant as somebody complaining that REDGamma doesn't look good just because that checkbox was selected. It's a starting point, that's all.

 

[jake blackstone]

No attack was made, and I didn't say I don't like your work. It just kind of surprised me when I saw your images feature a lot of crushed blacks and unnatural skin tones, but for some reason that's your main criticism here on this thread. Based upon your comments, you seem to be interested in forcing this to be a conversation about easy button presets and Matrix looks. I'm definitely not making any claims like that. All I'm saying is I might rather start from a nice looking LUT than from LOG, and in some instances this could be helpful. If that's not for you, great. However, making antagonistic comments about how this LUT looks ungraded is about as relevant as somebody complaining that REDGamma doesn't look good just because that checkbox was selected. It's a starting point, that's all.

Not sure where do you see ANY crashed blacks in my work (I'd leave that determination based on your monitor. I do hope you use calibrated broadcast monitor). Nevertheless, when you say "I might rather start from a nice looking LUT than from LOG", I have no idea what you're saying... You do understand the difference between the viewing LUT and Output LUT, right?

 

[jimhare]

You do understand the difference between the viewing LUT and Output LUT, right?

With SpeedLooks I believe it's one in the same. You apply it as an output LUT so it is used for both viewing and output. My assumption is it actually outputting REC709 after it does its magic.

You also apply an EPIC input LUT to make RED's RLF and RC3 what it likes as a LOG starting point.

Ideal? Of course not, but makes a nice starting point for those of us that don't know what we're doing! :yesnod:

 

[Jeff Whitehurst]

Once again, this isn't about criticizing your work or you. It's about your snarky comments about our ungraded images and insistence that using the Speedlooks LUT as a starting point is pointless.

Not sure where do you see ANY crashed blacks in my work

Nevertheless, when you say "I might rather start from a nice looking LUT than from LOG", I have no idea what you're saying... You do understand the difference between the viewing LUT and Output LUT, right?

I mean about the same thing as what you said here, except without the FU to the post people from RED part...
http://www.reduser.net/forum/showth...e-one-Graeme&p=1094907&viewfull=1#post1094907

Totally agree. IMO it's another FU to post people by Red. First, you need to spend $5k on RR, which by the way had no reduction in price in over three years, unlike their cameras. And no assistance at all with proper LUT for use with RedLogFilm, other than " just use standard Cineon LUT". Compare this approach with Arri, providing proper LUT at no charge to everyone for any type of software. So, now everyone uses Arri Alexa LUT for RedLogFilm and then everyone wonders why Red skin tones are funky And Red still can't understand why they're disliked by the post community...

Yes, it does look good with Alexa LUT, but not perfect, as Nick already pointed out. One of the things, that gets affected the most, IMHO are the skin tones. I find, that with Alexa LUT I have to work a little extra hard to get the skin tones right.
So, I would love for Red to release a series of LUTs, so we wouldn't even need to discuss it.

I find it absolutely fascinating, that we even need to have this conversation. How much time had passed from the introduction of RFL? A couple of years? And yet, no official response from Red on a proper LUT. WTF!!!??? They introduced RLF, stated, that it's perfectly adheres to Cineon standard and for the rest you're on your own. Come up with your own LUT for all iterations of RLF/RLF2/RLF3 legal and full for ADSK, Resolve, Nucoda, FilmLight, Nuke, AE and on and on and on. So, people in post chain start using third party solutions, so nothing matches between users and no one at Red cares. How difficult would be for Red to release an official LUT for main software platforms, so everyone operates on the same plane? LUT generator would be great too.

And yet, not a pip from Red or Graemme. Let's face it, if you're working with RLF, you're on your own. Nice!

 

[jake blackstone]

Once again, this isn't about criticizing your work or you. It's about your snarky comments about our ungraded images and insistence that using the Speedlooks LUT as a starting point is pointless.


View attachment 33051View attachment 33052



I mean about the same thing as what you said here, except without the FU to the post people from RED part...
http://www.reduser.net/forum/showth...e-one-Graeme&p=1094907&viewfull=1#post1094907

Apparently you're under impression, that every image should have a perfect "Marcie" look, despite what client asks. Promo, that you had used as an example, were run on all major networks for History channel promo of Swamp People. Another version of the commercial, including different, less severe, more traditional look, that you may find more to your liking, was used during this year Super Bowl promo. Just keep studying my web site. It's in there If you have an ax to grind against that look, please direct your ire toward Discovery Communications. I really don't think I need to justify my work to you. Apparently, once you ran out of arguments, you went directly to personal and professional attacks. If you had noticed, I did not...
But hey, every time someone like you does it, it just drives more views to my web site, so thanks

 

[jake blackstone]

With SpeedLooks I believe it's one in the same. You apply it as an output LUT so it is used for both viewing and output. My assumption is it actually outputting REC709 after it does its magic.

You also apply an EPIC input LUT to make RED's Camera RGB what it likes as a LOG starting point.

Ideal? Of course not, but makes a nice starting point for those of us that don't know what we're doing! :yesnod:

The idea is to map the linear RAW signal to a LOG curve, by selecting RLF curve. It is not an EPIC input LUT as you had stated. Viewing LUT is used to view that resulting LOG image as Linear representation. In other words, all grading is performed by using Log image and not Linear image, thus preserving full dynamic range of the signal. Applying Log2Lin conversion at the very beginning defeats the purpose of this process. In this case you're better of by simply selecting RG3 and thus escaping the double translation of Linear debayered image to RLF gamma and back with Log2Lin LUT

 

[jimhare]

You do me a great honor by assuming I would understand any of that! :laugh:

Actually I can sort of follow the logic here. Why interpret and then immediately reinterpret the R3D, right?

In this case I start as interpreting the R3D as RLF and RC3 through the RAW settings in Resolve to get a log-like start, then apply a Speedlook INPUT LUT that LookLabs provided called "EPIC", which reduces the contrast a bit further so it matches the expectations of the output LUT, then apply the actual Speedlook Clean3500 LUT as OUTPUT.

Seems like the only other way would be to know what we were doing without the help of the Speedlooks LUT, which we have already stated we don't.

All we know is by going through that process we can 'purty pictures' we otherwise wouldn't be able to do on our own.

If we are technically double-dipping in the translation pool that's a cost of admission we are willing to pay.

The idea is to map the linear RAW signal to a LOG curve, by selecting RLF curve. It is not an EPIC input LUT as you had stated. Viewing LUT is used to view that resulting LOG image as Linear representation. In other words, all grading is performed by using Log image and not Linear image, thus preserving full dynamic range of the signal. Applying Log2Lin conversion at the very beginning defeats the purpose of this process. In this case you're better of by simply selecting RG3 and thus escaping the double translation of Linear debayered image to RLF gamma and back with Log2Lin LUT

 

[jake blackstone]

You do me a great honor by assuming I would understand any of that! :laugh:

Actually I can sort of follow the logic here. Why interpret and then immediately reinterpret the R3D, right?

In this case I start as interpreting the R3D as RLF and RC3 through the RAW settings in Resolve to get a log-like start, then apply a Speedlook INPUT LUT that LookLabs provided called "EPIC", which reduces the contrast a bit further so it matches the expectations of the output LUT, then apply the actual Speedlook Clean3500 LUT as OUTPUT.

Seems like the only other way would be to know what we were doing without the help of the Speedlooks LUT, which we have already stated we don't.

All we know is by going through that process we can 'purty pictures' we otherwise wouldn't be able to do on our own.

If we are technically double-dipping in the translation pool that's a cost of admission we are willing to pay.

Yes, that's the idea
If SpeedLooks indeed does Log2Lin conversion, then, at least, you should apply it at the end of the grade. You can use a separate node at the very end. Using it at the very beginning does defeats the purpose of using Log images.

 

[Matt Ryan]

Yes, that's the idea
If SpeedLooks indeed does Log2Lin conversion, then, at least, you should apply it at the end of the grade. You can use a separate node at the very end. Using it at the very beginning does defeats the purpose of using Log images.

Hey Jake,

Yeah you are instructed to use the Stock LUT on the very last node (log2Lin). You are supposed to apply the Epic camera patch to RLF in node 1, which apparently gets it to logC to begin. Then they recommend doing your corrections on nodes 2- whatever, with the stock LUT on the final node untouched. Is that the proper workflow? I've been getting great results that way.

 

[jake blackstone]

In this case I start as interpreting the R3D as RLF and RC3 through the RAW settings in Resolve to get a log-like start, then apply a Speedlook INPUT LUT that LookLabs provided called "EPIC", which reduces the contrast a bit further so it matches the expectations of the output LUT, then apply the actual Speedlook Clean3500 LUT as OUTPUT.

Wait, what? There are TWO LUTs, one on the input and one on the output that you need to apply? That is very strange and I have no idea why they are doing that. I understand the need for the Output LUT, but the Input EPIC LUT????

 

[Matt Ryan]

Wait, what? There are TWO LUTs, one on the input and one on the output that you need to apply? That is very strange and I have no idea why they are doing that. I understand the need for the Output LUT, but the Input EPIC LUT????

the input Epic LUT is to convert RLF to logC. They have patches for different cameras to get them all to the logC start point, or at least as close as possible. They have an input Lin2Log lut for 5D footage etc. That way you can get multiple cameras close to matching and close to logC to begin...

 

[jake blackstone]

Hey Jake,

Yeah you are instructed to use the Stock LUT on the very last node (log2Lin). You are supposed to apply the Epic camera patch to RLF in node 1, which apparently gets it to logC to begin. Then they recommend doing your corrections on nodes 2- whatever, with the stock LUT on the final node untouched. Is that the proper workflow? I've been getting great results that way.

So, they are trying to emulate ALEXA look out of EPIC? Because that's what Log-C means. Beats me...

 

[Matt Ryan]

So, they are trying to emulate ALEXA look out of EPIC? Because that's what Log-C means. Beats me...

yeah.

 

[jake blackstone]

the input Epic LUT is to convert RLF to logC. They have patches for different cameras to get them all to the logC start point, or at least as close as possible. They have an input Lin2Log lut for 5D footage etc. That way you can get multiple cameras close to matching and close to logC to begin...

Kinda sounds like ACES workflow, except it's not ACES. All they are trying to do is it to bring everything emulating the look of ALEXA. I hope their emulation LUTs are good and trustworthy...

 

[Matt Ryan]

"SpeedLooks is optimized for the latest digital cinema cameras, however, you can use them on any footage recorded in LOG color space. Included is our specialized camera patch technology, allowing you to customize your look specifically for the camera you shot on. You can optimize for the Alexa, RED, C300, Canon 5D/7D and the Sony F series cameras. Our camera patch technology is an ideal way to give footage from multiple cameras a consistent look. SpeedLooks uses LOG C for our base color space - we call it ‘SpeedLog’. Our camera patch technology automatically converts your camera footage to SpeedLog, which creates a consistent and optimized look across all cameras and footage. If you’re including linear color space footage - shot on any HD camera in any format - into your production, we’ve included a special linear patch to help give you a consistency with LOG images."

http://looklabs.net/index.php?p=looks&id=1

 

[jake blackstone]

Can someone post Lin image of Epic before applying the input EPIC LUT and after. Something like TIFF image would be great. I'm curious to see what does ALEXA emulation in comparison to standard Epic image does it look like.

 

[Matt Ryan]

Can someone post Lin image of Epic before applying the input EPIC LUT and after. Something like TIFF image would be great. I'm curious to see what does ALEXA emulation in comparison to standard Epic image does it look like.

give me a sec.