Robert Ruffo
Forum Replies Created
-
Shortcuts have changed from previous version – maybe an issue for you.
-
[Peter Chamberlain] “We use the RED SDK, as everyone does, for r3d files. Other RAW images are debayered in GPU, then re-sized to the timeline res… where the image processing, inc temporal NR and speed changes, occurs.
So the GPU ram size and core count is important in many respects.
Peter”OK, so… who much memory do I need to use temporal noise red. on Red 5K for a 1080p timeline? In other words, when it comes to GPU RAM requirements, is it the resolution of the timeline that is important for Red footage, or the resolution of the source footage?
BTW – I must say the new curves in latest build are great – much less noise than ever.
-
True 8 Bit is fine, the problem is an Apple cinema is less than 8 bit and does not display all colors in the 8 bit palette.
Another problem is if you are using a LUT to correct your apple display, in 8 bit, there is your multiplier right there.
-
Despite quality issues and feel, the Wave is actually a faster/better panel in my opinion for Resolve than the beautiful Element, only because of how BM has mapped it
-
I always thought memory requirements were tied to timeline resolution, not source footage resolution. Is this not the case for the new NR? If it is not we can all forget using it with Dragon footage. Who about optical time-warps? DO they look at the source footage resolution, or the timeline?
-
You are confusing looking “good” with really being accurate. All plasma displays have serious problems – missing colors in the darks make false contours, and ABL is another biggie.
Extremely irregular color tracking is also hard to fix, even with a LUT
-
Like I said – at Company 3 and so on there is another guy/girl to make the mattes, if need be, on another system entirely. Those situations are ultra-high-end. In narrow deadline mid-budget commercials the colorist does not have access to external resources like that.
That said, high-end post is slowly moving away, not toward Resolve. It is low end to mid range house who are newly adopting Resolve as a replacement for Apple Color or the tools within Premiere, etc. I have not heard of anyone dumping Mystika or Baselight to switch to Resolve. I don;t think that has happened in many years. This is a sign they are falling behind, that their only continuing advantage is price, and from feedback I hear (which BM does not seem interested in, but whatever) the secondary qualification and alpha tools are the weak point, the tracker the strong point.
I did not say Resolve was prosumer, but simply that their new feature focus had that orientation, and it did not address urgent problems.
I would much rather not have to buy a new system and spend the time learning it. There are also many things I do love about Resolve and the new features in 10. I don’t want to have to say goodbye to them just because they can’t fix something so basic which would not be hard to fix at all. How hard would it be to split out an alpha channel with the combiner node function? Then I could apply an OFX depspekle plugin or heavy denoising and most problems solved right there. It’s almost infuriating that they don’t do this.
That’s why I hope that my nagging will pay off. Without customer feedback software companies are just throwing darts in the dark, although BM does not seem to appreciate them (at their eventual peril – not my problem after we switch).
-
We don’t own Mistika yet, so I can’t do that on that platform, but I can do better – I’ll show you a secondary pulled with prosumer Colorista which came out way cleaner than the same shot in Davinci. Colorista is a generally a huge pain in the ass, but it pulls secondaries very easily an cleanly. Give me a few days.
Mistika secondaries are much more powerful than Colorista’s and I brought my own footage (which I had worked on perviously in Resolve) to try it out with. Granted, it was not a side-by side situation, but I did remember how long it took in Resolve – it was not 30 seconds like in Mistika and it never looked pristine like that.
The Mistika alpha finessing tools are not ground breaking or anything – similar tools (not real-time, but still) have been around in After Effects, Combustion, etc. for years and years – but in a grader in real-time, assigned to a control surface, using them is a dream.
As for the price of Resolve – well Photoshop is even less, and it’s extremely sophisticated and functional. Any tool can pay for itself quite quickly if you have the clients and give them good reason to come back – what I think is most important is the value of my time, and my clients’ time.
-
Rohit, I posted again because quite frankly I was hoping things would improve in version 10. They did not.
Instead we have tons of other features which mean little to the way we work. There is no way we are about to start editing on Resolve – and why would we when we own and know many other NLEs already.I must say your somewhat curt response, also, does not impress me as a paying customer of several thousands of dollars of your products, both software and hardware. Last time you replied that Hollywood films used that HSL qualifier – I since found out that the films you mentioned did not – alpha channels for secondaries were assigned to a matte creator at another desk and then loaded in. Higher budget shows also get the look much closer in camera than some of my clients, and expect less from the color grade
Best way to stop someone from making a legitimate complaint is fix the problem, not tell them to shut up.
At least that’s how we deal with complaints from our customers, which is in fact why we are solving their complaint of slowness and noisiness of secondaries in our grading sessions by moving over to Mistika. SGO is famous for really listening to users and mining them for good improvement ideas, not flipping them off as you just did, and it shows in the elegant interface of their product.
I’m not asking for a new raytracing engine to be added to Resolve, I’m asking for simple alpha finessing tools – things like despekcle have public-domain software routines. Blur-choke is trivial, as is splitting out the alpha channel in the split/combiner node so that one could apply blur, a curve or a third party plugin to the alpha and thus have allow us more options and tricks for fixing the HSL qualifier’s inherent shortcomings.
It’s amazing you have not felt this needed fixing, as I am not the only one who feels that color-qualification is a bit old-school in Resolve.
-
Same issues here.
i also just spent some time with Mistika.
I must say, their tracker is almost as good, and everything else is far better, including a much, much better secondary keyer.