Hey all -
Performance wise, what does Resolume prefer on something that is larger than 1920x1080 -
- A still PNG Image
- Or, a still image encoded as DXV3 and made into a 5 second length MOV file?
Thanks!
PNG vs DXV3
Re: PNG vs DXV3
I was told a single frame DXV works better.
Re: PNG vs DXV3
I haven't tried png's since I had a really bad experience with a show where performance just stepped off a cliff after a half hour. I mean what started as 60fps degraded to 10fps bad.
PNG's I guess just inflate the allocated ram space for Resolume. So it's a 4gig limit since its a 32bit app. Once it pegs that limit with triggering one PNG after another Resolume slogged badly.
So the work around is render your png into a 1 frame dvx3 clip since Resolume will properly free up its resources like Ram when you move on to the next clip. However I have run across issues with a single frame and now render stills at 10-15 frames and the file size is still reasonable. So that is my recommendation.
I truly hate rendering png's to dvx3 just to deal with this btw. No DJ is going to hand you a Dvx video of their logo. And if you got a lota PNG's for your project it's ugggg.
PNG's I guess just inflate the allocated ram space for Resolume. So it's a 4gig limit since its a 32bit app. Once it pegs that limit with triggering one PNG after another Resolume slogged badly.
So the work around is render your png into a 1 frame dvx3 clip since Resolume will properly free up its resources like Ram when you move on to the next clip. However I have run across issues with a single frame and now render stills at 10-15 frames and the file size is still reasonable. So that is my recommendation.
I truly hate rendering png's to dvx3 just to deal with this btw. No DJ is going to hand you a Dvx video of their logo. And if you got a lota PNG's for your project it's ugggg.
Re: PNG vs DXV3
On versions before Arena 5.1 / Avenue 4.6, any and all PNG files in the deck were loaded into memory. This ensured super fast triggering, because the file didn't need to be "opened" to get the pixels. The pixels were already in memory.
With great power came great responsiblity, and we got lots of crash reports from people that filled up their memory by loading 12 megapixel images straight from their camera. So on Arena 5.1 / Avenue 4.6, we switched to "deferred loading". You can add as many PNGs to a deck as you want, they will only get loaded into memory when they're playing a layer. This ensures that it's very hard to run out of memory.
The downside is that PNGs now take a bit longer when triggered. Especially on larger images, the delay is noticeable. We felt this was a reasonable tradeoff between trigger time and stability. If you want instantaneous triggering, it's best to use a single frame DXV file.
When we're talking about performance in fps, there is zero difference between a PNG and a DXV. Theoretically, the PNG would be slightly faster, because the pixels are in memory, and don't have to come from disk.
If you run into issues where performance drops to 10 fps, it's best to contact us directly with your log file and work with us to solve the issue. Venting your frustration on our forum is perfectly fine, but keep in mind it doesn't help in getting your problems solved.
With great power came great responsiblity, and we got lots of crash reports from people that filled up their memory by loading 12 megapixel images straight from their camera. So on Arena 5.1 / Avenue 4.6, we switched to "deferred loading". You can add as many PNGs to a deck as you want, they will only get loaded into memory when they're playing a layer. This ensures that it's very hard to run out of memory.
The downside is that PNGs now take a bit longer when triggered. Especially on larger images, the delay is noticeable. We felt this was a reasonable tradeoff between trigger time and stability. If you want instantaneous triggering, it's best to use a single frame DXV file.
When we're talking about performance in fps, there is zero difference between a PNG and a DXV. Theoretically, the PNG would be slightly faster, because the pixels are in memory, and don't have to come from disk.
RAM usage does not affect fps performance. Adding a single PNG file with a DJ logo does not affect fps performance. Using single frame DXV files does not affect fps performance.I truly hate rendering png's to dvx3 just to deal with this
If you run into issues where performance drops to 10 fps, it's best to contact us directly with your log file and work with us to solve the issue. Venting your frustration on our forum is perfectly fine, but keep in mind it doesn't help in getting your problems solved.
Re: PNG vs DXV3
To be completely honest... I personally have never converted my png's or jpg's to dxv.... I work festivals with upwards of 15-20 DJ's per night, 4 nights in a row. I have a deck for all of them that don't have their own vj (95%).
I have yet to see ANY performance drop due to the image files (typically 6-10 per artist).
I would have to second what Joris has suggested, send a log file. It might be something else in your comp??
I know I had memory error's pop up in my comp before talking to the guys. It was a simple fix, I had A LOT of memory hog FX active at the same time and I honestly wouldn't have thought to look in that area.
Just from my personal experience, I typically have 2-5 images running simultaneously with auto mask on each one with 1-3 video clips playing behind them (even on my old 2012 MBP) and my fps didn't drop below 50.
Png's work just fine for me, but if you really want it to be optimized. Single frame DXV is where it's at.
I have yet to see ANY performance drop due to the image files (typically 6-10 per artist).
I would have to second what Joris has suggested, send a log file. It might be something else in your comp??
I know I had memory error's pop up in my comp before talking to the guys. It was a simple fix, I had A LOT of memory hog FX active at the same time and I honestly wouldn't have thought to look in that area.
Just from my personal experience, I typically have 2-5 images running simultaneously with auto mask on each one with 1-3 video clips playing behind them (even on my old 2012 MBP) and my fps didn't drop below 50.
Png's work just fine for me, but if you really want it to be optimized. Single frame DXV is where it's at.

Re: PNG vs DXV3
All you have to do is create a DXV3 preset (high with alpha), drop all your PNG files in Adobe Encoder and push the button. The default will be a 5 second clip.I truly hate rendering png's to dvx3 just to deal with this btw. No DJ is going to hand you a Dvx video of their logo. And if you got a lota PNG's for your project it's ugggg.
If you want to make the system even easier, use a Watch Folder in Adobe Media Encoder. The only trick is that a .png file will not trigger the process so you have to change the name to .mov. before you drop it in the watch folder.
Batch renaming can be done in a few second on large amount of files using a batch renamer such as Automator or Name Mangler on a Mac (I don't do that on PCs but there are a lot of options out there).
Re: PNG vs DXV3
I believe I did just that about a year ago. The e-mail chain was quite long and the suggestion / solution at the end of that chain was convert png's to dvx3. I remember talking to Joris about Ram continuing to rise as I went down the deck and fps got worse and worse back then. The issue might have had to do with how Resolume use to handle still images as Joris mentioned. In the end I had to batch render a ton of png's (I can't remember the number) to dvx3 using AE. ><Joris wrote: If you run into issues where performance drops to 10 fps, it's best to contact us directly with your log file and work with us to solve the issue. Venting your frustration on our forum is perfectly fine, but keep in mind it doesn't help in getting your problems solved.
The suggestion did mostly solve the problem but to be fair it was annoying.
It was dozens per artist in the case that melted the computer. To be clear it was a mix of png's and dvx clips set up in large decks. By the 3rd 150 column deck I was noticing major performance loss. Again this was a while back so maybe the way Resolume now handles still images by not keeping them in memory has helped sort of but not really. Even then how the pixels were stored in memory it was crushing the moment it was triggered.dinga wrote: I have yet to see ANY performance drop due to the image files (typically 6-10 per artist).
The computer was using 2 video cards which I thought I shared in that chain. But we all know that doesn't help performance either. But it shouldn't make things worse over time.