• The Ai software is a powerful media server that delivers fast, intuitive, easy control featuring market leading media playback quality, a fully customisable node-based system (Salvation Engine) suited to the requirements of your individual performance. Ai Media Servers facilitate the future of live entertainment, lighting and video and are designed for professional show designers, content producers and artists, enabling seamless delivery of multi-media video projects at any scale or venue.

• Ai is a QuickTime capable application, which means that most files which play in QT “should” play within Ai. However, please note that Interframe codecs (such as h264, mpeg2,4 etc) and other CPU based codecs will significantly reduce playback performance. For this reason, we always recommend using the AiM codec as this will give the very best results for your show.  Please note that our support engineers will always insist on the use of the AiM codec as this is the only codec we use internally for testing.
For Further information on the AiM codec please take a look at this detailed document – AiM Codec Analysis.

• The smooth playback is down to our revolutionary AiM codec (a cross platform Quicktime codec) and the use of the latest GPU technology bred for the gaming industry, but now re-developed with the support of AMD and Nvidia exclusively for Ai. This allows for much better playback performance and can easily handle media of 4k resolution and above. To encode your media using the AiM codec, first download it from (http://www.avolites.com/ai/AIM_Codec.zip). To install the codec: OSX – To install the AiM codec on OSX copy ‘AiMCodec.component’ to /Library/Quicktime Windows – To install the AiM codec on Windows XP / 7 copy AiMWin32.qtx to C:Program FilesQuickTimeQTSystem (on 32bit systems) or C:Program Files(x86)QuickTimeQTSystem (on 64bit systems)
  For Further information on the AiM codec please take a look at this detailed document - AiM Codec Analysis

• It is usually recommend that media is created with a width which is divisible by 16. The AiM codec is much happier that way. It might work fine on many systems with a media width which is not divisible by 16 but it would not be recommended; because on some systems we have seen problems when the media is created at a width which is not divisible by 16. I would suggest adding a couple of pixels into the void areas between the screens to pad the width out to a number which is divisible by 16, or perhaps remove a few pixels from the void areas, whichever has least impact. You’re also welcome to ignore this advice if you’re able to test every system involved in the playback systems and content creation thoroughly – its a GPU / OS dependent issue.

• To encode your mediawith the AiM codec, you need to use video software, such as Final Cut Pro,Premier, After Effects, etc. Within your software, load your movie(s), and thenchoose to export/render using the AiM codec. There are three choices when using this codec, Transparency, Optimise For and Threads.
• Transparency allows you to choose wether to discard or preserve the Alpha channel in your clips.
• Optimise For has 3 choices - Quality which gives you a full 24bit render, Performance which is a 16bit render that is equivalent to the DXV codec, and Superstream which is a 24 bit variation with more frame information and results in slightly larger files but also gives the best playback performance. 
• The threads option allows you to determine how many processor threads are used when encoding - an incorrect setting can result in renders not being completed, and therefore we would recommend choosing the max option which will automatically determine how many threads you have available to use.

• Ai is very flexible with screen resolutions – they can be user defined up to the maximum limit of your software licence level;

• Anjuna license has maximum canvas resolution of 2048 x 2048.

• Bondi license has maximum canvas resolution of 4096 x 4096.
  

• Miami license has no maximum canvas resolution limit.
  

• Ai makes use of FreeFrameGL effects – an open source effects standard freely available from many sources across the internet. They can be added to Ai by installing them to DistribModulesEffectsFreeFrameGL.

• Ai needs to be run as an administrator. Right click the program shortcut, choose properties and then in the compatibility tab, tick the box labelled “run this program as an administrator”.

• This generally happens when Ai is run on a system which does not meet the required specification for the graphics card – often this will be due to an Intel graphics chip with shared memory. unfortunately these are not supported in Ai as they lack some of the higher functions required of the GPU that Ai makes use of. Please refer to the minimum required specs for further information.

• Live video can be dropped into the roller que to select as a source by selecting one of the patches named ‘LiveVideo’ in DistribPatches. You can select additional patches from the same patches folder for SDI inputs which are numbered accordingly

• OS is 64 bit, software is 32 bit at the moment, there are no immediate plans to convert to 64 bit, the software is fast enough as it is. We may decide to do this conversion in the future if it seems necessary.

• The system can play back DXV movies, but we have our own new codec (called AiM) which is similar in some respects to DXV but it has some significant quality improvements. Colour reproduction is much better than DXV which suffers with banding in smooth gradients due to the format using 16 bit RGB colour where our codec uses 24 bit RGB. The trade off is that files are 30% bigger than DXV, although there are 2 options in the quicktime settings one for ‘optimise for performance’ which produces similar results to DXV with some minor colour improvements but with matching file sizes and a 2nd option to optimise for quality which gives the 24 bit RGB colour and 30% larger file sizes.

• Yes

• Yes

• The default setting present in Bootcamp for scaling may need to be changed.  With Ai closed right click on the Ai shortcut on the desktop, select ‘properties’ and then the ‘compatibility’ tab.  If ‘disable display scaling on high DPI settings’ is checked then uncheck it.  Then restart the system.

This can happen when a virtual surround sound processing program called Nahimic is installed, most commonly on gaming motherboards.
Uninstalling this program and then rebooting the machine should resolve this issue and allow Ai to run.

• Yes, Ai can be controlled using industry standard MIDI, DMX or ArtNet controllers or from our own range of dedicated control surfaces. We recommend using our Titan range of consoles for excellent integration and functionality with the Ai software.

• Ai integrates well with Avolites lighting consoles, however it is not locked solely to them and can be controlled by whichever desk you are most familiar with via the ArtNet output.

• Occasionally, the screen outputs can come up in the wrong order when connected to the system. This can easily be rectified by first going into the Windows Display Settings and reordering your screens in the correct order, e.g. 1, 2, 3, 4, etc. It is then a good idea to go into your graphics card’s display settings and also changing the order here as well. In most cases this will rectify the issue. It is important to ensure you have the screens connected in the right order from the graphics card. In the R-range of media servers, the top DisplayPort connector is Output 1.

Windows operating system

Intel Quad-Core CPU.
Dedicated graphics card from AMD or NVIDIA (on board Intel graphics will not work).

Storage
7200 rpm hard drive:    2x layers of 1080p.
1 SATA SSD:                           3x layers of 1080p.
2 SATA SSD RAID 0:              up to 6x 1080p or 1x 4k.
4 SATA SSD RAID 0:              up to 8x 1080p or 2x 4k.

NVMe SSD:                             8x 1080p or 2x 4k

Capture Cards

Datapath Video Capture Cards are recommended and tested to work in Ai

It is worth noting that video playback software is very demanding of the graphics card. As a good reference point for the capabilities of your hardware you may find PassMask Software's video card benchmark pages for high end GPU's useful. A PassMark G3D mark of 5000 or more could be considered a general minimum performance requirement for Ai, but in practice the actual requirements will vary a lot depending on how you use Ai.

 

• We have used the quadro series in the past and they are very good cards. Performance is not a lot better than standard Nvidia Gaming cards though in our application. What they do excel at though is long term reliability. These cards will last much longer than the gaming cards which can burn out after a year or two of continuous usage.

We don’t use CUDA at the moment. We do use a lot of GPU based processing but it is mostly done with shaders, and DXT texture compression.

• ArtNet, MIDI, and Ethernet

As of version 7, Ai has the ability to create complex screen layouts using our model merge feature and vary how the texture is applied to them using our unique canvas editor . For previous versions of Ai, it is best to use an external modelling application.

The safest formats to use are 3ds Max ( .3ds ) or Wavefront ( .obj ) files. We also support Collada ( .dae ), Blender 3D ( .blend ), Stanford Polygon Library ( .ply ), LightWave ( .lwo )

Currently Ai can play back stereo audio in WAV format. Also, if your video clips contain audio, we recommend re-encoding them with the AiM codec. The best settings are 16-bit stereo at 44.1kHz uncompressed. If encoding in Quicktime it is necessary to untick the ‘Big Endian’ option. Files rendered with different rates may not perform as expected. We are looking into the permissibility of incorporating other audio formats.

For the best results, we recommend using our own AiM codec with audio embedded. Alternatively, WAV files can be added to the timeline along with videos and then the result exported as a single timeline cue.