|
|
Post by foggy1956 on Jun 7, 2015 10:19:25 GMT -5
I can appreciate having manual a/v lipsync specifically in a video processor [ like my old dvdo duo with per input adj ] but even this was hit and miss with trying to line up the frames with the audio . Not up to igor's standards methinks  The nice thing about the xmc1 is theres no video processing so hdmi protocols can do whats needed ; if the other component follows its edid properly of course  Wondering if your running multiple frame rates foggy and some video processing is upscaling/delaying etc ? Set everything to 480i/60 as an experiment maybe or 1080i say to utilise the xmc1s bandwidth capability ; whatever suits.. Will do some checking, have noticed mostly when running has through jriver. Thanks for the tip |
|
|
|
Post by igorzep on Jun 7, 2015 17:15:09 GMT -5
I'm curious as to what real world issues the OP is having with the XMC-1, not with the design, implementation and lack of documentation but how it actually sounds. A lot of quite annoying clicks/thumps/sound skips here and there. The problem with perfect digital silence in the audio stream is reported to them with a good reproducible test case and they are investigating... There must be something, other than wanting to verify everything is "perfect," that he noticed while listening that is causing his dissatisfaction and desire to re-design Dirac and the XMC-1. I think Keith has given a very detailed explanation of what can and cannot be done using Dirac with the XMC-1. While the OP might not like it, it is what it is. Of course suggestions can be made, but it sounds like what is desired is beyond the scope of the Dirac/XMC-1 collaboration. The thing is that crossovers do affect sound very much. The difference in sound between properly integrated crossover and just random cut at desired frequency is pretty large (and I've heard how it can totally transform how the system sounds). When you buy a product that incorporate both - Room Correction and Bass Management - you expect they are thought out to work together and optimized to properly integrate with each other. This is not the case now and require quite a lot of understanding/skills and extraordinary adroitness with curve editor do make this integration happen. This is not quite what you expect from an 'automatic' tool. Again - the difference to make it work 'automatically' without extraordinary adroitness of a user is one biquad - not whole re-design of Dirac and XMC-1 and doesn't even need Dirac involvement. Another thing is the promoting impulse-response/timing correction which is in fact - nothing else but correcting effects to timing introduced by crossovers. The fact is, that the tweeter-to-mid crossover is the most noticeable thing if you look to the impulse response (the reason raw impulse response graphs are looking so impressive with Dirac), but also least noticeable in SQ. On the other side the same phase/timing distortion introduced by low frequency crossover is more noticeable (as it introduces more group delay) and the one crossing over the sub is the most audible, but also completely invisible on a raw impulse response graph (as it is totally dominated by high frequencies). So, while the claims of excess-phase correction are perfectly true for Dirac as a stand-alone product, you don't get the most important part of it on XMC-1 as the reason, causing the most important distortion is simply unknown to Dirac. So those claims/talks are pretty much just marketing tricks in relation with XMC-1. Passing the info about desirable crossovers to the Dirac would automatically solve both issues. This is what I would prefer to have, but this requires work on both sides - XMC-1 of which probably most complicated part is agreeing on how to transfer data from XMC-1 to Dirac... Or, alternatively it can be done completely on Dirac side by allowing the user to manually enter the desired crossover properties that will later be used with XMC-1. The math behind that then is then also trivial, and nowhere comparable to the complexity of the Dirac algorithms and in no way a complete re-design of Dirac. |
|
