Nuendo / SX Benchmarks :.

Blofelds DSP 40 - The Infamous Save / ReOpen Performance Droop :

It was evident right from the outset that we were seeing large variables directly related to the specific audio cards being employed.

2 distinct variables were noted , one being the number of plugins that could be loaded simply running the session respective to the audio hardware, and the second being the variable of the number of plugins that could be succesfully loaded after saving and re opening the session.

The second being the most controversial, as it seemed that nearly all audio hardware suffered from the experience , except Lynx audio cards.


Of course as the project progressed, and the collated and expanding data base continued to confirm more and more reports of the anomaly, the mystery deepened.

Without any evidence outside of the specific audio hardware, select sectors began questioning the validity of the test , and claimed it had to be a flaw in the test session itself.

I had already answered all queries in relation to the session being the possible cause , but that unfortunately did not appease all, and the resulting fallout still continues to this day.


However, over at the Sound On Sound Forum, One member named PhatRiff , decided to pull up his sleeves and develop a similar styled test that he could run across multiple DAW applications, using a 3rd party plugin, in doing so negating any other variables except the Audio Interface and the specific PlugIn being used.

His results showed that the now infamous Performance Droop accurred on 4 of the 5 Applications tested , proving that the issue was not reserved to the Blofelds DSP session, or its respective PlugIns, and could be repeated across multiple DAW applications, using a completely different set of PlugIns.

Performance Droop Across Multiple DAW Applications:    

The above data was collated from the information posted at the Sound on Sound Forum by PhatRiff , who developed and conducted the testing.

Thread and ongoing discussion can be found here

Five DAW applications were used, and the Waves IR PlugIn was selected as the reference PlugIn to be used across all the Applications. A RME Fireface was used in all testing @ a buffer setting of 64 Samples.

This is one of the units that displayed the anomaly in the Blofelds DSP40 session, so it is a good cross reference.

Multi Processing was also disabled to give a more even result, as several of the Applications are not Multi - Threaded.


Whats interesting is that the results were very consistant across 4 of the 5 Applications , With the maximum plugins when first loading the sessions being between 28-30, and the ReOpen results all being identical.

The One application that behaved differently, reporting the same result of PlugIns both Running the Session and after Save/ReOpen, was SAW.

Phatriff reported that on all of the other Applications, the maximum result was acheived after pushing thru some earlier minor crackling and drop outs in some instances, whereas SAW consistantly stopped playback as soon as any drop out was sensed.

That would explain the lower results on initial playback.


SAW's behavior definately highlights distinct differances in the way that the specific Audio engine is being employed , whether that behaviour is the most prefered would be debatable, but it certainly takes the guesswork at of where the actual drop outs start to occur.

It was also reported that the Performance Droop could be experienced by simply changing the buffer/latency setting back and forth and or resetting the audio driver, without needing to Save and ReOpen the Session.

The actual cause of the Performance Droop is still being investigated, with Phatriff developing a second more detailed and expanded test session.

Details are on the original SOS thread that was listed earlier on.

Page 1 | 2 | 3 | 4 |5 |6 |7|8 |9 |10 |11 |

© AAVIMT 2006