Background:
I did some quick digital-path testing on my AltitudeCI because I wanted to check the part of the unit that I use in playback: HDMI input to AES output. I am not using the AltitudeCI analog outputs, so I was not trying to test the DAC section or analog performance. I also don't feel like trying to sweep them as there's not a lot of space behind my front rack...
The test was inspired by the earlier Altitude32 ASR review, but this is not a direct duplicate of that test setup as I ran into REW limitations and didn't want to spend a lot of time, once I saw the outstanding results I was getting with room correction and a "Strong" target curve enabled, versus without.
The good news for the few tests I did, I do not see any obvious clipping, side bands, or strange HDMI behavior visible at the AES outputs that is of relevance.
Signal Path for Test:
JRiver Media Center playing playing wave test files → NVIDIA RTX 2080 Ti HDMI output → Trinnov AltitudeCI HDMI input → AltitudeCI AES output → RME HDSPe AES input → REW RTA capture
The RME HDSPe AES is best accessed through ASIO, while the Trinnov HDMI output from the PC is exposed through the normal Windows/Java device path in REW. REW’s driver mode is global, so I could not simply select Java for HDMI output and ASIO for RME input in the same REW instance. The workaround was to generate 96 k/24-bit -3dBFS WAV test files, play them in JRiver through my normal Trinnov HDMI zone, and capture the AltitudeCI AES output with the RME/REW analyzer path.
I used -3 dBFS test files for the main flat/no-correction checks. Captured levels matched the file levels, so the HDMI-to-AES path appeared to be at unity (for the flat measurements with room optimization turned off anyways).
I would not interpret the small far-out spurs as audible or practically significant.They are on the order of -175 to -180 dBFS, and the 2M FFT setting is intentionally revealing. A lower FFT length would likely hide them. They also appeared at 19kHz even if I perform a loop back test on the RME sound card by itself in REW, with the Trinnov totally removed from the signal chain. Unfortunately, I can't measure any better to remove them, as I have already turned on dithering which helped get rid of some... However, this is plenty good enough for me, and I can't swing Audio Precision gear plus have the Trinnov AltitudeCI
Bottom line:
In my setup, the AltitudeCI HDMI-in to AES-out path looks extremely clean at 96 kHz, even with my usual strong target curve listening preset plus optimization enabled... The very low spurs I have are reproducible, but are from my RME HDSPe AES card (as I did loopback measurements to prove this). I'm really happy to finally find a processor that does what I want, and to have decent HDMI that doesn't degrade audio quality (my last processor was in 2016).
The CI is highly recommended, but is not cheap at all. However, I figured, heck the room has 16 Benchmark AHB2 amps, so why not add a Trinnov processor too. There are actually very few options that could provide 32 AES input and output channels btw, with no down-sampling either (only tested up to 192kHz, and can't test beyond that). Of course, most movies are still mastered in 48kHz, but it's great to have an Atmos/DTS-X decoder that can work with my pro-audio based setup so well!
I will say that before when I tried room correction software, I quickly removed it as I found it to add slight artifacts. Not the case with the Trinnov, all I hear is improvement. It is one of the first processors I've had that didn't try to set my subs to some ridiculous value like 30 feet either.
In fact, the CI left most of my custom settings I had previously put into each Xilica XD4080, using an Earthworks M30 mic, Benchmark ADC1 and MPA1. And the 3d microphone is highly sensitive. Even if you rotate it as small as humanly possible, the slight rotation still shows up on the polar plot inside the CI's GUI that can map your speaker locations three dimensionally. It's that sensitive, and if anything, all the level and time delay it did improved from what I had meticulously done by hand.
The AtitudeCI I'm using is actually only 10 channels for now, and eventually if I can ever afford the channel upgrades necessary, I'd like to drive the remaining two height speakers, tri-amp the main bed and hook the subs individually to the Trinnov AltitudeCI. Right now, I'm still using 6 Xilica XD4080 Speaker Processors (4in, 8out) that are fed AES3. They act as the DSP for crossing to the 8 subs, tri-amping the 5 bed channels with FIR filters, all time alignment, all level alignment, etc...
Hopefully I can one-day remove the excellent Xilica XD4080's and add more Benchmark DAC3 B's connected directly to the Trinnov Altitude CI. This would actually be a lot better for headroom management, plus the DAC3 B analog outputs provide a few additional dB of headroom over the XD4080 analog outputs. I've been slowly adding the DAC3 B's as I find them used, since the Xilica XD4080's have digital inputs and outputs as well.
Flat results with no correction:
-180dBFS WAV File Test (Flat target curve, no optimization turned on):
-3dBFS WAV File Tests (Flat target curve, no optimization turned on):
Strong Target Curve Plus Room Optimization:
Notes for these tests with the "Strong" target curve and room optimization turned on: I had to set the Altitude's volume down -8dB just to prevent any clipping due to the room optimization adding a few dBs, plus the fact that the target curve is a little aggressive:
That's why I labeled these measurements as -11dBFS, while they are the same -3dBFS WAV files as before...
Some pics of the setup:
I did some quick digital-path testing on my AltitudeCI because I wanted to check the part of the unit that I use in playback: HDMI input to AES output. I am not using the AltitudeCI analog outputs, so I was not trying to test the DAC section or analog performance. I also don't feel like trying to sweep them as there's not a lot of space behind my front rack...
The test was inspired by the earlier Altitude32 ASR review, but this is not a direct duplicate of that test setup as I ran into REW limitations and didn't want to spend a lot of time, once I saw the outstanding results I was getting with room correction and a "Strong" target curve enabled, versus without.
The good news for the few tests I did, I do not see any obvious clipping, side bands, or strange HDMI behavior visible at the AES outputs that is of relevance.
Signal Path for Test:
JRiver Media Center playing playing wave test files → NVIDIA RTX 2080 Ti HDMI output → Trinnov AltitudeCI HDMI input → AltitudeCI AES output → RME HDSPe AES input → REW RTA capture
The RME HDSPe AES is best accessed through ASIO, while the Trinnov HDMI output from the PC is exposed through the normal Windows/Java device path in REW. REW’s driver mode is global, so I could not simply select Java for HDMI output and ASIO for RME input in the same REW instance. The workaround was to generate 96 k/24-bit -3dBFS WAV test files, play them in JRiver through my normal Trinnov HDMI zone, and capture the AltitudeCI AES output with the RME/REW analyzer path.
I used -3 dBFS test files for the main flat/no-correction checks. Captured levels matched the file levels, so the HDMI-to-AES path appeared to be at unity (for the flat measurements with room optimization turned off anyways).
I would not interpret the small far-out spurs as audible or practically significant.They are on the order of -175 to -180 dBFS, and the 2M FFT setting is intentionally revealing. A lower FFT length would likely hide them. They also appeared at 19kHz even if I perform a loop back test on the RME sound card by itself in REW, with the Trinnov totally removed from the signal chain. Unfortunately, I can't measure any better to remove them, as I have already turned on dithering which helped get rid of some... However, this is plenty good enough for me, and I can't swing Audio Precision gear plus have the Trinnov AltitudeCI
Bottom line:
In my setup, the AltitudeCI HDMI-in to AES-out path looks extremely clean at 96 kHz, even with my usual strong target curve listening preset plus optimization enabled... The very low spurs I have are reproducible, but are from my RME HDSPe AES card (as I did loopback measurements to prove this). I'm really happy to finally find a processor that does what I want, and to have decent HDMI that doesn't degrade audio quality (my last processor was in 2016).
The CI is highly recommended, but is not cheap at all. However, I figured, heck the room has 16 Benchmark AHB2 amps, so why not add a Trinnov processor too. There are actually very few options that could provide 32 AES input and output channels btw, with no down-sampling either (only tested up to 192kHz, and can't test beyond that). Of course, most movies are still mastered in 48kHz, but it's great to have an Atmos/DTS-X decoder that can work with my pro-audio based setup so well!
I will say that before when I tried room correction software, I quickly removed it as I found it to add slight artifacts. Not the case with the Trinnov, all I hear is improvement. It is one of the first processors I've had that didn't try to set my subs to some ridiculous value like 30 feet either.
In fact, the CI left most of my custom settings I had previously put into each Xilica XD4080, using an Earthworks M30 mic, Benchmark ADC1 and MPA1. And the 3d microphone is highly sensitive. Even if you rotate it as small as humanly possible, the slight rotation still shows up on the polar plot inside the CI's GUI that can map your speaker locations three dimensionally. It's that sensitive, and if anything, all the level and time delay it did improved from what I had meticulously done by hand.
The AtitudeCI I'm using is actually only 10 channels for now, and eventually if I can ever afford the channel upgrades necessary, I'd like to drive the remaining two height speakers, tri-amp the main bed and hook the subs individually to the Trinnov AltitudeCI. Right now, I'm still using 6 Xilica XD4080 Speaker Processors (4in, 8out) that are fed AES3. They act as the DSP for crossing to the 8 subs, tri-amping the 5 bed channels with FIR filters, all time alignment, all level alignment, etc...
Hopefully I can one-day remove the excellent Xilica XD4080's and add more Benchmark DAC3 B's connected directly to the Trinnov Altitude CI. This would actually be a lot better for headroom management, plus the DAC3 B analog outputs provide a few additional dB of headroom over the XD4080 analog outputs. I've been slowly adding the DAC3 B's as I find them used, since the Xilica XD4080's have digital inputs and outputs as well.
Flat results with no correction:
-180dBFS WAV File Test (Flat target curve, no optimization turned on):
-3dBFS WAV File Tests (Flat target curve, no optimization turned on):
Strong Target Curve Plus Room Optimization:
Notes for these tests with the "Strong" target curve and room optimization turned on: I had to set the Altitude's volume down -8dB just to prevent any clipping due to the room optimization adding a few dBs, plus the fact that the target curve is a little aggressive:
That's why I labeled these measurements as -11dBFS, while they are the same -3dBFS WAV files as before...
Some pics of the setup:
Last edited:
