Still seeing high end rolloff in Engine 1.3.1

Hello, I’ve been holding off the purchase of any Prime equipment until the high end rolloff issue was fixed. I was excited today to hear that 1.3.1 released which lists the high end response as an improvement. When I got home I downloaded the update to run my high end rolloff tests which consists of simply running white noise through engine, recording the output, and plotting the spectrum.

I use a windows computer and audacity to do the recordings. I found, even with engine 1.3.1 installed, I still get the high end rolloff happening.

As this was the reason preventing me from buying any prime hardware I, unfortunately, can’t check on the hardware itself. If anyone else can with the latest firmware I would like to see the results!

The high end roll off has definately gone from the players.

Also re-Reading the updates, it is ONLY the players which have has their high end roll off removed. The update details were very clear on that.

As Engine Prime is preparation software not performance software at a gig, the roll off on engine Prime isn’t as critical as on the 5000 and 5000m etc

I recognize this is a big ask but I figure I’ll ask anyways. Would you or anyone else be willing to provide a white noise sepctrum plot recorded from the players? If that’s too much even an audio recording of white noise from the players themselves would be fine.

I recognize this is a lot of work to ask a stranger on the internet to do so I understand if you simply don’t have the time :slight_smile:

This was in the update to the players, not the Engine management software.

1 Like

High-end roll-off is still there on the players compared to original test file, but less than before. Audibly better anyway.

If I had to guess, I’d say it’s because the apparently-shared audio processing… uhhmmm… engine is probably due for a major rewrite. The tweaking of what I assume were the players’ SRC filters was likely a short-term stop-gap, even though it netted us an additional 10dB on 44.1khz files and about 16dB on 96khz files at 20khz. This is not Elastique’s FS setting, though you do get very slightly more highs when you turn keylock off.

The IMD+noise in RMAA and weird harmonics (varying 3.5-order-ish at about 30dB down in succession?) with single-frequency tones and the impulse response’s lovely lack of ringing are probably from non-windowed realtime FIR filters in the core engine. With IIR filters you gotta supposedly choose from gradual with its lack of ringing plus very low distortion but getting too much roll-off early, not enough roll-off late, and/or aliasing coming through or you choose steep that extends your highs plus prevents aliasing but causes nasty impulse ringing. FIR filters lack the nasty ringing and are still steep enough to prevent most aliasing, but to suppress the IMD, signal-embedded-noise, and harmonics we’re measuring now would apparently cause a delay from the combination of feedback and feedforwarding with windowing, so that’s out.

But exactly why is it seemingly being constrained as if everything’s being downsampled lossily? And why are the impulse responses modulating in amplitude over time? If I’m wrong and it doesn’t get overhauled or it otherwise must be constrained this way, then maybe they can do like on audiophile kit and just give us a couple other filter options under Utility in addition to the present FIR one and we can choose our poison, and maybe they can even make the nonlinear & odd-order-ish harmonics on the FIR one instead more linear & even-order to sound more tubey and less like a guitar pedal if they can’t be suppressed further. I presume the audio engine has other issues than just the filter tradeoffs, though, and that we ought not to have to be even making such tradeoffs. But I could be wrong.

I don’t actually think it’s that unlistenable on very dynamic and/or minimalist recordings/productions, and while it’s always got a bit of processing signature, it doesn’t start to seem smeared, congested, transiently-blunted, or crunchy or muddy (depending on the content) in an overt way unless you’re playing really dense music. Who else here sometimes mashups shoegazer? Anyone, anyone? Guess I’m the only one :slight_smile:

1 Like

I should have recorded last night, but the players and x1800 do sound world’s better. I tested an original via interface to PA, X1800 Interface to PA, 1 SC5000 without update, 1 SC5000 with update. I am one happy panda.

Are you liking the big input pad on the X1800? Puts a 0dBFS test signal at -15.5dBFS which is showing at about 4dBVu-ish on the X1800. Gives us a nice healthy automatic headroom from the brickwall even without touching the gain/trims. I think last night I had DM’s Black Celebration at about 2 o’clock on the knob being conservative with the metering and it seemed appropriate-feeling for the work flow with those dynamics. There’s still enough boost left that we can crank the gain/trims during quiet parts for effect, too. I haven’t tried vinyl yet. Hopefully they can get us an MP2015 firmware update one of these days :slight_smile:

I didn’t have enough time to test headroom. Only audibly what I heard while mixing. The mixer already had loads of headroom as is. I’ll dig deeper tonight when I get home.

Headroom of the mixer can’t be changed. They just changed where we’re running the gain/trims. Didn’t you want that? I should mention I was talking knob at 12 o’clock with the test signal. The Rane needs it more, but one thing at a time.