Line 6 Xd-v75 Review

[Rafael Nowning]

Ihave read and heard mixed reviews regarding the reliability of the series, more particularly with regard to the chance of dropout occurring. My application will be musical theatre, and there are likely to be human bodies surrounding the transmitters on stage, blocking line of sight to the receivers. In particular, the actors will conceal the transmitters about their person, possibly in pockets (which the V75 manual says to avoid). I am very nervous about experiencing dropout during a show, which I have rarely encountered with the licensed UHF mics I regularly use.

Your concerns about line of sight are very valid. I would never recommend running any radio receivers with built-in antennae at table top height - human bodies are very good at blocking radio signals, particularly as the frequencies get higher. If you get your antennae high up in the air with a clear "view" of the stage, they'll be much more reliable. I use the P180 paddles up above head height which are great for this but may not fit in your budget. The antennae loop through outputs on the V75 are also useful as you can daisy chain several receivers from one pair of antennae. Having a cluster of several receiving antennae very close to each other is bad practice so best avoided.

Thank you very much for your insight, Sheriton. My existing antennas are always above audience head height, about 100 feet from the stage, as would the V75s be. I was more concerned about the beltpacks' antennas being smothered by actors' bodies, but if you're using them for the same application as me then I assume that this should be fine?

Bear in mind that you'll have hundreds of mobile phones between stage and receivers, all transmitting wifi and bluetooth in the same band as your L6 mics. I would put the recievers side of stage if you possibly can. They'll almost certainly work fine out front in an empty house but once the audience is in, it may be a different story and there's no way of testing them in those conditions without an audience.

I've seen signal strength vary a bit if an actor is tightly surrounded by others but nothing show stopping. Preventing the TX antenna contacting bare skin is, I believe, a good idea although costume practicalities can get in the way of that. The L6 TX antenna is just a bare wire inside that dome so it's perhaps less of an issue than those packs that utilise an external flexible wire antenna.

1. I would always recommend RF1 over RF2. It's more robust; you won't notice the difference in a quiet RF environment, but once you have an audience full of mobile phones in, the extra security it offers will be very useful. It uses four frequencies to transmit duplicated data whereas RF2 only uses two so it's a lot more resilient to interference.

1.5. It's unlikely that your UHF troubles are directly related. You're not leaving a L6 transmitter right next to your UHF receivers / antennae are you? Or clustering all of your receiving antennae very close to each other?

2. There's no gain adjustment but the headroom is enormous. I've never managed to make one clip. Some mics are more sensitive than others so it's technically possible that a real screamer eating the mic might cause a problem but I doubt it. The receiver outputs are set to roughly the same level as an equivalent wired mic output so as long as you treat the mic preamp gain the same as you would with any other mic, you should be fine.

No, UHF/L6 receivers are separate, and transmitters are all on stage. It turns out the actress in question was wearing her beltpack under a tight corset, thus pressing the antenna into her skin (and the corset possibly containing wires?). I separated the antenna from her body by using a neoprene pouch and the problem seems to have been resolved, RF is stronger.

Copyright: 2020 Skaar et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

The traditional physics approach to modeling typically involves four steps: (i) A hypothesis is formulated in terms of a candidate mechanistic mathematical model, that is, a model based on interactions between building blocks of the system, (ii) predictions of experimentally measurable quantities are calculated from the model, (iii) the predictions are compared with experiments, and (iv) if necessary, the hypothesis is adjusted, that is, a new candidate model is proposed. In neuroscience, a descriptive or statistical approach has been more common, in particular in systems neuroscience aiming to understand neural network behaviour in vivo. Here statistical techniques are used to look, for example, for correlations between measured neural activity and sensory stimuli presented to the animal to estimate receptive fields [1, Ch. 2]. While descriptive models can inform us about neural representations in various brain areas, mechanistic models may explain the biological mechanisms underlying these representations [2].

The Brunel network [28] consists of two local populations, one with excitatory and one with inhibitory neurons. These populations of size NE and NI, respectively, consist of leaky integrate-and-fire (LIF) neurons interconnected with current-based delta-shaped synapses. Inputs from external connections are modeled as uncorrelated excitatory synaptic input currents with activation governed by a fixed-rate Poisson process with rate νext.

Top row: First, the dynamics of a network is simulated using a point-neuron simulation (A), and the resulting spike times are saved to file. Orange and blue color indicate excitatory and inhibitory neurons, respectively. In a separate simulation, the obtained spike times are replayed as synaptic input currents onto reconstructed neuron morphologies representing postsynaptic target neurons (B, only one excitatory in orange and one inhibitory neuron in blue are shown). Based on the resulting transmembrane currents of the postsynaptic target neurons in this second simulation, the LFP is calculated (C). Bottom row: Prediction of LFPs from population firing histograms. Instead of running the full hybrid scheme, the LFP can be predicted by the convolution of the population firing histograms (lower figure in A) with kernels representing the average contribution to the LFP by a single spike in each population (lower figure in B). These kernels are computed using the hybrid scheme [29], see Methods.

The presently used choice of current-based synapses and morphologies with passive membranes in the multicompartment neuron models introduces a linear relationship between any presynaptic spike event and contributions to the LFP resulting from evoked currents in all postsynaptic multicompartment neurons. Thus the LFP contribution at position r from a single presynaptic point-neuron neuron j in population Y can, in general, be calculated by the convolution of its spike train with a unique kernel as . This kernel encompasses effects of the postsynaptic neuron morphologies and biophysics, the electrostatic forward model, the synaptic connectivity pattern, conduction delay and PSCs. (Note that to be in accordance with the notation used in [28], we here use a different notation than in [29] where instead i and X denoted presynaptic neurons, and j and Y denoted postsynaptic neurons.)

The resulting LFP due to spikes in a presynaptic population Y is then given by [29](9)The evaluation of this sum is computationally expensive for large population sizes. For our purposes where the calculation of LFP signals lasting seconds must be repeated tens of thousands of times to have training and test data for the CNNs, this scheme is not feasible.

Following [29, Fig 13] we instead use a firing-rate approximation and compute the LFP by a convolution of population firing rates and averaged kernels , that is,(10)As in [29], these averaged kernels were computed using the full hybrid-scheme. This was done by computing the LFP resulting from a fully synchronous activation of all the outgoing synapses from all neurons in the presynaptic population. Thus for the computation of the LFP kernel, we have where tY is the timing of the synchronous event in population Y. In the application of Eq 10, this computed kernel is then convolved with the population firing rates measured in the point-neuron simulations to compute the LFP.

By using this kernel approach the computational resources needed to run LFP simulations are reduced by several orders of magnitude compared to direct use of Eq 9. To test the accuracy of the approximation of using Eq 10 instead of Eq 9, we compared their LFP predictions for a set of example parameter sets and found in general excellent agreement between the resulting power spectra. A comparison is shown in the lower panels of the first figure in Results.

Two statistical measures were employed to probe the spiking network activity in the different regions of the parameter space. Simulations of 30.5 seconds of the activity were run and used to calculate the statistics, where the first 500 ms of the simulations were discarded. These simulations were run in addition to the ones entering the training and testing of the convolutional neural network (described in the next section), in order to have longer data sets for calculating reliable statistics characterizing the parameter regimes.

The mean network firing rate, including both the excitatory and inhibitory populations was calculated as(11)over all neurons i and their spikes l at spike times . The coefficient of variation (CV) of the inter-spike intervals (ISI) of individual neurons was used as a measure of the irregularity of firing [40]. The presently used mean CV was defined as(12)averaged over all neurons i.

The CNN architecture is illustrated in Fig 3 and fully described in Table 6, and was set up using the Keras machine learning framework [41] running on top of TensorFlow [42]. It consisted of three convolutional layers with 20 filters, each followed by max pooling layers, and two fully connected layers before the output layer. The rectified linear unit (ReLU) function f(x) = max(0, x) was used as the activation function for all layers apart from the output layer, and biases were only used in the fully connected layers. As input, it took the PSD of each LFP channel, a 6 by 151 matrix. The convolutions were done in one dimension, with kernels extending over all LFP channels. There were two fully connected layers, with 128 nodes each, before the output layer consisting of 3 nodes. Each node in the output layer corresponded to a single parameter η, g and J.

3a8082e126