The lion, Panthera leo, is a species of Felidae in Panthera. Its social associations, commonly organized as lion prides, make it useful for studying how a social group responds to other groups.^[ref]

A 1994 field study tested group assessment with a playback experiment. Recordings of a single female roaring or three females roaring together simulated unfamiliar intruders in territories of 21 lion prides in Serengeti National Park. Defending females were less likely to approach the three-intruder playback, and approached more cautiously when they did. Their own group's size and composition also influenced the response.^[ref]

This experiment changed an acoustic cue while recording a behavioral response, without requiring researchers to introduce actual rival animals. Comparing one opponent with three while accounting for the defenders' social group size tests a more precise question than whether lions sometimes approach a roar.

The result supports sensitivity to information associated with relative group strength. It does not by itself establish human-like arithmetic, conscious calculation or an identical strategy in every social setting. The useful scientific point is the connection between a controlled signal and a context-dependent decision to engage.

The tiger, Panthera tigris, is a species of Felidae in the taxonomic genus Panthera. A striped individual encountered in the field is both a member of that species and a particular animal with its own movement and detection history.^[ref]

This distinction made a useful measurement technique possible. A 1995 study in Nagarahole used camera traps along travel routes and identified individual tigers from photographs. Thirty-one photographic captures across nine sampling occasions represented ten different animals, not 31 animals. Their capture histories were then analyzed using mark and recapture methods that allowed different individuals to have different detection probabilities.^[ref]

For an original toy capture history, write 101 for an animal photographed on the first and third occasions, and 010 for another photographed only on the second. These are three detections but only two known individuals. Animals never photographed leave no such history, which is why an abundance estimate needs a detection model in addition to a tally.

The method connects a natural visible marking with statistical identification. It also exposes a limitation: placing cameras on travel routes samples encounters with those cameras, rather than guaranteeing that every tiger in the study region has been seen.

Overdispersion of detector-interval counts under shared drift is excess count variability caused by a common experimental condition changing the chance of an interval outcome. Let $M$ count empty physical detector-width intervals among $N$ readings. Conditional on a block's calibration state $H$, suppose the readings are independent with empty probability $p(H)$, so $M\mid H\sim \mathrm{Binomial}(N,p(H))$.

Writing $\overline{p}=\mathbb{E}[p(H)]$, the law of total variance givesThe first term is the binomial variance using the mean probability. The second comes from sharing one fluctuating condition across the block. Individual indicators have covariance $\mathrm{Var}(p(H))$ across different readings, even though they are independent after conditioning on $H$.

For an original mixture example, suppose equally likely block conditions have empty probabilities $0.01$ and $0.09$. Then $\overline{p}=0.05$ and $\mathrm{Var}(p(H))=0.0016$. With $N=100$, the expected count is five, but its variance is $4.75+15.84=20.59$, with standard deviation about $4.54$. An independent binomial model with probability $0.05$ would give standard deviation only about $2.18$.

The effective repetition count for detector intervals with shared drift can consequently be much smaller than the number of stored traces. Repeating measurements within one unchanged calibration block does not supply the same information about drift as acquiring independent blocks. Record acquisition grouping and reference reuse before pooling interval categories.

The negative covariance of empty and positive detector-interval counts applies to a fixed-probability multinomial model. Shared drift adds cross-repetition dependence and can change that count covariance as well as each count's variance. In the example, the two probability states are stipulated; a physical model would relate them to the receiver's drift, variance-difference estimator and uncertainty calibration.

NIST retains shared influences in uncertainty propagation and explains that confidence interpretations depend on the assumed distributions.^[ref] Observed overdispersion can motivate checking the block model, but does not by itself identify which physical influence caused it.

Randomized endpoint inclusion at a zero-duration atom adds an independent random decision about whether an interval contains the atom, allowing an exact included probability under a specified distribution. Let the duration $L\ge 0$ have zero mass $p_0$ and a continuous positive part. For a nominal central level $1-\alpha$, put $a=\alpha /2$, $b=1-\alpha /2$ and assume $a\le p_0<b$. Choose upper endpoint $u$ with $\mathrm{Pr}(L\le u)=b$.

The coverage excess from central quantiles at a zero-duration atom shows that the closed set $[0,u]$ has probability $b$, while $(0,u]$ has probability $b-p_0$. Draw an independent $U\sim \mathrm{Uniform}(0,1)$ and reportThen the joint probability over both $L$ and the auxiliary random draw isThe independence assumption is essential to the multiplication $r{p}_0$.

For an original example, let $p_0=0.20$ and $\alpha =0.05$. The endpoint is included with probability $r=0.875$. The positive part contributes $0.775$ probability and randomized zero inclusion contributes $0.175$, totaling $0.95$. An individual reported set is still either closed, with duration-model probability $0.975$, or open at zero, with probability $0.775$. Exact $0.95$ applies to the randomized procedure jointly, not conditionally to each realized set.

The distribution sensitivity of randomized zero-endpoint probabilities matters if $p_0$ or the upper endpoint is estimated inaccurately. Plugging a guessed atom probability into $r$ does not establish the same equality for the true distribution. The pinned numerical reference documents a standard uniform distribution and its random-variates method, while its normal cdf/ppf methods evaluate the Gaussian component when the duration comes from a clipped normal input.^[ref]

This is a mathematical probability construction. It does not remove the atom physically, improve measurement precision, or create a worst-case guarantee on missed optical energy. Its randomization rule and probability interpretation must be specified as part of the procedure.

Cavia is a taxonomic genus within Caviidae. It includes Cavia porcellus, the domestic guinea pig, and other species such as Cavia aperea. A genus name therefore identifies a broader group than the domestic species alone.^[ref][ref]

An identification example illustrates the information carried by the name. A sample labelled Cavia sp. is assigned to the genus but has not been given a species identification in that label. Replacing it with Cavia porcellus adds a specific claim that requires supporting evidence. Conversely, combining several differently identified Cavia samples under the genus name deliberately discards some taxonomic detail.

The same care applies to experimental results. A measurement made on one guinea pig preparation is evidence about that preparation and its stated biological source. Membership in Cavia helps organize comparisons with related species, but does not establish identical physiology, experimental history or response magnitude. The rank relationship is stable within the cited classification even when individual species treatments or names are revised.

Dolichotis is a taxonomic genus of rodents in Caviidae. It includes Dolichotis patagonum, the Patagonian mara. The alternative common name Patagonian hare does not place it in Lagomorpha: its family relationship is to rodents such as Cavia.^[ref][ref]

Field observations reveal an unusual combination of social scales. A 1992 study of Dolichotis patagonum described pair relationships together with communal breeding: young from several pairs shared a warren. Larger groups had greater collective vigilance and higher pup survival, although each pair spent less time sitting alert at the warren.^[ref]

An invented time-budget example shows why these measurements differ. Suppose each of two pairs watches during a different ten-minute interval: each pair watches for ten minutes, while the group has twenty minutes with a watcher present. If both watch simultaneously for ten minutes, individual effort is unchanged but collective coverage halves. Real coverage requires recording overlap, not merely adding individual durations. The field association between group size and survival motivates explanations involving predation; it does not by itself isolate a single cause.

The input-time horizon of a delayed optical receiver record is the latest physical input time that could influence the recorded output, given the receiver's causal onset delay. If $h(t)=0$ for $t<d$ and the last output time is $T$, convolution depends only on input timesThus $T-d$ is a necessary timing limit on information in that record. Recoverability within the limit additionally depends on bandwidth, noise, sampling and the response's invertibility. Causal indistinguishability of delayed optical inputs on a finite record shows why later input changes cannot be inferred from the existing output samples.

For several channels with last recorded times $T_j$ and known onset delays $d_j$, the latest possible input horizon isAny input perturbation starting strictly after $H$ is invisible in every stored channel. A channel's latest output timestamp alone is insufficient to compare its input horizon with another channel's.

For an original example, channel 1 records through $120\mathrm{ns}$ with delay $15\mathrm{ns}$, while channel 2 records through $100\mathrm{ns}$ with delay $2\mathrm{ns}$. Their input limits are $105$ and $98\mathrm{ns}$; the combined limit is $105\mathrm{ns}$. Knowing both response functions exactly still supplies no observation of an input pulse beginning later than that limit.

A receiver-latency uncertainty in the inferred input-time horizon also changes the timing claim. If channel 1's delay is known only to lie between $14$ and $16\mathrm{ns}$, its possible horizon ranges from $104$ to $106\mathrm{ns}$. An input beginning at $104.5\mathrm{ns}$ could affect the record for one allowed delay but arrive too late for another. Using the conservative endpoint $104\mathrm{ns}$ avoids claiming timing access that depends on an unresolved latency. It still does not guarantee successful reconstruction of every earlier input.

NIST characterized detector, cable and oscilloscope as a combined response chain using a short known optical pulse.^[ref] Such a calibration must relate output timing to the physical input time origin. The delays above are illustrative, not quoted apparatus values. De-embedding known latency aligns the record's time reference; it cannot extend observations into an unrecorded future interval.

Cost-weighted allocation of baseline-response tolerances chooses how tightly to enforce each baseline constraint when different tolerances require different effort. The allocation of a deterministic bias budget among baseline constraints gives the physical restriction ${\sum }_j{B}_j{\delta }_j\le E$. An optimization also needs an explicit cost model; the bias bound alone does not choose a unique allocation.

For one illustrative model, let $0<{\delta }_j\le {\delta }_j^{\mathrm{ref}}$ and minimizeTighter tolerances cost more, and the logarithm compares quantities in matching units. This is a stipulated convex effort model, not an empirical law for every numerical solver or measurement procedure.

When the budget binds and no reference upper bound is active, a constraint multiplier gives $-a_j/{\delta }_j+\lambda B_j=0$. ThereforeThe positive curvature $a_j/{\delta }_j^2$ establishes the unique interior minimum. A more expensive-to-tighten condition receives more of the allowed physical bias, while its coefficient bound converts that share to response units.

For an original two-constraint example, take $E=10\mathrm{nV}$, offset bound $B_1=10\mathrm{mV}$, drift-rate bound $B_2=100V/s$ and costs $a_1=1$, $a_2=3$. The optimum allocates $2.5$ and $7.5\mathrm{nV}$, giving ${\delta }_1=2.5\times 10^{-7}$ and ${\delta }_2=7.5\times 10^{-11}\mathrm{s}$. If their reference limits are $10^{-5}$ and $10^{-8}\mathrm{s}$, both are interior. Compared with equal physical bias shares, this optimum reduces the chosen cost by $\log (1/2)+3\log (3/2)\approx 0.523$ cost units.

Active-set clipping in cost-weighted baseline-tolerance allocation is needed if an interior solution exceeds a reference limit. If all reference limits already satisfy the bias budget, no tightening is needed in this cost model. Otherwise cap active components and redistribute the remaining budget among the others.

NIST's pulse-energy baselines motivate controlling residual offset and drift responses.^[ref] The selected tolerances control deterministic leakage within the stated coefficient bounds. Noise variance and unmodelled background shapes remain separate parts of the measurement analysis.

Causal indistinguishability of delayed optical inputs on a finite record means that an observation ending at a fixed time cannot discriminate inputs whose differences have not yet reached its receivers. The unobserved delayed-pulse counterexample to a fitted tail bound proves this for one calibrated convolution channel. The same information limit holds for any finite set of causal linear channels.

Let channel $j$ have response $h_j$ with $h_j(t)=0$ for $t<0$, and record $y_j=P\ast h_j$ only for $t\le T$. Every perturbation $q$ supported strictly after $T$ satisfiesfor every channel. Thus the finite-record observation operator has a whole family of invisible input directions. This is compatible with nonnegative optical power: choose any nonnegative delayed pulse and compare $P$ with $P+q$. No cancellation between unphysical negative powers is required.

Known propagation or detector latency can hide part of the input even before the record ends. If $h_j(t)=0$ for $t<d_j$, an input starting at $S$ produces no response before $S+d_j$. The input-time horizon of a delayed optical receiver record therefore depends on the channel's time reference, not just the last stored sample time.

For an original example, suppose two ideal delay channels measure $y_1(t)=P(t-2\mathrm{ns})$ and $y_2(t)=P(t-7\mathrm{ns})$, both recorded through $T=100\mathrm{ns}$. A pulse starting at input time $99\mathrm{ns}$ first appears at output times $101$ and $106\mathrm{ns}$. Both stored records are unchanged by that pulse, even though it began before the nominal record endpoint. Additional causal channels with smaller delays can extend the input horizon; none can observe an input that begins after their recording ends.

Offline processing of identical stored records still gives identical results, even if the processing filter itself uses samples on both sides of each stored time. To distinguish the inputs, obtain additional observations or supply independently justified restrictions on later illumination. Exact knowledge of the receiver response alone supplies neither.

NIST's pulse calibration characterized the detector, cable and oscilloscope as a combined response chain using a short optical input.^[ref] Such characterization is important for aligning the physical input and recorded output times. The two pure delays above are chosen models, not measured delays of that apparatus. A future-valid optical envelope or documented source cutoff remains additional information needed for an all-future missed-energy bound.

Caviidae is a taxonomic family of rodents containing the taxonomic genus Cavia, which includes the domestic guinea pig. The inspected NCBI family tree also contains Dolichotis. Thus Caviidae is broader than either one of those genera, while remaining within Rodentia.^[ref]

For a classification example, a record identified as Cavia porcellus belongs to Cavia and Caviidae. A record assigned to Dolichotis can share the family while belonging to a different genus. The shared family does not make the two records the same species, and a family-level sequence label does not identify the domestic guinea pig specifically.

This matters when comparing biological experiments. A reported response in a named species and preparation cannot be promoted automatically to a measured property of every member of Caviidae. The family supplies evolutionary and taxonomic context, while the experimental organism must still be identified at the resolution needed for the question. The cited classification fixes which provider's family usage is being followed; no numerical species count is needed to understand the distinction.

Allocation of a deterministic bias budget among baseline constraints assigns how much residual signal-response and background-leakage error each constraint may contribute. Let the allowed total bias be $E>0$, and let coefficient bounds $B_j>0$ multiply response errors $r_j$. Choose nonnegative fractions $f_j$ with ${\sum }_j{f}_j\le 1$ and setIf every computed response obeys $|r_j|\le {\delta }_j$, the triangle inequality gives $|\mathrm{Bias}|\le {\sum }_j{B}_j{\delta }_j\le E$. The response-normalized tolerance for baseline-feasibility residuals checks those component limits. Positive tolerances require positive allocated fractions; a zero allocation demands exact cancellation in the stated model.

For an original example, take $E=12\mathrm{nV}$ and bounds of $2\mathrm{mV}$ for signal amplitude, $10\mathrm{mV}$ for constant background and $100V/s$ for linear drift. Allocate fractions $1/6$, $1/3$ and $1/2$. The respective bias shares are $2$, $4$ and $6\mathrm{nV}$, giving tolerancesEqual bias shares would instead allow $4\mathrm{nV}$ each, relaxing the signal-response tolerance to $2\times 10^{-6}$ while tightening the drift tolerance to $4\times 10^{-11}\mathrm{s}$. Both allocations guarantee the same total bound when their assumptions hold; their numerical and experimental demands differ.

An allocation should reflect attainable constraint accuracy and the cost of tightening each response. Cost-weighted allocation of baseline-response tolerances can formalize that tradeoff. Equal numerical tolerances across differently dimensioned conditions do not mean equal physical bias contributions. If a bound $B_j$ is uncertain, use a justified upper bound or include its uncertainty explicitly before calling the total a guarantee.

NIST's pulse-energy apparatus used before/after measurements for baseline and drift correction.^[ref] The bounds above are illustrative design choices. This budget controls deterministic leakage within the assumed background family. It is separate from measurement-noise variance, whose covariance contributions must still be propagated.^[ref]

Coverage excess from central quantiles at a zero-duration atom is the additional distributional probability included when a closed quantile interval captures an entire point mass at zero. Here coverage means probability under the stated duration model. For $L=\tau \max (0,X)$ with $X\sim N(\mu ,s^2)$, the zero mass is $p_0=\Phi (-\mu /s)$ and the positive part is continuous. Let $a=\alpha /2$, $b=1-\alpha /2$, and use the lower quantiles from quantiles of a clipped Gaussian acquisition duration.

For the closed central interval $[q_a,q_b]$, its probability isTo verify this, write the probability as $F(q_b)-F(q_a^{-})$. When both endpoints are positive, the lower left limit is $a$ and the upper value is $b$. Once the lower endpoint is zero, its left limit is zero; if the upper endpoint is zero too, the interval contains exactly the zero atom. Endpoint inclusion is therefore part of the probability calculation.

For a nominal 95 percent central construction, zero masses $0.01$, $0.20$ and $0.99$ give actual probabilities $0.95$, $0.975$ and $0.99$ respectively. In the middle case, the excess is $0.025$, not the whole $0.20$ atom: the interval already includes the probability between zero and the upper quantile.

As an original dimensional example, set $\tau =20\mathrm{ns}$, $s=0.02$ and $\mu /s\approx 0.84162$, giving $p_0=0.20$. The nominal central 95 percent interval is approximately $[0,1.1206]\mathrm{ns}$ and contains probability $0.975$. SciPy documents the normal cdf and inverse-cdf ppf used to evaluate these Gaussian-component values.^[ref]

A one-sided upper interval or a separately specified probability construction can target a different included probability. Randomized endpoint inclusion at a zero-duration atom would require an explicit randomization rule, not a change made silently after seeing an endpoint equal zero. NIST explains why interval interpretation requires the input distributions and the chosen construction.^[ref] Distributional probability for this conditional acquisition model is separate from a worst-case optical-energy guarantee or a confidence procedure for an unknown fixed parameter.

Joint time-and-readout perturbations at a Raman design tie describe the directional response of the best acquisition variance when both available wall time and per-frame readout cost change. A tie generally has different derivatives on opposite sides, so there need not be one gradient for the optimum. The active branch derivatives must be combined along the actual perturbation direction.

Use the original balanced-noise model in one-sided derivatives at an optimal Raman frame-count switch. With $a=10^{-6}\mathrm{s}$, per-exposure cap $c=0.080\mathrm{s}$ and readout $d$, the locally competing branches areThey tie along $W=6c+8d$. At $(W_{\ast },d_{\ast })=(0.60,0.015)\mathrm{s}$, define $K=4a/(6c{)}^2\approx 17.3611\times 10^{-6}{\mathrm{s}}^{-1}$. The three-pair branch is locally constant, while the four-pair gradient is $(-K,8K)$.

For a path $W=W_{\ast }+ϵu_W$, $d=d_{\ast }+ϵu_d$ with $ϵ\downarrow 0$, the optimum has directional derivativeThe sign of the net live-time change $u_W-8u_d$ selects the improving branch. This conclusion assumes the branches remain feasible and nonactive designs have a positive value gap nearby.

For original finite examples, adding8ms of budget while adding1ms to each readout leaves $W-8d$ exactly unchanged, so these two branches remain tied. Adding10ms while adding1ms per readout increases that live time by2ms: the four-pair variance becomes $4\times 10^{-6}/0.482\approx 8.2988\times 10^{-6}$, below the three-pair value $8.3333\times 10^{-6}$. The first-order improvement is $-K(0.002)\approx -3.4722\times 10^{-8}$. Adding only6ms with the same readout increase instead makes the four-pair branch worse; the three-pair optimum remains unchanged locally.

The resource distinction has a physical motivation. The primary stimulated-Raman experiment used alternating pump states and shuttered detector readout; its exposure and readout intervals were separate parts of the acquisition sequence.^[ref] The rounded timings and noise parameter above are a mathematical planning example, not that instrument's measured values. Changing exposure caps, state-switching costs or noise terms requires the corresponding branch model to be recomputed.

The negative covariance of empty and positive detector-interval counts arises because a fixed number of measurement intervals is divided into mutually exclusive categories. Let $E_i,P_i$ indicate that interval $i$ is respectively physically empty or strictly positive. The same interval cannot satisfy both, so $E_i{P}_i=0$ andFor $N$ independent repetitions with common probabilities, the counts $N_E={\sum }_i{E}_i$ and $N_P={\sum }_i{P}_i$ therefore haveThe joint probability of empty and strictly positive detector-width intervals gives the underlying multinomial model. The sign describes category-count dependence, not physical antagonism between two detectors.

For an original boundary example, take $N=100$ and $p_E=p_P=0.025$. Each count has variance $100(0.025)(0.975)=2.4375$, while their covariance is $-0.0625$. The total tail-category count has variance $2.4375+2.4375-0.125=4.75$. Treating the two counts as independent would give $4.875$ instead. The covariance is needed even though the individual measurement repetitions are independent.

Conditioning on the total is important. If the number of repetitions is random and independent of the category assignments, the law of total covariance givesAn independent total with a Poisson distribution has equal mean and variance, making this covariance zero; a more variable total can make it positive. This is a different sampling protocol, not a contradiction of the fixed-total formula.

Shared calibration drift is another change of model. Repetitions may then have dependent category indicators, and cross-repetition covariance terms must be added. Overdispersion of detector-interval counts under shared drift can make a simple multinomial analysis inadequate.

NIST's measurement guidance retains common influences in uncertainty propagation and requires assumptions for confidence-level interpretation.^[ref] State whether the repetition total is fixed, how the categories are defined and which calibration influences are shared before interpreting their empirical count covariance.

Panthera is a taxonomic genus within Felidae, including the lion, Panthera leo, and the tiger, Panthera tigris. It is distinct from Felis, the taxonomic genus containing the domestic cat. A genus groups related species; membership does not make those species ecologically or behaviorally interchangeable.^[ref]

Molecular comparisons show why reconstructing those relationships is more interesting than sorting animals by appearance. A genomic study compared single-nucleotide polymorphisms, other nuclear markers and mitochondrial sequences across living Felidae. Different parts of their genomes did not always support the same branching history, and the authors found evidence consistent with ancient hybridization among some lineages.^[ref]

For an illustrative reasoning exercise, suppose two gene regions support different phylogenetic trees. One must examine processes that can generate that disagreement rather than select a tree merely because the corresponding animals look alike. A named taxonomic genus summarizes a classification; individual gene histories supply evidence that can complicate a simple picture of wholly separate ancestry.

The lion's group interactions and the tiger's photographic capture histories also illustrate a different distinction: comparative ancestry concerns evolutionary relationships, while those behavioral and ecological measurements concern living individuals.

Signal-subspace preservation for several spectral amplitudes requires a compression to retain every noise-weighted mean direction relevant to those amplitudes. Let $\mathbf{y}$ have mean $A\mathbf{\theta }$, fixed positive-definite noise covariance $C$, and a full-column-rank shape matrix $A$. Compress it as $B\mathbf{y}$, with $B$ full row rank. The original and compressed amplitude-information matrices areFor Gaussian noise these are Fisher information matrices. Without that distributional assumption, they still describe the corresponding known-covariance linear estimation problem.

Whiten the shapes as $T=C^{-1/2}A$ and put $M=B{C}^{1/2}$. The orthogonal projection onto the row space of $M$ is $P=M^T(M{M}^T{)}^{-1}M$, givingThis difference is positive semidefinite. It vanishes exactly when every column of $T$ lies in that row space, equivalentlyThus preserving one selected shape does not establish preservation of the whole signal subspace. Linear covariance propagation supplies $BC{B}^T$ for the compressed observations.^[ref]

For an original example, take $C=\mathrm{diag}(1,4,9)$ and shape columns $(1,2,0{)}^T$ and $(0,0,3{)}^T$. Then $J=\mathrm{diag}(2,1)$. Choosinggives $BA=\mathrm{diag}(2,1)$ and $BC{B}^T=\mathrm{diag}(2,1)$, hence $J_B=J$. Keeping only the first compressed row retains information2 for the first amplitude but destroys sensitivity to the second. An unweighted first row $(1,1,0)$ would instead reduce the first information entry to1.8.

At least $p$ independent compressed directions are needed to preserve $p$ independent shape amplitudes. One construction is $B=A^T{C}^{-1}$, whose compressed signal has mean $J\mathbf{\theta }$ and covariance $J$. When $J$ is invertible, $J^{-1}B\mathbf{y}$ has mean $\mathbf{\theta }$ and covariance $J^{-1}$. In the fixed-covariance Gaussian model, expanding the likelihood function shows that its parameter dependence is retained through $A^T{C}^{-1}\mathbf{y}$.

Unknown nuisance shapes, changing covariance or additional spectral parameters can alter the subspace that needs preservation. The proof concerns the specified family of mean shapes and fixed noise model, rather than guaranteeing that one compression preserves every future spectral question.

Failure of sign recovery from magnitude-only calibration data occurs when opposite signed calibration models induce exactly the same distribution of every recorded magnitude. Under this symmetry, more repetitions can estimate the magnitude distribution accurately without deciding the missing sign. This is the limitation relevant to sign-sensitive measurements that resolve a covariance-axis ambiguity.

Let $Z=sm+ϵ$ with sign $s\in \{-1,+1\}$, known $m>0$ and symmetric noise $ϵ$ having density $f(x)=f(-x)$. If the recorded value is $Y=|Z|$, then for $y>0$Symmetry gives $g(y\mid +1)=f(y-m)+f(y+m)=g(y\mid -1)$. Independent repetitions multiply identical densities, so their likelihood ratio remains one. With equal sign priors and equal decision costs, no decision rule based solely on these magnitudes can have average sign-error probability below1/2.

For normal distribution noise with standard deviation $\sigma$, the magnitude has a folded normal distribution:where $\varphi$ is the standard normal density. The formula depends on $|m|$, not its sign. The pinned numerical reference provides the equivalent standardized folded-normal density and its scale transformation.^[ref]

For the original calibration scale $m=0.20\mathrm{V}$ and $\sigma =0.050\mathrm{V}$, recording a signed value can distinguish the two models strongly, but recording only its magnitude cannot do so under the symmetric model. This is an information loss in the measurement map, rather than a shortage of averaging trials.

The symmetry assumption is essential. As an original counterexample, let noise equal−1 with probability2/3 and+2 with probability1/3, giving zero mean but an asymmetric distribution. With $m=1$, the positive-sign magnitudes are0 or3, while the negative-sign magnitudes are2 or1. Their disjoint supports permit sign recovery if this asymmetric noise model is independently established. The calibrated asymmetry itself supplies sign-reference information absent from the symmetric case.

Similarly, a genuinely signed projection or a known phase reference can break the original ambiguity. An unverified readout polarity or unknown amplitude sign can preserve it. Specify which signed information the apparatus actually records before interpreting a magnitude fit as an oriented calibration.

A response-normalized tolerance for baseline-feasibility residuals measures a pulse estimator's constraint errors against physically chosen response limits. For residual $\mathbf{r}=F^T\hat{\mathbf{a}}-\mathbf{b}$, assign positive tolerances ${\delta }_j$ in the corresponding response units and calculateThe criterion $\eta \le 1$ means every specified response limit is satisfied. Numerical feasibility residual for baseline constraints explains why a raw unscaled norm alone has no universal physical interpretation.

These tolerances can enforce a deterministic bias budget. If signal and background coefficients are bounded by $|\theta |\le B_s$ and $|{\beta }_j|\le B_j$, thenThe second inequality follows when the normalized test passes. The allocation of a deterministic bias budget among baseline constraints therefore connects a numerical acceptance criterion to a stated physical error allowance.

For an original example, bound the signal amplitude by $1\mathrm{mV}$, constant background by $20\mathrm{mV}$ and drift rate by $100V/s$. Choose signal-response tolerance $2\times 10^{-6}$, constant-response tolerance $2\times 10^{-7}$ and linear-time-response tolerance $4\times 10^{-11}\mathrm{s}$. Their maximum bias contributions are respectively $2$, $4$ and $4\mathrm{nV}$, totaling $10\mathrm{nV}$. If all actual residuals are half these tolerances, $\eta =0.5$ and the corresponding bias bound is $5\mathrm{nV}$.

The normalized test can be invariant to constraint units. Multiplying one equation and its required response by a nonzero factor $q_j$ multiplies its residual by $q_j$; replacing its tolerance by $|q_j|{\delta }_j$ keeps the ratio unchanged. The associated coefficient bound changes inversely, so the physical product $B_j{\delta }_j$ is unchanged as well.

NIST's before/after pulse baselines supply a measurement setting where residual offset and drift responses matter.^[ref] The example bounds and tolerances are stipulated, not specifications of that apparatus. This deterministic leakage test is separate from noise variance, which needs the actual measurement covariance.^[ref] Backgrounds outside the assumed family also need separate treatment.

The condition for preserving a spectral signal under linear compression specifies when compression retains the covariance-weighted signal measure of a known shape. Let the mean be $\theta \mathbf{s}$, with fixed positive-definite noise covariance $C$, and compress by a full-row-rank matrix $B$. The measures in noninvertible spectral compression versus a coordinate change obeyThus the rows of the compression must retain the noise-weighted signal direction, not necessarily the unweighted plotted shape.

For a short proof, whiten the shape as $\mathbf{t}=C^{-1/2}\mathbf{s}$ and put $M=B{C}^{1/2}$. The parent's orthogonal projection formula givesEquality holds exactly when $\mathbf{t}$ lies in the row space of $M$. Multiplying that subspace relation by $C^{-1/2}$ yields the stated condition. Linear covariance propagation gives the compressed covariance $BC{B}^T$.^[ref]

One useful construction is the scalar matched filter statistic $z={\mathbf{s}}^T{C}^{-1}\mathbf{y}$. Its mean is $\theta J$ and variance $J$, so $z/J$ has variance $1/J$, the original optimal linear amplitude precision. In the Gaussian known-covariance model, this statistic also retains the likelihood's dependence on that single amplitude.

For an original example, choose $C=\mathrm{diag}(1,4)$ and $\mathbf{s}=(1,2{)}^T$. Then $J=2$ and the retaining row is $B=(1,1/2)$. Its signal response is two and noise variance two, giving $J_B=4/2=2$. An unweighted sum $B=(1,1)$ instead has response three and variance five, giving $J_B=9/5=1.8$: ten percent of this signal measure is lost.

Signal-subspace preservation for several spectral amplitudes requires retaining the corresponding family of weighted mean directions when all those amplitudes are to be preserved. Unknown nuisance shapes can also change the relevant estimation problem. A compression designed for one known shape does not automatically preserve every spectral question or uncertainty component.

Quantiles of a clipped Gaussian acquisition duration assign probability levels to a recording extension whose distribution has a point mass at zero. Under the model of Gaussian input uncertainty at a clipped acquisition-time threshold, let $X\sim N(\mu ,s^2)$, $s>0$, and $L=\tau \max (0,X)$ for fixed $\tau >0$. Its zero mass is $p_0=\Phi (-\mu /s)$.

Using the lower generalized quantile $q_p=\inf \{\ell :\mathrm{Pr}(L\le \ell )\ge p\}$ for $0<p<1$ givesEquivalently, $q_p=\tau \max (0,\mu +s{\Phi }^{-1}(p))$. The constant branch is essential: directly using a negative untruncated normal quantile would assign an impossible negative duration. SciPy's normal-distribution documentation identifies ppf as the inverse cumulative distribution function used for the Gaussian part of this calculation.^[ref]

For an original threshold example, take $\mu =0$, $\tau =30\mathrm{ns}$ and $s=0.01$. Half the probability is at zero, so the median is zero. The 90th and 95th percentiles are approximately $0.3845$ and $0.4935\mathrm{ns}$. The one-sided interval $[0,q_{0.95}]$ has probability $0.95$ under this stated model.

A nominal equal-tail 95 percent construction behaves differently: $[q_{0.025},q_{0.975}]=[0,0.5880]\mathrm{ns}$ actually contains probability $0.975$. Its lower endpoint includes the entire zero atom instead of excluding a 2.5 percent lower tail. This coverage excess from central quantiles at a zero-duration atom is a property of the mixed distribution, not a rounding error.

Quantiles describe uncertainty in the conditional acquisition requirement. They are not automatically a frequentist confidence set for an unknown fixed physical duration, nor a worst-case guarantee on unrecorded optical energy. NIST explains why an interval's probability interpretation requires the underlying distributions and assumptions.^[ref] Propagate any uncertainty in decay time or other envelope parameters jointly before applying this fixed-$\tau$ formula.