BE1021546B1 - METHOD AND SYSTEM FOR STORING WAVEFORM DATA. - Google Patents

Abstract

The method of the invention comprises the steps of: receiving waveform data (100); generating a main signal Y0; save main data relating to the main signal Y0; defining an input signal to be folded Yjn which is said main signal Y0 or a folded signal which was generated just before; generate a new folded signal Yf which is an overall reproduction, of lower resolution, of an input signal to be folded Yjn, and which has fewer samples than the latter, by performing the following steps: apply a folding algorithm to the entire input signal to be folded Yjn, such that: its samples satisfying folding rules are selected and its samples not satisfying said folding rules are deleted; then grouping the selected samples to form said new folded signal Yf; save sub-data relating to said folded signal Yf.

Description

TECHNICAL FIELD [0001] The present invention relates to a method for storing waveform data. More specifically, and in accordance with a first aspect, the present invention relates to a method for storing, from waveform data, useful information for efficient use of a waveform viewer. According to a second aspect, the invention relates to a device for storing, from waveform data, useful information for a waveform viewer.

Description of the Prior Art [0002] A waveform viewer is used to view waveform data. Waveform data is generally composed of a series of values representing the evolution of the amplitude of at least one variable in accordance with a parameter which is, for example, and most often time. Waveform data can come from, for example, measurements made on circuits or data extracted from a circuit design, or from an electronic design automation (EDA) tool performing a simulation of circuit design.

[0003] Generally, the complete set of waveform data is too large to fit on the display area of a waveform viewer. The display area of the waveform viewer has only a limited number of pixels, available to display a waveform that is defined by the dimensions and the pixel resolution of the viewer window of the viewer. waveform. Typically, the waveform viewer may be user-controlled to display all or part of the waveform. The user can zoom in or out to see the full or partial waveform at different resolution levels. It is likely that only a limited portion of all waveform data can be displayed on the waveform viewer at any point in time. Therefore, it is not necessary to provide the waveform viewer with all the bulky data of the waveform, when only a smaller subset is to be displayed.

The method disclosed in US Pat. No. 8,325,188 addresses some of these problems. In accordance with this method, waveform data is first segmented to fragment all of the waveform data into smaller portions or segments. Then these segments are compressed. Various compressor mechanisms can be used. The next action is to index the segmented and multiple resolution levels of the waveform data. The index data allows the system to identify and access the appropriate waveform data for the required resolution level to be displayed on the waveform viewer.

This method has several disadvantages. When a user wishes to display a long portion of a signal from the data of the waveform, it is necessary, in the first place, to reassemble the different segments that have been generated and saved, in order to obtain the said long part of the waveform. signal that the user wishes to display. Therefore, this method leads to an access to useful information, which is not simple: we must reassemble the various segments that have been generated and saved, in order to obtain the useful and sought-after information. In addition, this way of saving information from waveform data is not well suited to the different levels of screen resolution that have been selected by a user of a waveform viewer. . As a result, this method of saving information from waveform data is not efficient enough for subsequent display of data relating to these waveform data, on a waveform viewer. 'wave. SUMMARY OF THE INVENTION [0006] In accordance with a first aspect, an object of the invention is to provide a method for storing useful information from waveform data, which method provides subsequent access to such information. information, which is simpler and more efficient to display, on a waveform viewer, data relating to these waveform data. For this purpose, the inventors propose the process that follows. A method for storing, from waveform data, useful information for a waveform viewer, and comprising the steps of: (i) receiving said waveform data; (ii) generating, from these waveform data, a main signal Yo having N0 samples, Y0 (i) i = 1, ... N0 which are Novators of said main signal Yo, at N0 values of a reference signal X; (iii) backing up primary data relating to said main signal Yo; (iv) defining an input signal to be folded, Yin, which is: - if no folded signal from said main signal Yo has been generated before, said main signal Yo, - or a folded signal from said main signal Yo that was generated just before; (v) generating a folded signal Yf: - which is a global reproduction, of lower resolution, of said input signal to be folded, Yjn, - which has Nf samples which are Nf values of said folded signal Yf, at Nf values of said signal reference number X, and - which has fewer samples than said input signal to be folded, Yin, by performing the following steps: a) applying a fallback algorithm, substantially to the whole of said input signal to be folded, Yjn in such a way that: samples of said input signal to be folded, Yjn satisfying fallback rules, are selected; that samples of said input signal to be folded, Yjn, not satisfying said fallback rules, are deleted; b) then grouping the selected samples in step a) to form said refolded signal Yf, (vi) saving subdata relating to said refolded signal Yf, (vii) optionally repeating steps (iv) to (vi) .

The terms "useful information for a waveform viewer" mean information that can be used by a waveform viewer, to provide a user of such a waveform viewer, relevant information. Relevant information means, for example, information that could help a user to know if an electronic circuit is working properly. With the method of the invention, different folded signals Yf are generated. Each folded signal Yf is a reproduction of the main signal Yo in its entirety (or its entirety), since each folded signal Yf is a reproduction, of lower resolution, of practically (substantially) all of each input signal. to fold Yin. And this input signal to fold Υ, η is the main signal Yo at the beginning of the iterative process. In other words, the overall shape (or envelope) of the various input signals to fold Υ, η and, consequently, of the main signal Yo, is preserved during the generation of the different folded signals Yf. Therefore, unlike the method of US Pat. No. 8,325,188, a signal of interest is not, with the method of the invention, divided into segments for storage purposes. The entire envelope of the main signal Yo is considered and kept at each fallback level. The different fallback levels concern the different folded signals Yf that have been generated. Therefore, unlike the method of US Pat. No. 8,325,188, it is not necessary to reassemble different segments in order to view a signal in its entirety (no segment of such a signal is generated). Finally, the method of the invention is simpler for subsequent access to relevant information, such as data concerning a signal in its entirety.

[0008] Main data relating to the main signal Yo and sub-data relating to the various folded signals Yf, which are saved during steps (iii) and (vi), can be used for the display, by performing, with the method of the invention, minimal operations of reconstruction. Depending on the limits and the resolution of the window that is selected by a user of a waveform viewer, it is possible to determine a number of pixels that is available to display a signal between these boundaries. Then, it is possible for a central computer unit, for example to choose the folded signal Yf (or the main signal Yo) which is best suited to display the signal between these limits. No additional decompression steps are necessary. Saved data can be used for display, performing minimal reconstruction operations. As a result of the foregoing, the method of the invention is more efficient for displaying, on a waveform viewer, data relating to data of a waveform.

The method of the invention has other advantages. Since each folded signal Yf has the same main envelope as the main signal Y0, an overall image of a signal can still be displayed and is always available to the user. This facilitates a fast global analysis of a signal. And, subsequently, useful information can be found more easily and quickly by a user of a waveform viewer. With the method of the invention, data of a waveform is encoded in such a way that excessive access and data transfer, beyond what is actually necessary for display, are minimized. .

Preferably, main data relating to said main signal Yo are saved in a main data file, during step (iii). Preferably, sub-data relating to said folded signal Yf are saved in a sub-data file, during step (vi).

[0011] Preferably, said No samples of said main signal Y0 represent amplitudes of a digital signal. The samples of the folded signals are then also amplitudes of a digital signal. Preferably, said No samples of said main signal Yo represent amplitudes of an analog signal. Similarly, the samples of the folded signals are also amplitudes of an analog signal. Preferably, said main signal Yo represents a combination of signals contained in said data of the waveform. This preferred embodiment has the following advantages in particular: clarity and ease of reading information which are then displayed on a waveform viewer. Such a preferred embodiment is preferably used in dealing with digital signals. A combination of bits makes it possible to form a bus. A digital signal can then be transformed into an analog signal to which min algorithms can be applied. / max.

Preferably, said main signal Y0 is generated using a transition detection method, such that said main signal mainly contains information concerning transitions. Said method of transition detection can be compared to the value change data transfer method (VCD method) which is known to those skilled in the art. With said transition detection method, transitions are first detected. Then, only data related to these detected transitions are selected to generate a main signal Yo. By using this preferred embodiment, the size of the main signal Yo can be reduced, as can the size of the preferred main data file, as well as that of the folded signals Yf and the preferred subdata files.

Preferably, said main data relating to said main signal Yo are saved during step (iii), using a method that records transitions. This method can be compared to a process similar to the VCD method. In this preferred embodiment, the main signal Yo is preferably a signal extracted from the data of the waveform or, more preferably, a combination of signals extracted from the data of the waveform. By storing (or saving) the main data of step (iii) according to such a preferred method, it is possible to reduce the size of the main data file when such a primary data file is used. More generally, it is then possible to reduce the size of the memory needed to save main data relating to the main signal Y.sub.O. Preferably, subdata relating to the folded signals Yf are also saved during step (vi). , using the same method. Similarly, the size of the sub-data file (s) can also be reduced when using such or such sub-data file (s). More generally, it is then possible to reduce the size of the memory necessary to save sub-data relating to a sub-signal or sub-signals Yf. Preferably, temporal information is also saved in addition to the information relating to the transitions. More preferably, values relating to time intervals during which the main signal Y0 or folded signals Yf is or are constant, are saved with a first value of said main signal Yo or with a first value of said folded signals Yf such time intervals. Then, it is easier to reconstruct the main signal Y0 or folded signals Yfl from information relating to the transitions.

[0014] Preferably, said fallback algorithm comprises the following steps: - dividing said input signal to fold, Yin, each portion comprising essentially Np successive samples of said input signal to be folded, Ym, the number Np being £ 3; - apply, for each of said portions, the following steps to select two samples in each of said portions: o a sample, whose absolute value is equal to or greater than that of the other samples, is selected, o another sample, whose value absolute is equal to or smaller than that of the other samples, is selected, and o the other samples are deleted.

Np is an integer. Np is less than the number of samples of the input signal to fold Yin. Each portion essentially comprises Np successive samples of the input signal to fold Yj ". This means that small portions (for example one, two or three) could comprise more or fewer samples, since it is not always possible to divide the input signal to fold, Υ, η, in portions which all include Np successive samples. This preferred embodiment is particularly well suited when dealing with analog signals or with data relating to analog signals.

Preferably, said folding algorithm comprises the following steps: - dividing said input signal to be folded, Yjn, each portion comprising essentially N successive samples of said input signal to be folded, Υη, the number NP being £ 3; applying for each of said portions the following steps for selecting two samples in each of said portions: the sample Yn (q) having the lowest index q is selected; o the sample γη (ρ) having the highest index p is selected; o the other samples are deleted.

Np is an integer. Np is less than the number of samples of the input signal to fold Yin. Each portion essentially comprises Np successive samples of the input signal to fold Υ, η. This means that small portions (eg one, two or three) may have more or fewer samples, since it is not always possible to divide the input signal to fold, Yin, into portions which include all Np successive samples. This embodiment is particularly well suited when dealing with digital signals or with data relating to digital signals.

[0016] Preferably, said fallback algorithm comprises the following steps: - dividing into portions said input signal to be folded, Υ, η, each portion comprising essentially Np successive samples of said input signal to be folded, Yin, the number Np being £ 3; - applying, for each of said portions, the following steps for selecting a sample in each of said portions: o detecting whether a transition occurs among the successive samples of the portion; if so, a sample Yn (b) of the portion is selected, where b is neither the lowest nor the highest index in the portion, the other samples of the portion are deleted; - if it is not the case, no sample is selected.

Np is an integer. Np is less than the number of samples of the input signal to fold Yin. Each portion essentially comprises Np successive samples of the input signal to fold Yin. This means that small portions (eg one, two or three) may have more or fewer samples, since it is not always possible to divide the input signal to fold, Yin, into portions which include all Np successive samples. This preferred embodiment is particularly well suited when dealing with digital signals or with data relating to digital signals. Preferably, a transition is detected by comparing amplitudes of two successive samples. If two successive samples have different amplitudes, we can conclude that a transition takes place. Otherwise, there is no transition.

[0017] Preferably, the method of the invention also makes it possible to access said useful information, to display it on a waveform viewer, and furthermore comprises the following steps: - providing a range ΔΧ of said reference signal X; providing a display window width of said waveform viewer; providing a number of pixels Νρ, χ available for display in said width of said display window; selecting, between said main signal Yo and said folded signals Yf, one of those having, in said range ΔΧ of said reference signal X, a number of samples closest to said number of pixels Npix.

This preferred embodiment of the method of the invention also provides access to said useful information for display on a waveform viewer. By using such a method, access to useful information is more effective than that provided in other methods known to those skilled in the art.

[0018] Preferably, the method of the invention also makes it possible to access said useful information, to display it on a waveform viewer, and furthermore comprises the following steps: - providing a range ΔΧ of said reference signal X; providing a display window width of said waveform viewer; providing a number of Npix pixels available for display in said width of said display window; selecting, between said main signal Yo and said folded signals Yf, one of those having, in said range ΔΧ of said reference signal X, a number of samples equal to or greater than said number of pixels Npix.

Preferably, said main data relating to said main signal Yo are saved in a main data file during step (iii), said main data file comprising at least one element having a value, of number n, said reference signal X (n) and a value of a sample Y0 (n) of said main signal, at said value, of number n, of said reference signal X. The recovery of a sample of the main signal Yo corresponding to a given value of the reference signal X is easier and faster with this preferred embodiment. This preferred embodiment of the method of the invention is therefore more efficient. A comparison of said given value of the reference signal X, with said value, of number n, of said reference signal X (n) can in fact be carried out, in order to quickly identify in which part (s) of the file of main data is located the sample of the main signal Y0 corresponding to said given value of the reference signal X.

[0020] Preferably, said sub-data relating to said folded signal Yf are saved in a sub-data file during step (vi), said sub-data file comprising at least one element having a value, of number n, of said reference signal X (n) and a value of a sample Yf (n) of said folded signal, at said value, of number n, of said reference signal X. Recovery of a sample of a signal folded Yf corresponding to a given value of the reference signal X is easier and faster with this preferred embodiment. This preferred embodiment of the method of the invention is therefore more efficient. A comparison of said given value of the reference signal X, with said value, of number n, of said reference signal X (n) may in fact be carried out, in order to quickly identify in which part (s) of the subfile file (s) is located the sample of said folded signal Yf corresponding to said given value of the reference signal X.

According to a second aspect, the invention relates to a device for storing, from waveform data, useful information for a waveform viewer, and comprising: - means for input to receive waveform data; processing means for generating, from said waveform data, a main signal Yo having No samples, Yo (i) i = 1, ... No, which are No values of said main signal Y0, at N0 values of a reference signal X; backup means for storing main data relating to said main signal Yo; pretreatment means for defining an input signal to be folded, Yjn, which is: o if no folded signal originating from said main signal Yo has been generated before, said main signal Yo, o or else a folded signal originating from said main signal Yo and which was generated just before; processing means for performing the following steps: a) applying a folding algorithm to substantially all of said input signal to be folded, Υη, in such a way that: o samples of said input signal to be folded, Yjn, satisfying fallback rules, be selected; o samples of said input signal to be folded, Yin, not satisfying said fallback rules, are deleted; b) subsequently grouping the selected samples in step a) to form a folded signal Yf; said folded signal Yf being: o a global reproduction, of lower resolution, of said input signal to be folded, Yin, o having Nf samples which are Nf values of said folded signal Yf, at Nf values of said reference signal X, and o having fewer samples than said input signal to be collapsed, Yin; backup means for saving sub-data relating to said folded signal Yf.

Preferably, said preprocessing means, said processing means and said backup means can operate iteratively, in order to define successive input signals to be folded Y, n, in order to generate successive folded signals Yf and to save such successive folded signals Yf.

[0023] Preferably, said successive folded signals Yf are saved in different sub-data files 11.

Preferably, the device of the invention further comprises an access unit for selecting, between said main signal Yo and said folded signals Yf, one of those having, within a range ΔΧ of said reference signal X , a number of samples that is closest to a number of available Npix pixels in a display window width of a waveform viewer. Preferably, said access unit is able to choose between said main signal Yo and said folded signals Yf, one of those having, in a range ΔΧ of said reference signal X, a number of samples which is larger or equal to the number of pixels NPjX available in a display window width of a waveform viewer.

[0025] Preferably, said main data relating to said main signal Yo are saved in a main data file. Preferably, said no samples of said main signal Y0 represent amplitudes of a digital signal. More preferably, they represent amplitudes of an analog signal. Preferably, said main signal Yo represents a combination of signals contained in said waveform data. Preferably, said processing means are adapted to generate said main signal Yo by using a transition detection method, such that said main signal Yo essentially contains information concerning transitions. Preferably, said backup means are able to save main data relating to said main signal Yo using a method that records transitions. Preferably, said processing means are adapted to apply, to generate said folded signal Yf, a folding algorithm comprising the following steps: - portioning said input signal to be folded, Yin, each portion comprising essentially Np successive samples of said input signal to fold, Υη, the number Np being> 3; - apply, for each of said portions, the following steps to select two samples in each of said portions: • a sample, whose absolute value is greater than or equal to that of the other samples, is selected, • another sample, whose value absolute is smaller or equal to that of the other samples, is selected, and • the other samples are deleted.

In other preferred versions, the processing means are adapted to apply, to generate said folded signal Yf, the preferred fallback algorithms that have been presented with the method according to the first aspect of the invention.

BRIEF DESCRIPTION OF THE FIGURES [0026] These and other aspects of the invention will be explained in more detail by way of examples and with reference to the accompanying figures in which:

FIG. 1 shows in particular a diagram illustrating the main steps of the method of the invention;

Fig. 2 shows an example of a digital master signal and a corresponding folded signal;

Fig. 3 shows an example of an analog main signal and a corresponding folded signal;

Figure 4 shows an example of a preferred sub-data file;

Figure 5 shows a preferred embodiment of the device of the invention;

Figure 6 shows another preferred embodiment of the device of the invention.

The drawings of the figures are neither drawn to scale nor proportioned. In general, like components are indicated by the same reference numerals in the figures.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS [0027] FIG. 1 shows a diagram illustrating the main steps of the method of the invention. The first step (step (i)) is to receive waveform data 100. The waveform data 100 may come from measurements made on circuits or from data taken from a circuit design or for example, come from an EDA tool performing a circuit design simulation (the EDA tools are known to those skilled in the art). The waveform data 100 is generally composed of a series of values representing the evolution of the amplitude of at least one variable according to a parameter which is, for example, time. In what follows, this parameter is called a reference signal X. This reference signal X is preferably a clock signal known to those skilled in the art. Preferably, the waveform data 100 comprises one or more physical signals such as temporal changes relating to voltages, current flowing in electronic devices. In the general scheme shown in FIG. 1, the waveform data 100 comes from a data source 1 (i.e., comes from, for example, an EDA tool or a unit transmitting measurement results).

In a second step (step (ii)), a main signal Yo is generated from these waveform data 100. This main signal Y0 has No samples: Yo (i) i = 1 , ... N0, where No is an integer. These No samples of the main signal Yo are No values of said main signal Yo, N0 values of the reference signal X. The main signal Y0 can be of different types. When the waveform data 100 comprises changes in different signals with respect to the reference signal X, the main signal Y0 may be one of these different signals. The main signal Y0 may also be a combination of various signals contained in the waveform data 100. These various signals may have different numbers of samples corresponding to various reference signals X. For example, one may have a first signal YYi contained in the waveform data 100 and corresponding to a first reference signal XXi, and a second signal YY2 contained in the waveform data 100 and corresponding to a second reference signal XX2, the signals XXi and XX2 being different in a general case. Similarly, it is possible to generate a main signal Yo by combining the signals YYi and YY2. Preferably the first of these two signals then defines a common (or shared) reference signal X having a higher resolution (or a larger number of values) than that of XXi and XX2.

Methods known to those skilled in the art can be used to generate a main signal Yo which is a combination of various signals contained in the waveform data 100: for example, two measured signals can be added . Other examples of the main signal Yo are presented below.

Once the main signal Yo has been generated, data relating to this main signal are saved, preferably in a main data file 10: this constitutes step (iii). The main signal Y0, in its entirety, can be saved in the main data file 10. Preferably, specific methods are used to save the main signal Yo, in order to reduce the size of the memory required. This is particularly interesting when the main signal Yo represents a whole waveform data signal 100, which means a signal that has not been subjected to any method, to reduce its size. The main signal Yo, in its entirety, is then preferably unsaved. Data relating to this main signal Yo are then rather saved. A transition detection method which can be compared to a value change data transfer method (VCD) is preferably used to generate such data relating to this main signal Yo, which are saved. The VCD process is known to those skilled in the art.

Steps (iv) - (vi) are preferably repeated iteratively, in order to generate different folded signals Yf. However, only one iteration could be applied, so that only a folded signal Yf is generated. A folded signal Yf is an overall reproduction (or duplication), of lower resolution, of an input signal Υ, η to fold. During the first iteration, the input signal Υ, η to fold is the main signal Yo. Then, the input signal Υ, η to fold is a folded signal Yf that has been generated during a previous iteration. Each folded signal Yf has Nf samples which are Nf values of this folded signal, at Nf values of the reference signal X. At each successive iteration, the number Nf is smaller. Each folded signal Yf has fewer samples than the input signal Yin to collapse, from which it is generated. Each folded signal Yf is generated in step (v) by applying a fallback algorithm to substantially (or substantially) the entire input signal Yjn to fold. The term substantially (or essentially) means here that the fallback algorithm is applied to at least 80% or preferably at least 90% of the samples of said input signal to fold Yin. More preferably, the fallback algorithm is applied to at least 95% of the samples of the input signal to replicate Yin. Even more preferably, the fallback algorithm is applied to all samples of the input signal to replicate Yjn. As a result of the fallback algorithm, the fact that samples of the input signal to be folded Yj ", satisfying fallback rules, are selected, while samples of the input signal to fold Yjn, not satisfying fallback rules, are removed. By grouping the samples that are selected by the fallback algorithm, a new folded signal Yf can be obtained. For purposes of illustration, a theoretical example is considered. It is assumed that the input signal to refold Yin is the main signal

Yo which has No samples: Yo (1), .... Yo (10) ..... Yo (No). If Yo samples (1),

Yo (10) and Yo (No) are selected by the fallback algorithm, the folded signal Yf generated from the main signal Yo has the following samples: Yo (1), Yo (10), Y0 (N0). In this case, Nf = 3.

The inventors propose various fallback algorithms for generating a folded signal Yf which is an overall reproduction, of lower resolution, of the input signal to fold Yin and which has fewer samples than said input signal to fold Yjn . Some examples are given below. During step (vi), sub-data relating to the folded signal Yf that has been generated is saved, preferably in a sub-data file 11. As discussed above, in connection with the main data file 10, different ways to save subdata can be used. Raw folded signals Yf can be saved or, preferably, data relating to said folded signals Yf are saved when it is desired to reduce the size of the memory required. A method similar to the VCD method can be used for example to save data relating to the folded signals Yf.

[0033] Preferably, main data is saved using a compression algorithm. More preferably, one of the following compression algorithms is used: ZIP (Alphabetical Algorithm), Zlib, LZMA, libzip, mimiz. Preferably, subdata are saved using a compression algorithm. More preferably, one of the following compression algorithms is used: ZIP (Alphabetical Algorithm), Zlib, LZMA, libzip, mimiz.

FIG. 2 shows an example, by way of illustration, of a main signal Yo and of a folded signal Yf generated with the method of the invention when the main signal Y0 is a digital signal. From this figure, we see that the overall envelope of the main signal Yo is preserved in the folded signal Yf.

Figure 3 shows another example, by way of illustration, of a main signal Yo and a folded signal Yf generated with the method of the invention when the main signal Yo is an analog signal. The central portion of Figure 3 shows samples of the main signal Y0 which are selected by the fallback algorithm.

In a preferred embodiment, the main signal Y0 is a reproduction (or an image) of a signal of the waveform data 100. This means, for example, that certain mathematical operations can be applied to a signal of waveform data 100, to generate the main signal Yo. Preferably, the main signal Yo is generated using a transition detection method, such that said main signal Yo contains mainly information concerning transitions. This method of transition detection can be compared to the value change data transfer method (VCD) which is known to those skilled in the art. If a transition is detected, a corresponding sample of the signal is selected to generate the main signal Yo. If no transition is detected, then the index number indexing the different samples of the signal is simply incremented without any sample storage of the signal. Therefore, this preferred method stores only transitions. Various methods known to those skilled in the art can be used to detect transitions. When dealing with a digital signal, a transition is easy to detect, since it represents a change from 0 to 1 or from 1 to 0. As detailed below, when a transition detection method is used for generating a main signal Yo, time information, such as a time interval during which the main signal Yo is constant, is preferably also used to generate said main signal or to define said main data which is saved during step (iii). Even if only information about transitions is first selected, a "real" signal can then be easily reconstructed from such temporal information.

[0037] Now, we present different fallback algorithms. They can be used alone or in combination to generate the folded Yf signals. According to a first preferred example, such a folding algorithm comprises the following steps: - dividing into portions the input signal to be folded, Yin, each portion comprising essentially Np successive samples of said input signal to be folded, Yin, the number Np being> 3; - applying, for each of said portions, the following steps to select two samples in each of said portions: o a sample, whose absolute value is equal to or greater than that of the other samples, is selected, o another sample, whose absolute value is equal to or less than the other samples, is selected, and o the other samples are deleted.

According to a second preferred example, the fallback algorithm comprises the following steps consisting in: - splitting the input signal to fold Yjn, each portion comprising essentially Np successive samples of said input signal to be folded, Yin, the number Np being> 3; - to apply, for each of said portions, the following steps to select two samples in each of said portions: o the sample Yin (q) having the lowest index q is selected; o the sample Yjn (p) with the highest index p is selected; o the other samples are deleted.

Therefore, this second preferred example corresponds to a method that selects the first and the last sample of each portion.

According to a third preferred example, the fallback algorithm comprises the following steps consisting in: - dividing said input signal to be folded, Yjn, each portion comprising essentially Np successive samples of said input signal to be folded back Yin, the number Np being> 3; - applying, for each of said portions, the following steps for selecting a sample in each of said portions, consisting of: o detecting whether a transition occurs among the successive samples of the portion; if so, a sample Yin (b) of the portion is selected, where b is neither the lowest nor the highest index in the portion, the other samples of the portion are deleted; if not, no sample is selected.

Preferably, a transition is detected by comparing two successive samples. If the amplitude of two successive samples changes, we can then deduce that a transition occurs. Otherwise, there is no transition.

Preferably, the first and second preferred examples for the fallback algorithm, which have been previously described, are used in a combined manner. More preferably, the first, second and third preferred examples for the fallback algorithm, which have been previously described, are used in combination.

For each of these three preferred examples, when the input signal to fold Υ, η comprises 1024 samples, said input signal is preferably divided into 64 portions, each portion comprising substantially Np = 16 successive samples .

By grouping the different Nf samples of the signal Yjn, which have been selected by the fallback algorithm, the folded signal Yf can finally be obtained, preferably using the following preferred procedure: the first sample of the folded signal Yf (1) corresponds to the selected sample of Υ, η having the lowest index: Yf (1) = Yn (k), where k is the lowest index among the indexes of the samples of the signal Υη, said samples having been selected by the fallback algorithm; the last sample of the refolded signal Yf (Nf) corresponds to the selected sample of the signal Y, n having the highest index: Yf (Nf) = Yin (m), where m is the highest index among the indices of the samples of the Yin signal, said samples having been selected by the fallback algorithm; the other selected Nf-2 samples of the signal Yin are organized between Yf (1) = Yin (k) and Yf (Nf) = Yin (m), so that indices j of these selected samples of the signal Yin increase by invariant way between k and m. k and m are integers.

According to another preferred embodiment, the method of the invention also makes it possible to access useful information stored from the waveform data 100, to display them on a waveform viewer. 200. A waveform viewer 200 is known to those skilled in the art. This preferred embodiment comprises the following preferred steps, consisting in: providing a range ΔΧ of the reference signal X between a minimum value Xmin and a maximum value Xmax, ΔΧ = Xmax-Xmin (this step is preferably performed by a user selecting a region of interest, preferably using a pointing device such as a mouse); to provide a display window width of said waveform viewer 200; to provide a number of pixels Npix available for display in said display window (preferably, NPiX corresponds to a number of pixels in a horizontal direction); to select, between the main signal Y0 and the folded signals Yf, which have been generated by the method of the invention, one of those having a number of samples in said range ΔΧ of said reference signal X which has a number of samples that is closest to said number of NPjX pixels.

According to another preferred embodiment, this last step may be replaced by that consisting in: - choosing between the main signal Y0 and the folded signals Yf which have been generated by the method of the invention, one of those having a number of samples in said range ΔΧ of said reference signal X which has a number of samples which is equal to or greater than said number of Npix pixels.

In the latter preferred case, it is therefore possible that the signal selected for display has a number of samples greater than the number of NPiX pixels. Next, the following display procedure is preferably followed: several samples of the selected signal are displayed with the same pixel (horizontal) over the entire width of the display window of the waveform viewer 200.

Therefore, according to these preferred embodiments that have been described in the previous paragraph, it is possible to more easily access the signal that is best suited for the display window selected by a user. Such easy access is possible, in particular thanks to the specific generations and backups of the main signal Y0 and the folded signals Yf. In order to choose the most suitable signal between the signal Y0 and the various signals Yf, the number of samples in each of these signals, namely the signal Yo and the various signals Yf, in the range ΔΧ must be determined. It is then possible to determine a difference value for each signal, namely the signal Yo and the various signals Yf, that is to say the difference between the number of pixels NPjX and the number of samples of each of these signals in the ΔΧ range. A signal with the smallest difference value is selected. Since the number of samples of the selected signal is not generally equal to the number of pixels Npix, a processing step is generally applied thereafter, in order to have a value of the signal selected at each pixel of the window. display. The interpolation, which is known to those skilled in the art, is preferably applied when the selected signal has a number of samples smaller than the number of pixels Νρ, χ.

According to another preferred embodiment, main data relating to the main signal Yo, which are saved during step (iii), and sub-data relating to each folded signal Yf, which are saved in course of step (vi), include temporal information. For example, when the Yo and Yf signals are backed up with a process similar to the VCD process, information about transitions is saved. For example, for the main signal Yo, the index i is incremented, and each sample Y0 (i) is compared with the previous sample Y0 (i-1). If Yo (i) is different from Y0 (i-1), the sample Yo (i) is saved. Then the index i is further increased. Assuming that all samples from Yo (i + 1) to Yo (i + 10) have the same value as Y0 (i), but Y0 (i + 11) is different from Y0 (i), the number 10 is then a main data relating to the main signal Yo, principal data which is saved with the sample Yo (i) during step (iii). The number 10 corresponds to an interval of the reference signal X, during which the main signal Yo is constant, which interval is hereinafter called "constant interval". When the method of the invention also provides access to useful information stored from the data 100 of the waveform, for display on a waveform viewer 200, the method then proceeds as follows. By choosing between the signal Yo and the various folded signals Yf which is best suited for display, the method of the invention first determines, preferably, a "real" number of samples in said range ΔΧ of reference signal X. Said "real" number can be determined from the constant interval. If a main (or folded) signal Yo is saved with a method similar to the VCD method, a real number is obtained by combining the number of saved data relating to transitions, and said "constant" intervals. The number of saved data relating to transitions does not actually correspond to the exact number of samples of the signal Yo. Taking into account the different "constant intervals" during which the main (or folded) signal Y0 is constant, it is possible to determine an exact number of samples of the corresponding signal and then to choose the one which is best suited for the display, which means the one that has a number of samples that is for example the closest to the number of pixels Npix.

According to another preferred embodiment, the generation of the main signal Y0 is such that it comprises a temporal information such as that described in the preceding paragraph. Therefore, the different folded signals Yf also include such time information.

According to another preferred embodiment, the preferred main data file 10 comprises at least one element having a value, of number n, of said reference signal X (n), and a value of a sample Y0 ( n) of said main signal, at said value, of number n, of said reference signal X. Preferably, the sub-data file 11 comprises at least one element comprising a nth value of said reference signal X (n), and a a value of a sample Yf (n) of the folded signal corresponding to said nth value of said reference signal X. More preferably, the main data file 10 and all the sub-data files 11 are constituted in this manner. Fig. 4 illustrates a sub-data file 11 which is constituted in this manner.

According to a second aspect, the invention relates to a device 20 (or system 20) for storing, from waveform data 100, useful information for a waveform viewer 200. The device 20 of the invention can take different forms or be presented according to different embodiments. Preferably, the device 20 is a computer or a central unit. A personal computer can be used. Figure 5 schematically illustrates a preferred embodiment of the device 20 of the invention, in combination with a data source 1 and a waveform viewer 200.

The device 20 of the invention comprises input means 21 for receiving waveform data 100. Examples of input means 21 are an acquisition card, a USB interface, an Ethernet interface . As shown in FIG. 5, the waveform data 100 is preferably provided by a data source 1. Examples of data sources 1 have been given previously. The device 20 of the invention also comprises a processing means 22 for generating, from these waveform data 100, a main signal Y 0 having No samples: Y o (i) i = 1, ... No. These No samples are Novators of said main signal Y 0, at N 0 values of a reference signal X. Examples of main signal Y 0 and reference signal X have been provided previously and apply here for the device 20 of the invention . Examples of processing means 22 are a processing unit or a computation unit of a computer, an onboard PC, a mini-computer, a tablet, ASIC logic circuits, programmable networks of the FPGA type, processors DSP type, a workstation, a system on a chip, an ASSP type processing. Preferably, the processing means 22 is part of a sensor that collects waveform data 100.

The device 20 of the invention comprises backup means 23 for saving, in a main data file 10, main data relating to the main signal Yo. Examples of backup means 23 are internal or portable hard drives, a USB key, a CD-ROM, another device that is part of cloud computing.

The device 20 of the invention comprises a preprocessing means 24 for defining an input signal to be folded, Yin, which is: o if no folded signal from said main signal Yo has been generated before, said signal main Y0, o or a folded signal from said main signal Y0 which has been generated just before.

The pretreatment means 24 and the processing means 22 described above may be identical. Examples of pretreatment means 24 are a processing unit or a computation unit of a computer, logic circuits of ASIC type, programmable networks of FPGA type, microcontrollers, processors of DSP type.

The processing means 25 of the device 20 makes it possible to generate one or more folded signals Yf which is / are an overall reproduction, of lower resolution, of the input signal to be folded, Υ, η, and which has Nf samples which are Nf values of said folded signal Yf, at Nf values of said reference signal X, and which has fewer samples than said input signal to fold Υη. The generation of the folded signal Yf, by the processing means 25, is carried out by carrying out the following steps: a) applying a folding algorithm to substantially all of said input signal to be folded, Yin, in such a way that: samples of said input signal to be folded, Yjn, satisfying fallback rules, are selected; o samples of said input signal to be folded, Yjn, not satisfying said fallback rules, are deleted; b) then grouping the selected samples in step a) to form said folded signal Yf.

The pretreatment means 24, the previously-described processing means 22, and the processing means 25 described in this paragraph may be identical. Examples of processing means are a processing unit or a computing unit of a computer. All the preferred technical features described above, with reference to step (v) of the method according to the first aspect of the invention, apply here by analogy, mutatis mutandis. In particular, various fallback algorithms may be used, for example the three preferred fallback algorithms described above, with reference to step (v) of the method according to the first aspect of the invention.

The device 20 of the invention also comprises backup means 26 for saving, in a sub-data file 11, sub-data relating to a folded signal Yf. The various examples for the main signal Yo and for the folded signals Yf, which have been presented previously, in connection with the method of the invention, apply here for the device 20. Similarly, the different ways of constituting the file 10 and the sub-data file 11, which have been presented previously in connection with the method of the invention, apply here for the device 20. As a general remark, the characteristics described in connection with the The method of the invention applies by analogy, mutatis mutandis, for the device 20, and when they are applicable.

Preferably, the pretreatment means 24, the processing means 25 and the backup means 26 can operate iteratively, in order to define successive input signals to fold Yjn, in order to generate successive folded signals Yf and in order to save such successive folded signals Yf in different sub-data files 11. Preferably, the device 20 thus comprises a sub-unit for detecting when a folded signal Yf has a number of samples which is equal to a predetermined minimum value. The device 20 is then preferably able to stop the iteration process to stop the generation of folded signals Yf.

Main data relating to the main signal Y0 and sub-data relating to the folded signals Yf, respectively saved in the main data file 10 and in the sub-data file 11, can be used thereafter, to display information useful in a waveform viewer 200. In general, a waveform viewer 200 is placed at a user station.

FIG. 6 shows another preferred embodiment of the device 20 of the invention, also in combination with a data source 1 and a waveform viewer 200. In this preferred embodiment, the same backup means 23 are used to save, in a main data file 10, main data relating to the main signal Y0, and to save, in one or more sub-data file (s) 11, sub-data relating to the signal / with signals folded Yf. This preferred device 20 comprises an access unit 27 for choosing between said main signal Yo and said folded signals Yf, one of those having, in a range ΔΧ of said reference signal X, a number of samples which is the most near a number of available pixels Npix in a display window width of a waveform viewer 200, in said range ΔΧ of said reference signal X. Therefore, in accordance with this preferred embodiment, it It is possible to easily access the signal that is best suited for the viewing window selected by a user. Such easy access is possible, in particular thanks to the specific generations and backups of the main signal Y0 and the folded signals Yf. In order to choose the most suitable signal between the main signal Yo and the various folded signals Yf, the access unit 27 must first determine the number of samples in each of these signals, namely the signal Yo and the different signals Yf, in the range ΔΧ. The access unit 27 can then determine a difference value for each signal Yo and for the different signals Yf, which value is the difference between the number of pixels NPiX and the number of samples of each of these signals, in the range ΔΧ. A signal having the smallest difference value is selected by the access unit 27. Since the number of samples of the selected signal is not generally equal to the number of pixels Npix, a processing step is in progress. general applied next, in order to have a signal value selected at each pixel of the display window. The interpolation, which is known to those skilled in the art, is preferably applied. When a compression algorithm is used to back up master data and / or sub-data in master data files 10 and sub-data files 11, decompression of these files is preferably performed before determining a value. of difference for each signal Y0 and for the different signals Yf, difference value which is the difference between the number of pixels Νρ, χ and the number of samples of each of these signals, in the range ΔΧ. Preferably, the access unit 27 and the processing means (22; 25) for generating a main signal Yo and folded signals Yf are one and the same unit. This one and the same unit may also correspond to preprocessing means 24, to define the input signal to fold Yin.

Preferably, a fallback algorithm is applied until the number of samples, in a folded signal Yf, is equal to or less than a threshold. Preferably, this threshold is equal to 1920. More preferably, this threshold is equal to 800. Even more preferably, this threshold is equal to 366.

Preferably, the fallback algorithm is such that each folded signal Yf comprises a local preview information of the main signal Yo.

Preferably, more than one main signal Yo is generated during step (ii) of the method of the invention, or else by the processing means 22 of the device 20 of the invention. Different sets of folded signals Yf are then generated, one set for each main signal Yo.

The present invention has been described in terms of specific embodiments which are provided by way of illustration of the invention and should not be construed as limiting. More generally, those skilled in the art will appreciate that the present invention is not limited by what has been particularly shown and / or described herein in the foregoing. Reference numbers in the claims do not limit the extent of their protection. The use of the verbs "to understand", "to contain", "to be composed of" or any other variant, as well as their respective associations, does not exclude the presence of elements other than those specified. The use of the article "a", "a" or "the", "the" preceding an element does not exclude the presence of a plurality of such elements.

The invention can also be described as follows. The method of the invention comprises the steps of: receiving data 100 of the waveform; generating a main signal Yo; saving main data relating to the main signal Y0; defining an input signal Yin to fold, which is said main signal Yo or a folded signal which has been generated just before; to generate a new folded signal Yf which is an overall reproduction, of lower resolution, of an input signal Yjn to be folded and which has fewer samples than the latter, by following the steps of: applying a fallback algorithm the entire input signal Yin to be folded, so that its samples satisfying fallback rules are selected; that his samples not satisfying the said fallback rules are deleted; then grouping the selected samples to form the new refolded signal Yf; to save subdata relating to said folded signal Yf. Therefore, storage of a waveform is based on envelope coding occurring at each successive foldback level.

Claims (29)

claims A method for storing, from waveform data (100), useful information for a waveform viewer (200) and comprising the steps of: (i) receiving said data of a waveform wave (100); (ii) generating, from said waveform data (100), a main signal Y0 having N0 samples, Y0 (i) i = 1, ... N0, which are N0 values of said main signal Y0, at No values a reference signal X; (iii) backing up primary data relating to said main signal Yo; (iv) defining an input signal to be folded, Yin, which is: - if no folded signal from said main signal Y0 has been generated before, said main signal Yo, - or a folded signal from said main signal Yo that was generated just before; (v) generating a folded signal Yf: - which is a global reproduction, of lower resolution, of said input signal to be folded, Υ, η, - which has Nf samples which are Nf values of said folded signal Yf, at Nf values of said reference signal X, and - which has fewer samples than said input signal to refold, Yjn, by performing the following steps: a) applying a fallback algorithm to substantially all of said input signal to be folded, Yin in such a way that: samples of said input signal to be folded, Yin satisfying fallback rules are selected; that samples of said input signal to be folded, Yin, not satisfying said fallback rules, are deleted; b) then grouping the selected samples in step a) to form said refolded signal Yf, (vi) saving subdata relating to said refolded signal Yf, (vii) optionally repeating steps (iv) to (vi) . 2. Method according to claim 1 characterized in that main data relating to said main signal Y0 are saved in a main data file (10) during step (iii). 3. Method according to claim 1 or 2 characterized in that sub-data relating to said folded signal Yf are saved in a sub-data file (11) during step (vi). 4. Method according to any one of the preceding claims, characterized in that said N0 samples of said main signal Yo represent amplitudes of a digital signal. 5. Method according to any one of claims 1 to 3 characterized in that said No samples of said main signal Yo represent amplitudes of an analog signal. The method according to any one of claims 1 to 3 characterized in that said main signal Yo represents a combination of signals contained in said waveform data (100). 7. Method according to any one of the preceding claims, characterized in that said main signal Yo is generated by using a transition detection method, in such a way that it essentially contains information concerning transitions. 8. Method according to any one of claims 1 to 6 characterized in that said main data relating to said main signal Y0 are saved during step (iii), using a method that records transitions. 9. Method according to any one of the preceding claims, characterized in that said folding algorithm comprises the following steps: - dividing said input signal to be folded, Yin, each portion comprising substantially Np successive samples of said input signal to collapse, Yin, the number Np being £ 3; - apply, for each of said portions, the following steps to select two samples in each of said portions: o a sample, whose absolute value is greater than or equal to that of the other samples, is selected, o another sample, whose value absolute is smaller or equal to that of the other samples, is selected, and o the other samples are deleted. 10. Method according to any one of the preceding claims, characterized in that said folding algorithm comprises the following steps: - dividing said input signal to be folded, Ym, each portion comprising essentially Np successive samples of said input signal to fold, Ym, the number Np being £ 3; applying for each of said portions the following steps for selecting two samples in each of said portions: the sample Yn (q) having the lowest index q is selected; o the sample Ym (p) having the highest index p is selected; o the other samples are deleted. 11. Method according to any one of the preceding claims, characterized in that said folding algorithm comprises the following steps: - dividing said input signal to be folded, Υη, each portion comprising essentially Np successive samples of said input signal to fold, Ym, the number Np being £ 3; - to apply, for each of said portions, the following steps for selecting a sample in each of said portions: o detecting whether a transition occurs among the successive samples of the portion; if so, a sample Yjn (b) of the portion is selected, where b is neither the lowest nor the highest index in the portion, the other samples of the portion are deleted; if not, no sample is selected. The method of any preceding claim for also accessing said useful information for display on a waveform viewer (200), and further comprising the steps of: - providing a range ΔΧ of said reference signal X ; providing a display window width of said waveform viewer (200); providing a NPiX pixel number available for display in said width of said display window; selecting, between said main signal Y0 and said folded signals Yf, one of those having, in said range ΔΧ of said reference signal X, a number of samples closest to said number of pixels Npix. The method of any one of claims 1 to 11, to also access said useful information for display on a waveform viewer (200), and further comprising the steps of: - providing a range ΔΧ of said signal reference X; providing a display window width of said waveform viewer (200); providing a number of pixels Np, x available for display in said width of said display window; selecting, between said main signal Y0 and said folded signals Yf, one of those having, in said range ΔΧ of said reference signal X, a number of samples equal to or greater than said number of pixels Νρ, χ. 14. Method according to any one of claims 1 to 13 characterized in that said main data relating to said main signal Yo are saved in a main data file (10) in step (iii), said data file principals (10) comprising at least one element comprising a nth value of said reference signal, X (n), and a value of a sample, Yo (n), of said main signal, at said nth value of said reference signal X. 15. Method according to any one of claims 1 to 14 characterized in that said sub-data relating to said folded signal Yf are saved in a sub-data file (11) during step (vi), said file sub-data (11) comprising at least one element having an nth value of said reference signal X (n), and a value of a sample Yf (n) of said folded signal, at said nth value of said signal reference X. Apparatus (20) for storing, from waveform data (100), useful information for a waveform viewer (200), and comprising: - input means (21); ) to receive waveform data (100); processing means (22) for generating, from said waveform data (100), a main signal Yo having N0 samples, Yo (i) i = 1, ... N0 which are N values of said main signal Yo, at N0 values of a reference signal X; backup means (23) for storing main data relating to said main signal Y0; pretreatment means (24; 22) for defining an input signal to fold Yin, which is: o if no folded signal from said main signal Yo has been generated before, said main signal Yo, o or else a folded signal from said main signal Y0 and generated just before; processing means (25; 22) for performing the following steps: a) applying a folding algorithm to substantially all of said input signal to be folded, Yin, so that: o samples of said input signal to replicate, Yin, satisfying fallback rules, be selected; o samples of said input signal to be folded, Yjn, not satisfying said fallback rules, are deleted; b) subsequently grouping the selected samples in step a) to form a folded signal Yf; said folded signal Yf • being an overall reproduction, of lower resolution, of said input signal to be folded, Υ, η, • having Nf samples which are Nf values of said folded signal Yf, at Nf values of said reference signal X, and • having fewer samples than said input signal to be folded, Yjn; backup means (26; 23) for saving sub-data relating to said folded signal Yf. 17. Device (20) according to the preceding claim characterized in that said preprocessing means (24), said processing means (25) and said backup means (26) can operate iteratively to define input signals to to fold successive Yin, to generate successive folded signals Yf and to save such successive folded signals Yf. 18. Device (20) according to claim 16 or 17 characterized in that said successive folded signals Yf are saved in different sub-data files (11). The device (20) according to any of claims 16 to 18, further comprising an access unit (27) for selecting between said main signal Yo and said folded signals Yf, one of those having, in a ΔΧ range of said reference signal X, a number of samples that is closest to a number of available Npix pixels in a display window width of a waveform viewer (200). Apparatus (20) according to any one of claims 16 to 18, further comprising an access unit (27) for selecting between said main signal Y0 and said folded signals Yf, one of those having, in a ΔΧ range of said reference signal X, a number of samples that is greater than or equal to the number of NPiX pixels available in a display window width of a waveform viewer (200). 21. Device (20) according to any one of claims 16 to 20 characterized in that said main data relating to said main signal Y0 are saved in a main data file (10). 22. Device (20) according to any one of claims 16 to 21 characterized in that said No samples of said main signal Y0 represent amplitudes of a digital signal. 23. Device (20) according to any one of claims 16 to 22 characterized in that said N0 samples of said main signal Y0 represent amplitudes of an analog signal. 24. Device (20) according to any one of claims 16 to 21 characterized in that said main signal Yo represents a combination of signals contained in said waveform data (100). 25. Device (20) according to any one of claims 16 to 24 characterized in that said processing means (22) are adapted to generate said main signal Yo using a transition detection method, such that said signal main Yo basically contains information about transitions. 26. Device (20) according to any one of claims 16 to 24 characterized in that said backup means (23) are able to save main data relating to said main signal Y0 using a method that records transitions. 27. Device (20) according to any one of claims 16 to 26 characterized in that said processing means (25; 22) are adapted to apply, to form said folded signal Yf, a fallback algorithm comprising the following steps: - divided into portions said input signal to be folded, Yjn, each portion comprising essentially Np successive samples of said input signal to be folded, Yjn, the number Np being £ 3; - apply, for each of said portions, the following steps to select two samples in each of said portions: • a sample, whose absolute value is greater than or equal to that of the other samples, is selected, • another sample, whose value absolute is smaller or equal to that of the other samples, is selected, and • the other samples are deleted. 28. Device (20) according to any one of claims 16 to 27 characterized in that said processing means (25; 22) are adapted to apply, to form said folded signal Yf, a fallback algorithm comprising the following steps: - divided into portions said input signal to be folded, Yjn, each portion comprising essentially Np successive samples of said input signal to be folded, Yin, the number NP being> 3; - applying, for each of said portions, the following steps for selecting two samples in each of said portions: • the sample Yin (q) having the lowest index q is selected; • the sample Υη (ρ) with the highest index p is selected; • the other samples are deleted. 29. Device (20) according to any one of claims 16 to 28 characterized in that said processing means (25; 22) are adapted to apply, to form said folded signal Yf, a fallback algorithm comprising the following steps: - divided into portions said input signal to be folded, Υ, η, each portion comprising substantially Np successive samples of said input signal to be folded, Yin, the number Np being 3 3; - applying, for each of said portions, the following steps to select a sample in each of said portions: • detecting whether a transition occurs among the successive samples of the portion; if so, a sample Yin (b) of the portion is selected, where b is neither the lowest nor the highest index in the portion, the other samples of the portion are deleted; if not, no sample is selected.