For example, I have 100 Byte PCM buffer and want to increase it to 300 Byte.
what i tried:
asume original buffer was 9, 4, 1, 7, 5
insert 0 - 9 0 0 4 0 0 1 0 0 7 0 0 5 0 0
average - 9 7 5 4 3 2 1 3 5 7 7 6 5 5 5
insert 0 in back - 9 4 1 7 5 0 0 0 0 0 0 0 0 0 0
They all had weird noise in result audio file.
How can I change length of buffer without effect sound?
Is there any formula I can use?
Usually linear interpolation works. What is the bit-resolution of your PCM file? If it is 16 bits (pretty typical), you'll have to first convert two bytes into a single value before applying the interpolation, and then disassemble the values back to bytes. You will need to know the byte order, as it can be either little-endian or big-endian.
EDIT: I should have added that the pitch will drop with this method of lengthening the file, unless the playback frame rate increases. To stretch out a sound in time without affecting its pitch is considerably more complicated.
Related
I have a program that will ask the user for a number as many times as they want so long as it is within the range 0-10.
I want to then output the frequency of each input for that range including if the number wasn't entered by the user at all. So by default it would be 0.
Here's an example of what I'm looking for.
Input:
1
5
1
2
5
8
9
5
5
Wanted Output:
0 - 0
1 - 2
2 - 0
3 - 0
4 - 0
5 - 4
6 - 0
7 - 0
8 - 1
9 - 1
10 - 0
So far I have a list that stores the users inputs, and a dictionary with keys going from 0-10, with each having a value of 0. I've found code that gives the frequency of the users input but it's unordered and only for the numbers inputed. I'm not sure what to do next.
I'm currently looking through the DEFLATE compression specification and am confused about this part:
0 - 15: Represent code lengths of 0 - 15
16: Copy the previous code length 3 - 6 times.
The next 2 bits indicate repeat length
(0 = 3, ... , 3 = 6)
Example: Codes 8, 16 (+2 bits 11),
16 (+2 bits 10) will expand to
12 code lengths of 8 (1 + 6 + 5)
17: Repeat a code length of 0 for 3 - 10 times.
(3 bits of length)
18: Repeat a code length of 0 for 11 - 138 times
(7 bits of length)
If I'm understanding correctly, 0-15 are the lengths of the Huffman codes for the code length sequences. However, I do not understand what 16-18 is supposed to be. Thanks for your help!
The 16-18 codes are instructions to the decoder to generate several lengths, either zeros or the repeats of the last length.
So for example:
18(12) 14 4 3 3 3 4 4 5 17(3) 5 16(9) 7
becomes:
0 0 0 0 0 0 0 0 0 0 0 0 14 4 3 3 3 4 4 5 0 0 0 5 5 5 5 5 5 5 5 5 5 7
where the numbers in parentheses are coded as 7, 3, and 2 bits respectively immediately after the Huffman code for that symbol.
I am trying to understand the the MCTAL output of a spherical TMESH tally. What I want is to create one tally bin that has the following boundaries 1.9 cm and 2.1 cm in the radial direction, 88 to 92 degrees in theta and 180 to 360 degrees in the phi direction. my input for the tally is
C tally card spherical mesh energy tally
TMESH
SMESH1:p DOSE 1 1 1 1.0 PEDEP MFACT 1 1 0 1.0
CORA1 1.9 2.1
CORB1 88 92
CORC1 180 360
Now what I expect is one result for that volume what I get are eight values as shown below.
ntal 1
1
tally 1 -1 -3
0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
f 4 0 1 2 2
1.90000E+00 2.10000E+00
0.00000E+00 8.80000E+01 9.20000E+01
0.00000E+00 1.80000E+02 3.60000E+02
d 1
u 1
s 2
m 1
c 1
e 1
t 1
vals
5.57481E-04 0.0067 7.68088E-09 0.0493 8.24471E-03 0.0046 1.38395E-07 0.0639
5.53931E-04 0.0046 7.44313E-09 0.0287 8.24244E-03 0.0042 1.27868E-07 0.0553
I am assuming that these eight vals correspond to the eight points that that are listed under f. Does TMESH only give one values for individual points on a grid or can it be used to create a volume within which to obtain a result? lastly to what points do what vals correspond to ?
The matrix bellow the vals is true value of your meshtally result.
but
you must load data to Matlab and reshape it to your mesh tally matrix
With your SMESH setup you score both dose and energy deposition. This causes two bins along the segment axis (the "s 2" record in your mctal). Then, you have only 1 bin along the radial direction (1.9-2.1 cm) and actually TWO bins along each of the angular directions (0-88, 88-92, and 0-180, 180-360) which sums up to 2^3 = 8 bins. The mctal file format is described in the manual: it'a 11-dimension loop. In your case only the s, j and k axes are divided, so it's actully a 3D loop (in this exact order: s being the outer, k - the inner loop). Therefore the value for your volume is either the 4th (1.38395E-07 0.0639) or last (1.27868E-07 0.0553) record depending on whether you need dose or energy deposition.
I have a bag-of-words representation of a corpus stored in an D by W sparse matrix word_freqs. Each row is a document and each column is a word. A given element word_freqs[d,w] represents the number of occurrences of word w in document d.
I'm trying to obtain another D by W matrix not_word_occs where, for each element of word_freqs:
If word_freqs[d,w] is zero, not_word_occs[d,w] should be one.
Otherwise, not_word_occs[d,w] should be zero.
Eventually, this matrix will need to be multiplied with other matrices which might be dense or sparse.
I've tried a number of methods, including:
not_word_occs = (word_freqs == 0).astype(int)
This words for toy examples, but results in a MemoryError for my actual data (which is approx. 18,000x16,000).
I've also tried np.logical_not():
word_occs = sklearn.preprocessing.binarize(word_freqs)
not_word_occs = np.logical_not(word_freqs).astype(int)
This seemed promising, but np.logical_not() does not work on sparse matrices, giving the following error:
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all().
Any ideas or guidance would be appreciated.
(By the way, word_freqs is generated by sklearn's preprocessing.CountVectorizer(). If there's a solution that involves converting this to another kind of matrix, I'm certainly open to that.)
The complement of the nonzero positions of a sparse matrix is dense. So if you want to achieve your stated goals with standard numpy arrays you will require quite a bit of RAM. Here's a quick and totally unscientific hack to give you an idea, how many arrays of that sort your computer can handle:
>>> import numpy as np
>>> a = []
>>> for j in range(100):
... print(j)
... a.append(np.ones((16000, 18000), dtype=int))
My laptop chokes at j=1. So unless you have a really good computer even if you can get the complement (you can do
>>> compl = np.ones(S.shape,int)
>>> compl[S.nonzero()] = 0
) memory will be an issue.
One way out may be to not explicitly compute the complement let's call it C = B1 - A, where B1 is the same-shape matrix completely filled with ones and A the adjacency matrix of your original sparse matrix. For example the matrix product XC can be written as XB1 - XA so you have one multiplication with the sparse A and one with B1 which is actually cheap because it boils down to computing row sums. The point here is that you can compute that without computing C first.
A particularly simple example would be multiplication with a one-hot vector. Such a multiplication just selects a column (if multiplying from the right) or row (if multiplying from the left) of the other matrix. Meaning you just need to find that column or row of the sparse matrix and take the complement (for a single slice no problem) and if you do this for a one-hot matrix, as above you needn't compute the complement explicitly.
Make a small sparse matrix:
In [743]: freq = sparse.random(10,10,.1)
In [744]: freq
Out[744]:
<10x10 sparse matrix of type '<class 'numpy.float64'>'
with 10 stored elements in COOrdinate format>
the repr(freq) shows the shape, elements and format.
In [745]: freq==0
/usr/local/lib/python3.5/dist-packages/scipy/sparse/compressed.py:213: SparseEfficiencyWarning: Comparing a sparse matrix with 0 using == is inefficient, try using != instead.
", try using != instead.", SparseEfficiencyWarning)
Out[745]:
<10x10 sparse matrix of type '<class 'numpy.bool_'>'
with 90 stored elements in Compressed Sparse Row format>
If do your first action, I get a warning and new array with 90 (out of 100) nonzero terms. That not is no longer sparse.
In general numpy functions do not work when applied to sparse matrices. To work they have to delegate the task to sparse methods. But even if logical_not worked it wouldn't solve the memory issue.
Here is an example of using Pandas.SparseDataFrame:
In [42]: X = (sparse.rand(10, 10, .1) != 0).astype(np.int64)
In [43]: X = (sparse.rand(10, 10, .1) != 0).astype(np.int64)
In [44]: d1 = pd.SparseDataFrame(X.toarray(), default_fill_value=0, dtype=np.int64)
In [45]: d2 = pd.SparseDataFrame(np.ones((10,10)), default_fill_value=1, dtype=np.int64)
In [46]: d1.memory_usage()
Out[46]:
Index 80
0 16
1 0
2 8
3 16
4 0
5 0
6 16
7 16
8 8
9 0
dtype: int64
In [47]: d2.memory_usage()
Out[47]:
Index 80
0 0
1 0
2 0
3 0
4 0
5 0
6 0
7 0
8 0
9 0
dtype: int64
math:
In [48]: d2 - d1
Out[48]:
0 1 2 3 4 5 6 7 8 9
0 1 1 0 0 1 1 0 1 1 1
1 1 1 1 1 1 1 1 1 0 1
2 1 1 1 1 1 1 1 1 1 1
3 1 1 1 1 1 1 1 0 1 1
4 1 1 1 1 1 1 1 1 1 1
5 0 1 1 1 1 1 1 1 1 1
6 1 1 1 1 1 1 1 1 1 1
7 0 1 1 0 1 1 1 0 1 1
8 1 1 1 1 1 1 0 1 1 1
9 1 1 1 1 1 1 1 1 1 1
source sparse matrix:
In [49]: d1
Out[49]:
0 1 2 3 4 5 6 7 8 9
0 0 0 1 1 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 1 0
2 0 0 0 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0 1 0 0
4 0 0 0 0 0 0 0 0 0 0
5 1 0 0 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0 0 0 0
7 1 0 0 1 0 0 0 1 0 0
8 0 0 0 0 0 0 1 0 0 0
9 0 0 0 0 0 0 0 0 0 0
memory usage:
In [50]: (d2 - d1).memory_usage()
Out[50]:
Index 80
0 16
1 0
2 8
3 16
4 0
5 0
6 16
7 16
8 8
9 0
dtype: int64
PS if you can't build the whole SparseDataFrame at once (because of memory constraints), you can use an approach similar to one used in this answer
I am working on a school project, which is a simulation of logical gates. I can implement and run the simulation with ease, but i need help with showing the output.
Right now, i print everything to the console, like this:
sample frequency: 50
###############################################
IN NOT(1) OUT
IN1:0 IN1:3 IN1:5
IN2:0 IN2:0 IN2:0
OUT:3 OUT:5 OUT:0
0 1 -1 -1
50 1 -1 -1
100 1 0 0
150 0 0 0
200 1 1 1
250 1 0 0
300 1 0 0
350 1 0 0 (IN = 1, delay is 1 so we can see
400 0 0 0 the correct output of NOT element in line 400 <-> 350 + 1*50)
450 1 1 1
500 1 0 0
550 1 0 0
600 1 0 0
650 0 0 0
700 0 1 1
750 1 1 1
800 1 0 0
850 1 0 0
900 1 0 0
950 1 0 0
1000 1 0 0
on the left, there is the simulation time (step). In each step, the values are printed out and new set of inputs is generated.
where there is -1, this means undefined output.
The 3rd row ( IN NOT(1) OUT ) means that there are 3 elements, 1 input, 1 NOT gate and an output. The value in brackets means the delay of the element, so an element with delay value of X will show the correct output after X*sample_freq (excluding the 0 time).
The rows after mean:
IN1 - the index of the node that is read as input 1
IN2 - the index of the node that is read as input 2
OUT - the index of the output node
In this situation, the IN is giving its output to node #3. The NOT element reads its input from node #3 and gives some output to node #5. The overall output of this system is the OUT element, which reads from #5.
Here is the file that specifies the topology:
3 (number of elems)
IN 0 0 3 (no inputs for input element obviously)
NOT 3 0 5 (reads from #3 and outputs to #5)
OUT 5 0 0 (reads from #5 and this is the end point of the system)
There can obviously be more elements, IN's and OUT's, but lets stick to this for the sake of simplicity.
And what i want to see as the result is: X-axis tells the simulation time (0 - 1000, step is 50), y axis tells the output value of each element in the system and the elements write their output one above the other, see this picture as an example.
Can you tell me how to create this kind of gnuplot script, that transforms the output of my application into the desired plot?
Thank you!
ok, I have found a solultion myself, here it is:
first, I had to transform the output of the app a bit, so that it looks like this:
0 1 2 4
49 1 2 4
50 1 2 4
99 1 2 4
100 0 2 4
149 0 2 4
150 0 3 5
199 0 3 5
200 1 3 5
249 1 3 5
250 1 2 4
299 1 2 4
300 0 2 5
349 0 2 5
350 1 3 5
399 1 3 5
400 0 2 4
449 0 2 4
450 1 3 5
499 1 3 5
the extra sim time steps make the edges look almost square, I also separated each column by 2 (added 0 to column #2, added 2 to column #3, added 4 to column #4 and so on), so that it is drawn one above each other and the simple command to plot this is:
plot 'out.txt' using 1:2 with lines, 'out.txt' using 1:3 with lines, 'out.txt' using 1:4 with lines
plus some set xtics, set ytics and other cosmetic stuff
now I have to deal with naming the lines with the names of the elements and voila.