I am trying to encode 2 videos side by side, sync'd by the audio of the 2 clips. I can successfully encode the 2 videos side by side and select the audio from one of the input streams. However the system we are using to record the 2 videos does not start and stop the recording at the same time (could be up to a second different between cameras). Basically we are using a CCTV system to capture what's going on in a room from multiple angles. We export the 2 cameras between 2 timestamps and due to the way the system records the videos the start of the 2 clips are not the same point in time.
e.g. Export videos between 09:00:00:000 and 09:10:00:000
Video 1 - exports from 08:59:59:123 to 09:10:00:123
Video 2 - exports from 08:59:59:789 to 09:10:00:789
Therefore when video 1 and video 2 are stitched together side by side, they are out of sync by 666ms (which is very noticeable in the encoded video)
Both input streams have (near) identical audio and are both in the exact same format. We are currently placing these videos into Premiere Pro and syncing these videos by the audio and exporting them side by side, however we have a project where we need to do a lot of these in quick succession and this is not really an option. We need to look at scripting this.
Does anyone know if FFMPEG can do this? Or anything else?
Any info would be greatly appreciated.
You can use audio-offset-finder in bash file to calculate offset, cut of the head from one of the video, stitch them together ( like stated here ).
You would need to extract audio streams into separate files and use finder to calculate offset.
offset=`audio-offset-finder --find-offset-of file1.wav --within file2.wav`
How to get file information like sampling rate, bit rate etc of .raw audio files using terminal in linux? Soxi works for .wav files but it isn't working for .raw.
If your life depended on discovering an answer you could make some assumption to tease apart the unknowns ... however there is no automated way since the missing header would give you the easy answers ...
The audio analysis tool called audacity allows you to open up a RAW file, make some guesses and play the track
http://www.audacityteam.org
In audacity goto File -> Import -> Raw Data...
Above settings are typical for audio ripped from a CD ... toy with trying stereo vs mono for starters.
Those picklist widgets give you wiggle room to discover the format of your PCM audio given that the source audio is something when properly rendered is recognizable ... would be harder if the actual audio was noise
However if you need a programmatic method then rolling your own solution to ask those same questions which appear in above window is possible ... is that what you need or will audacity work for you ? We can go down the road of writing code to play off the unknowns mentioned in #Frank Lauterwald's comment
To kick start discovering this information programmatically, if the binary raw audio is 16 bit then each audio sample (point on the audio curve) will consume two bytes of your PCM file. For mono audio then the following two bytes would be your next sample, however if its stereo then these two following bytes would be the sample from the other channel. If more than two channels then just repeat. Typical audio is little endian. Sampling rate is important when rendering the audio, not when programmatically parsing raw bytes. One approach would be to create an output file with a WAV header followed by your source PCM data. Populate the header with answers from your guesswork. This way you could listen to this output file to help confirm your guesses.
Here is a sample 500k mono PCM audio file signed 16 bit which can be imported into audacity or used as input to rolling your own identification code
The_Constructus_Corporation_Long_Street-ycexQvMy03k_excerpt_mono.pcm
Our are looking for a way to develop an automated system for our ALSA/Linux based product to test our codex card automatically. For this we envision the following steps:
1. Play an existing known wave file
2. Record sound from the MIC.
3. If both speaker and MIC are working fine (as they are placed close to each other) the recorded file should have some version of the played sound
4. Automatically analyse if the recorded file has some version of sound file.
Question: Is there an API, which can help analyse the recorded to find sound level or other parameters to help detect that both MIC and speaker in proximity are working fine.
I am using pygame to play .wav files and want to change the pitch of a particular .wav file as each level in my game progresses. To explain, my game is a near copy of the old Oric1 computer OricMunch Pacman game, where there are a few hundred pills to be munched on each level, and for every pill that is munched a short sound is played, with the pitch of the sound increasing slightly for each pill eaten/munched.
Now here is what I have tried:
1) I have used pythons wave module to create multiple copies of the sound file, each newly created file having a slight increase in pitch (by changing the 3rd parameter in params() the framerate, sometimes referred to as the sample frequency) for each cycle of a for loop. Having achieved this, I could then within the loop create multiple sound objects to add to a list, and then index through the list to play the sounds as each pill is eaten.
The problem is even though I can create hundreds of files (using the wave module) that play perfectly with their own unique pitches when played using windows media player, or even pythons winsound module, pygame does not seem to interpret the difference in pitch.
Now interestingly, I have downloaded the free trial version of Power Sound Editor which has the option to change the pitch, and so I’ve created just a few .wav files to test, and they clearly play with different pitches when played in pygame.
Observations:
From printing the params in my for loop, I can see that the framerate/frequency is changing as intended, and so obviously this is why the sounds play as intended through windows media player and winsound.
Within pygame I suspect the reason they don’t play with different pitches is because the frequency parameter is fixed, either to the default settings or via the use of pygame.mixer.pre_init, which I have indeed experimented with.
I then checked the params for each .wav file created by the Power Sound Editor, and noticed that even though the pitch sound was changing, the frequency stayed the same, which is not totally surprising since you have to select 1 of 3 options to save the files, either 22050, 44100 or 96000Hz
So now I thought time to check out the difference between pitch and frequency specifically in relation to sound, since I thought they were the same. What I found was it seems there are two principle aspects of sound waves: 1) The framerate/frequency And 2) The varying amplitude of multiple waves based on that frequency. Now I far from clearly understand this, but realise the Power Sound Editor must be altering the shape/pitch of the sound by manipulating the varying amplitude of multiple waves, point 2) above, and not by changing the fundamental frequency, point 1) above.
I am a beginner to python, pygame and programming in general, and have tried hard to find a simple way to change sound files to have gradually increasing pitches without changing the framerate/fundamental frequency. If there’s a module that I can import to help me change the pitch by manipulating the varying amplitude of mutiple waves (instead of changing the framerate/sample frequency which typically is either 22050 or 44100Hz), then it needs to take relatively no time at all if being done on the fly in order to not slow the game down. If the potential module opens, changes and then saves sound files, as opposed to altering them on the fly, then I guess it does not matter if it’s slow because I will just be creating the sound files so I can create sound objects from them in pygame to play.
Now if the only way to achieve no slow down in pygame is to create sound objects from sound files as I have already done, and then play them, then I need a way to manipulate the sound files like the Power Sound Editor (again I stress not by changing the framerate/sample frequency of typically 22050 or 44100) and then save the changed file.
I suppose in a nut shell, if I could magically automate Power Sound Editor to produce 3 to 4 hundred sound files without me having to click on the change pitch option and then save each time, this would be like having my own python way of doing it.
Conclusion:
Assuming creating sound objects from sound files is the only way not to slow my game down (as I suspect it might be) then I need the following:
An equivalent to the python wave module, but which changes the pitch like Power Sound Editor does, and not by changing the fundamental frequency like the wave module does.
Please can someone help me and let me know if there’s a way.
I am using python 3.2.3 and pygame 1.9.2
Also I’m just using pythons IDLE and I’m not familiar with using other editors.
Also I’m aware of Numpy and of various sound modules, but definitely don’t know how to use them. Also any potential modules would need to work with the above versions of python and pygame.
Thank you in advance.
Gary Townsend.
My Reply To The First Answer From Andbdrew Is Below:
Thank you for your assistance.
It does sound like changing the wave file data rather than the wave file parameters is what I need to do. For reference here is the code I have used to create the multiple files:
framerate = 44100 #Original .wav file framerate/sample frequency
for x in range(0, 25):
file = wave.open ('MunchEatPill3Amp.wav')
nFrames = file.getnframes()
wdata = file.readframes(nFrames)
params = file.getparams()
file.close()
n = list(params)
n[0] = 2
n[2] = framerate
framerate += 500
params = tuple(n)
name = 'PillSound' + str(x) + '.wav'
file = wave.open(name, 'wb')
file.setparams(params)
print(params)
file.writeframes(wdata)
file.close()
It sounds like writing different data would be equivalent or similar to how the Power Sound Editor is changing the pitch.
So please can you tell me if you know a way to modify/manipulate wdata to effectively change the pitch, rather than alter the sample rate in params(). Would this mean some relatively simple operation applied to wdata after it’s read from my .wav file. (I really hope so) I’ve heard of using numpy arrays, but I have no clue how to use these.
Please note that any .wav files modified in the above code, do indeed play in Python using winsound, or in windows media player, with the pitch increase sounding as intended. It’s only in Pygame that they don’t.
As I’ve mentioned, it seems because Pygame has a set frequency (I guess this frequency is also sample rate), that this might be the reason the pitch sounds the same, as if it wasn’t increased at all. Whereas when played with e.g. windows media player, the change in sample rate does result in a higher sounding pitch.
I suppose I just need to achieve the same increase in pitch sound by changing the file data, and not the file parameters, and so please can you tell me if you know a way.
Thank you again for helping with this.
To Summarise My Initial Question Overall, Here It Is Again:
How do you change the pitch of a .wav file without changing the framerate/sample frequency, by using the python programming language, and not some kind of separate software program such as Power Sound Editor?
Thank You Again.
You should change the frequency of the wave in your sample instead of changing the sample rate. It seems like python is playing back all of your wave files at the same sample rate (which is good), so your changes are not reflected.
Sample rate is sort of like meta information for a sound file. Read about it at http://en.m.wikipedia.org/wiki/Sampling_rate#mw-mf-search .
It tells you the amount of time between samples when you convert a continuous waveform into a discrete one. Although your (ab)use of it is cool, you would be better served by encoding different frequencies of sound in your different files all at the same sample rate.
I took a look at the docs for the wave module ( http://docs.python.org/3.3/library/wave.html ) and it looks like you should just write different data to your audio files when you call
Wave_write.writeframes(data)
That is the method that actually writes your audio data to your audio file.
The method you described is responsible for writing information about the audio file itself, not the content of the audio data.
Wave_write.setparams(tuple)
"... Where the tuple should be (nchannels, sampwidth, framerate, nframes, comptype, compname), with values valid for the set*() methods. Sets all parameters... " ( also from the docs )
If you post your code, maybe we can fix it.
If you just want to create multiple files and you are using linux, try SoX.
#!/bin/bash
for i in `seq -20 10 20`; do
sox 'input.wav' 'output_'$i'.wav' pitch $i;
done
I want to generate .mp4 file using Direct show samples (AMCap). But i don't know how to implement this.
Can anyone please help me about this?
Thanks in advance,
Dhaval Kariya
AMCap Sample captures and displays video. No encoding and choices of multiplexing into files (only basic capture/recording through a basically obsolete helper interface).
Video capture application.
This sample application demonstrates the following tasks related to
audio and video capture:
Capture to a file
Live preview
Allocation of the capture file
Display of device property pages
Device enumeration
Stream control
The items above might be confusing as they mention capture and file allocation. This is a trail of 15 years old history when file capture was a big deal. The helper object to initialize capture targets AVI and ASF/WMV only, you neither can extend it to support other formats, nor you need to.
You need to check how to store video/audio into files (see below) and follow the same steps in building the pipeline with MPEG-4 encoders and multiplexer. You will need to use a third party MPEG-4 multiplexer for MP4 file format because Windows does not provide you with such out-of-the-box usable component.
See:
Capturing Video to a File
Free DirectShow Mpeg-4 Filters