Does Background Noise Affect Audio Watermark Detection? (Including Motion and Music Genres)
A valid concern when considering an audio watermark for your music, stream or broadcast is how robust the watermark is in harsh environments.
The first thing to know is that different watermark technologies have varying CDR’s (correct detection rates) in different environments.
At Intrasonics, we use echo-modulation audio watermarking, enabling accurate detection in environments that include background noise and motion.
In this post, we are going to look at the following variables:
- Background or crowd noise
- Music Genres
Let’s discuss the challenges these variables pose for an audio watermark and some solutions.
What Environment is Best for Detection of Watermarked Audio?
When you invest in watermarking for your audio files, you want to know that they will be easily detected when using them.
Contact our team today for a free demo of our robust watermarking technology that stands up to harsh audio environments.
First, let’s give some consideration to the best environments for this to be possible.
There are essentially two main rules that you need to think about when it comes to detecting audio watermarks:
- Detection of audio watermarks will work best, the closer you or the user is to the sound source, whether it is a radio, computer, tablet, phone, or television.
- Detection of audio watermarks also works best when there are limited or no loud and disruptive noises in the background.
How Does Crowd Noise Affect Audio Watermark Detection?
Taking the point we made above a little further, you may wonder what impact, if any, that crowd noise can have on the detection of audio watermarks.
According to some tests conducted in the Journal on Applied Signal Processing, volume levels of ambient noise and crowd noise can affect how well audio watermarks are detected.
Note that it largely depends on the phone or device location and the crowd noise volume.
The study conducted also used Sonic Watermarking technology, rather than using Intrasonics more robust, echo-modulation watermarking.
- For example, if the crowd noise is 8dB louder than the audio containing the watermark, there should be no real problem detecting the watermark. However, it is worth noting that some of the audio speech is likely to become harder to hear and understand.
- If the crowd noise is more than 12dB louder than the audio containing the watermark and the full audio clip is hard to hear and understand, it will be harder to detect the audio watermark. That does not necessarily mean it will not detect it at all, just that it may take a longer amount of time.
It is also worth noting that the above would only be in a best-case scenario when the audio containing the watermark plays through high-quality speakers and a suitable audio track is played.
So, to have the best chance of success when detecting audio watermarks with a crowd or ambient background noise, speakers’ volume should be set to no more than around -8dB quieter than that noise.
How Does Motion Affect Audio Watermark Detection? (Doppler Effect)
Motion can play a hugely disruptive part in preventing audio watermarks from being successfully detected.
The Doppler effect is one example of this.
The doppler effect involves the idea that audio experiences a kind of shift in frequency when the audio receiver and the sender of the audio are not perfectly still or at least moving at the same speed.
To illustrate this, think about the last time an ambulance drove past you when the siren was sounding when you were stationary.
The siren’s pitch seems to change as the ambulance approaches and continues to shift as it passes by and eventually drives further away.
The audio watermarking type that is most affected by the doppler effect is Spread Spectrum watermarking.
Spread Spectrum watermarking involves the watermark, as the name hints at, being spread across the file’s entire Spectrum.
This method appeals to those wanting to intertwine audio watermarks within their content, making it challenging to separate one from the other.
As an added security, the audio watermark is at the same frequency as the audio file compression algorithms, but this presents problems in audio watermark detection.
For example, suppose the sound source, a speaker, and the device trying to detect the spread spectrum audio watermark are not entirely still.
In that case, the Doppler effect will render any attempts to detect it almost useless.
Watermark Detection Success with Background Noise is Also Dependent on Music Genre
Another problem that can affect the detection of audio watermarks is the genre of music involved.
The same study referenced above also showed that when pop music is used, what is known as the CDR (Correct Detection Rates - the rate at which audio watermarks correctly identified audio), is higher if the noise addition is no more than 20dB.
When using orchestral or classical music, the noise addition was a problem only if it was higher than 35dB.
Why is this the case?
The reason is that orchestral and classical music has a much wider dynamic range than pop music and is also composed of more sections of low volumes.
The quieter sections tend to degrade quicker than louder sections when the additional noise is comparatively louder.
Although the CDRs in the quieter sections can be increased, it sacrifices transparency.
And while using the masking effect of background noise can increase the robustness of the audio watermark, noise without using a masking effect is a problem yet to be solved.
How Intrasonics Watermarking Stacks Up Against These Factors
The audio watermarking used by Intrasonics is the more modern, robust and efficient echo-modulation based technology.
Echo modulation is specially devised to be more robust against the factors we have outlined above that can affect how easily audio watermarks are detected.
For example, noise, distance, the Doppler effect and motion are less of a problem when using echo-modulation than other forms of watermarking, such as Spectrum spread watermarking.
As we have shown, background noise can affect audio watermark detection and is a problem you will want to overcome.
If you choose Intrasonics echo-modulation audio watermarking, you are going to benefit from one of the most robust solutions on the market.
Our echo modulation watermarking technology is more capable of resisting the disruptive effects of noise, not being close enough to the sound source and the Doppler effect.
Although it is hard to remove the problem entirely, we can help to increase watermark detection accuracy noticeably. See how it works here
Don’t take our word for it. Contact our team today for a free demo of our technology.
< Previous Page