simulating 'phone' audio

J

j1clark@ucsd.edu

Veteran
I had a need to create a 'phone' sounding portion of dialog. Since it was for a class project, and I didn't really have time to test things out, all I did whas filter the voice differently than the dialog that was from the character 'at this end of the line'.

One idea that I didn't try, what basically to pass my audio through a 'real' phone, and record the result. Is that what people in audio would do? Or would they apply some well known processing techniques?
 
j1clark@ucsd.edu said:
I had a need to create a 'phone' sounding portion of dialog. Since it was for a class project, and I didn't really have time to test things out, all I did whas filter the voice differently than the dialog that was from the character 'at this end of the line'.

One idea that I didn't try, what basically to pass my audio through a 'real' phone, and record the result. Is that what people in audio would do? Or would they apply some well known processing techniques?
Click to expand...

Use the equalizer to put a sharp low-cut filter at about 380 Hz and a sharp high-cut at about 3400 Hz. This pretty well mimics the bandwidth that the phone system is set to deliver.
 
Steve House said:
Use the equalizer to put a sharp low-cut filter at about 380 Hz and a sharp high-cut at about 3400 Hz. This pretty well mimics the bandwidth that the phone system is set to deliver.
Click to expand...

Right, but you may want to play around with those frequencies to get the best results on a given voice. I usually add a midrange boost too, around 1KHz or so, but you'd need at least a 3-band EQ for that. Heavy post-EQ compression usually helps as well since the EQ can make the voice very peaky dynamically. So crush away... : )
 
Ted Spencer said:
Right, but you may want to play around with those frequencies to get the best results on a given voice. I usually add a midrange boost too, around 1KHz or so, but you'd need at least a 3-band EQ for that. Heavy post-EQ compression usually helps as well since the EQ can make the voice very peaky dynamically. So crush away... : )
Click to expand...

I'll try this. I did do the 'band pass' type filter, didn't quite use the parameters Steve had, but basically it sounded more like a rolled off 'room' voice, than an over the phone voice.

I was also thinking about using something that would allow me to set a threshhold and attack time, to cut into 'p', 'b', and other phonemes that get munched in phones.

The other thought was to add some 'different' background noise/tone to relocate the sound as well.
 
j1clark@ucsd.edu said:
I'll try this. I did do the 'band pass' type filter, didn't quite use the parameters Steve had, but basically it sounded more like a rolled off 'room' voice, than an over the phone voice.
Click to expand...

Try narrowing the passed range (raise the low cut freq and lower the hi cut freq) too. Play around with them and don't be shy about making it extreme.

I was also thinking about using something that would allow me to set a threshhold and attack time, to cut into 'p', 'b', and other phonemes that get munched in phones.
Click to expand...

Dunno about that, but a good compressor is key. Like I said, crush...

The other thought was to add some 'different' background noise/tone to relocate the sound as well.
Click to expand...

Ambience-wise, sure, like traffic in the background, but mainly what you need is the filtering. That will say "phone" immediately, and it will probably be enough. The bandpassed noise may not read very clearly once all the filtering is applied.
 
Or two other options. iZotope makes a plugin called Vinyl that is meant to mimic records, mostly in not great shape. It will do a lot and add in noise if you want. AND it's free and I think cross platform. For just EQ I like the Roger Nichols (Elemental Audio) plugin Uniqualizer. It has presets for phone filtering that adds comb filtering to bandpass and sounds pretty good. It is not free but it is a very flexible filter.
 
Noiz2 said:
Or two other options. iZotope makes a plugin called Vinyl that is meant to mimic records, mostly in not great shape. It will do a lot and add in noise if you want. AND it's free and I think cross platform. For just EQ I like the Roger Nichols (Elemental Audio) plugin Uniqualizer. It has presets for phone filtering that adds comb filtering to bandpass and sounds pretty good. It is not free but it is a very flexible filter.
Click to expand...

Given my budget... I'll be looking at the freebees... and I may have a scratchy vinyl need coming up. With that in mind, I've been looking at Audacity, but didn't recall a comb filter being one of the set of operations... perhaps there's a plug-in...

I think I'll record a test clip through a 'real' phone so I have some sort of handy reference...
 
Speaking from a theoretical POV, the time-division multiplexing on a phone line is pretty similar to using a low digital sample rate, i.e. instead of 44.1 or 48 kHz, consider trying 11kHz. Of course, the biggest change will be losing high-end, which is what your low-pass filter will be doing anyway.
 
ProfessorU said:
Speaking from a theoretical POV, the time-division multiplexing on a phone line is pretty similar to using a low digital sample rate, i.e. instead of 44.1 or 48 kHz, consider trying 11kHz. Of course, the biggest change will be losing high-end, which is what your low-pass filter will be doing anyway.
Click to expand...

I did exactly the above with Audacity - resample @ 11kHz mono - gives a very good result. Also change your bit depth to 16
 
j1clark@ucsd.edu said:
Given my budget... I'll be looking at the freebees... and I may have a scratchy vinyl need coming up. With that in mind, I've been looking at Audacity, but didn't recall a comb filter being one of the set of operations... perhaps there's a plug-in...

I think I'll record a test clip through a 'real' phone so I have some sort of handy reference...
Click to expand...

Izotope Vinyl will do the trick for a vinyl record effect - you can even specify what era of record player it is and how much dust is on the record!- and best of all, it's free for the download.
 
Of course you could do it the No Cost way and path of least resistance by using your camera to record the actor who call's you on a hands free/speaker phone. No filters, no mucking around and you record the audio from the speaker. Bam!
 
j1clark@ucsd.edu said:
Given my budget... I'll be looking at the freebees... and I may have a scratchy vinyl need coming up. With that in mind, I've been looking at Audacity, but didn't recall a comb filter being one of the set of operations... perhaps there's a plug-in...

I think I'll record a test clip through a 'real' phone so I have some sort of handy reference...
Click to expand...


You can make a comb filter by putting a very short delay on the clip. mix dry/wet to taste. I havn't done so in Audacity, but I'm sure there's a way.

Or, download a free comb filter plug in.
 
I just learned that Audio Ease, the makers of the AltiVerb convolution reverb softwarre, also has a product called "Speakerphone" that alters aaudio to make it sound like its coming through a wide variety of devices and in a huge variety of locations. Need it to sound like its coming through the PA in Grand Central Station mixed in with crowd walla, trains arriving, and police sirens pulling up to the curb? Just dial it in on the presets! Amazing online demo.
 
Last edited:
steadicamsteel said:
Of course you could do it the No Cost way and path of least resistance by using your camera to record the actor who call's you on a hands free/speaker phone. No filters, no mucking around and you record the audio from the speaker. Bam!
Click to expand...

Point one... that takes all the fun out of things...

Point two... I will probably record the voice track I'm working with through a real(tm) phone.

Point three... this is a learning exercise to become more familiar with 'things audio'.

Most of my signal processing activities to date have been 'how to extract a signal from shi*'... and not in the audio world... but in RF... image processing, wyrd sensor processing, whatever...

So the idea of creating audio to specific requirements is new to me, at least in so far as what tools and processing are available...


I'll be looking at some of the suggests from the others to see what I can do with the tools I can get a hold of... and I'll probably post my results at some point.
 
Steve House said:
I just learned that Audio Ease, the makers of the AltiVerb convolution reverb softwarre, also has a product called "Speakerphone" that alters aaudio to make it sound like its coming through a wide variety of devices and in a huge variety of locations. Need it to sound like its coming through the PA in Grand Central Station mixed in with crowd walla, trains arriving, and police sirens pulling up to the curb? Just dial it in on the presets! Amazing online demo.
Click to expand...

I don't have it, yet, but I have heard very good things about it from folks in the biz.
 
You must log in or register to reply here.
Back
Top