Here's one for the digital experts: I do a lot of interviews in which the other party is on a cell phone, and I think I've noticed a difference over time in how they work.
If the call is conducted at a busy time of day, and there are clear signs of bandwidth becoming limited, it used to be that there would be a corresponding number of dropouts in the conversation. Lately (and I call mostly the Chicago and Los Angeles areas) I've detected something different: it sounds like they're just reducing the audio bitrate. It starts to sound like it's coming from a cheap microphone.
In the beginning cellphones were all analogue. The sound quality can vary due to distance between the cellphone and the cellphone tower.
Today cellphones have almost all gone digital. You cannot buy an analogue cellphone anymore and even if you still have an analogue cellphone you might not be able to find a service provide which still runs the old analogue cellphone systems.
Digital cellphone system is almost like digital TV. You either get it
100% or you don't get it at all, because the signals are all in ones and zeros.
The difference in sound quality is most likely due use of "Voice over IP" (people dialing you using Skype or other WiFi connection Apps to avoid airtime charges), or use of low quality Bluetooth headset when talking to you (most cellphones have built-in Bluetooth technology).
VoIP is a frequent problem in recording for articles. That's why I still have a conventional landline at my end, recording directly into the computer. But I also use my cell phone with an earbud microphone into my Sony digital recorder -- the difference is almost undetectable.I don't bother with the in-line adapter on a cell phone anymore.
Having played with different bit rates over the years, it really does sound like a bit-rate reduction to me. I don't think my interviewees are using Bluetooth. I never use it for anything that has to be transcribed, myself.
Ah, well, it's not something that seriously bothers me because it only happens once in a while. I was thinking about it today because I had to transcribe a half-hour conversation myself (which I hate to do), because my transcribing service couldn't make it out. They have a good collection of audio filters, but they couldn't do it. I could, because I know what the other guy was saying.
I studied voice and data communications in 1970 and took a refresher class in the phone system in the 1990's when an interest in infrastructure didn't brand one a potential terrorist. I haven't looked into it since and don't intend to now.
My cellular internet modem drops out completely when the system hits voice call capacity. There are a number of ways they could reduce the audio data rate when that happens such as duplicating frames or dropping the least significant bit, which they do anyway to some extent to transmit the on/off hook signal. That's why the dialup maximum is 56K instead of 64K.
Digital TV drops a lot of data intentionally to stuff 20MB of HD video into 6MB of TV channel. They send a full "key" frame and then only the changes to it, until the next full frame. It works fine when the background is static or changes slowly. It isn't quite fast enough for football play closeups so the TV may display larger solid blocks momentarily.
Speech can be similarly trimmed down to taking and sending one sample every 50 milliseconds or so and just repeating it until the next sample, because most of us can't change the shape of our vocal tracts any faster than that. The intelligibility is still tolerable although JFK demanded and got a higher sound quality digital phone that needed special dedicated lines.
The underlying tech of digital phones dates from a secure channel between Roosevelt and Churchill.
"The voice encoding used the fact that speech varies fairly slowly as the components of the throat move. The system extracts information about the voice signal around 25 times a second."
Interesting. If you've played with recording systems, computer or otherwise, that record digitally and that allow you to select from a menu of bitrates, you've encountered what I'm talking about. Any decent digital recorder today has optional record-time settings, which actually are different bit rates.
AFAIK the landline phone system is locked into the bit rates of G.711 Mu- and A-Law encoding.
A cell phone has to be compatible with a landline, or even an old rotary dial phone.
I don't know exactly how they might compress it, or interpolate dropped samples. The military gear I mentioned transmitted a simplified spectrum analysis of the speech instead of the sequential real-time amplitudes.
Aha! Maybe this is the clue. I should have looked it up but I was lazy and thought one of you guys should know:
"For example, engineers at Nokia found that when they compressed voice data to 5.15 kilobits per second, which cellphones do automatically
"Standard compression rates vary from 12.2 kb/s to 4.75 kb/s, depending on the volume of voice traffic and the strength of the wireless signal. Calls compressed to speeds as low as 7.95 kb/s can still sound almost as good as a landline connection. But beyond that,
"...one proposal would require all TV stations, including LPTV, to give up all spectrum, with subsidized multichannel services replacing over-the-air TV, even after viewers spent a great deal of money on the DTV transition."
I'm on Verizon, both for landline and for cell phones. But what I'm describing isn't happening at my end. I'm using landline for recording interviews, 90% of the time. It's something happening at the other end, when they're using cell phones.
Hey, I'm headed for Chicago this coming Tuesday. If I get a break, I'll give you a call.
That ain't right---you can see how it really works with progressive GIFs. Basically, you send the crude approximation of your signal (image in this case) first, and then if you have bandwidth/time/money/patience you show more detail. A good example is here:
You could think of this as using very low bandwidth aka low bitrate connection in the beginning (sending just few 'bits' of the image), and then increasing the number of sent bits per unit time to get more detail.
Ed was talking about the phone companies using adaptive audio bitrate, which is how the phone company manages their limited aggregate cell tower bandwidth. They basically can serve small number of customers at high bitrate that gives very good sound quality. If, however, they get a lot of cellphones trying to use an oversubscribed tower, they run out of bandwidth. They COULD drop the call (drop from 100% to nothing at all, as you say), but the customers tend to complain loudly about that.
Instead, the phone company came up with a brilliant idea: they switch everyone to a low, crappy bitrate, so they can handle a lot of cellphones at lower quality. It's much harder for the customers to quantify their complaints now ("maybe momma is just mumbling"). Geniuses.
You are talking about "digitizing" voice or video.
I am talking about the "transmission mode" of analogue cellphone and digital cellphone, and the equivalent technology of digital TV and analogue TV.
Canada has phased out analogue TV broadcast a few years ago. All TV broadcast over the airwaves are now digital in Canada.
It is true that when you are transmitting in digital form, the signals are in ones and zeros, so that you either get it clearly or you don't get it at all.
Analog Versus Digital TV: What's the Difference? // How Digital Television Works
DTV is more advanced than the older analog technology. Unlike analog television, which uses a continuously variable signal, a digital broadcast converts the programming into a stream of binary on/off bits?sequences of 0s and 1s. This is the same way that computers store information in data files; each bit represents a small part of the picture, and all the bits combine to reproduce the original picture.
The primary advantage of digital broadcasting is that these binary bits recombine to reproduce an exact copy of the original material. The picture and sound received from a digital transmission are always identical to the original source.
Even better, over-the-air digital signals don?t weaken over distance, as analog signals do. As long as the signal can be received, the picture is perfect, with no degradation or ghosting. Because digital signals are composed of binary bits, a 1 is always a 1, and a 0 is always a 0. There is no fuzziness or snow in the picture, no ghosts caused by interference.
In addition, digital is a more efficient technology. A digital transmission requires less bandwidth than does a similar analog broadcast; this lets local television stations broadcast two, three, or even four digital channels in the space of a single analog channel. This ?multicasting? technology means you?ll receive more variety in programming from your local stations?all delivered with superior digital quality. \\