Batboard

Hello All

I have 2 mototrbo radios and like many others i think that the digital audio on the radios sounds a bit robotic, because the audio i so heavly compressed in the DVSI vocoder ambe +2 but the other day i happend i listen to an audio sample of both the mototrbo and the icom d-star after one another.
and it sounds to me like the d-star audio dont have the same robotic voice compare to mototrbo and like that the d-star has a is more natural reproduced voice..
are the anyone there can take time to listen to the audio clips and give my there opinion ?

i have looked at the spec. for both mototrbo and d-star and they both use the AMBE + 2 vocoder from DVSI with 2400bps voice data + 1200 bps FEC

anyone have any good explanation how that can be ?
maybee a more recent firmware update may change that ?

mototrbo audio: http://www.hamradio-dv.org/mototrbo/mototrbo.wav

D-star audio : http://www.hamradio-dv.org/icom/kb6cus-id1.wav

Maelv....

From listening to the clips, I much prefer the final digitized audio product from the Icom/D-Star. Then again, I prefer the analog audio from my JT1000 rx over the analog audio from my HT1550XLS. I'm just guessing that Motorola is processing the audio more before it is being digitized.

Well it could be that mototrbo is more compressed than D-star since twice as much traffic can occur on the same frequency with mototrbo.

Plus on that hamradio-dv site, it sounds like the guy doing the test has a cold or something because his voice sounds very nasal.

The Motorola product is also 2 slot TDMA vs D* that is FDMA. While they are using the same vocoder the demodulation the process is different.

Josh wrote:Well it could be that mototrbo is more compressed than D-star since twice as much traffic can occur on the same frequency with mototrbo.

i think that they should be equal compressed because
d-star has a channel width of 6,25 Khz
and mototrbo has a channel width of 12,5 Khz but is divide in 2 time slots = 12,5 khz /2 = 6,25 khz

KS4VT wrote:The Motorola product is also 2 slot TDMA vs D* that is FDMA. While they are using the same vocoder the demodulation the process is different.

yes but the vocoder is receiving a similar bit pattern in both standards, 2400 bit compressed voice and 1200 bit FEC...

KS4VT do you belive it is because motorola uses an other type of Digital to analog converter that i sounds that different ?
the sample rate maybe is less in mototrbo in the demodulation process ?

W2MB wrote:I'm just guessing that Motorola is processing the audio more before it is being digitized.

yes maybe thats why, but then motorola should bee able to make is sound more natural later if the want to by changing in the voice processing software in the firmware



Maelv...

I have noticed that Mototrbo audio that is using security encoding does sound quite compressed.

However standard audio sounds much better. Certainly not hi-fi but very good.

It's so nice not to have signal fade, white noise and static like heard on analog. Digital audio takes a little getting used to but crystal clear audio right out the the edge of system coverage is pretty neat.

Randy

Necro-threading brings up some of the coolest stuff. I liked those comparison wav files. It would have been interesting to hear the same guy on the two different products because their voices were in different octaves. The first guy was a bit more baritone than the second, and that seems to always affect subjective quality tests even in existing analog systems. It would also be interesting to hear the radios down at the bottom of performance when the ber starts climbing.

Yes it is interesting to watch the received signal level on RDAC of the 8300 repeater involved. It is quite amazing to see very low RSSI levels but have clear audio. Audio doesn't start getting muddy sounding until it is right around -125 or so.

I have done some comparison between a VHF analog and UHF Mototrbo repeater running about the same power output and same antenna height. At fringe locations, the VHF was naturally weak and noisy. The Mototrbo repeater sounded the same as it did while parked under the tower.

While not being a huge supporter of digital, my opinion has certainly been changed.

Randy

I have used trbo systems heavily and our audio does not sound like the trbo audio in the wav file? Our audio sounds pretty good, however every once and a while a particular persons voice does distort a little bit.

maelv wrote:

Josh wrote:Well it could be that mototrbo is more compressed than D-star since twice as much traffic can occur on the same frequency with mototrbo.

i think that they should be equal compressed because
d-star has a channel width of 6,25 Khz
and mototrbo has a channel width of 12,5 Khz but is divide in 2 time slots = 12,5 khz /2 = 6,25 khz

Maelv...

TDMA is a phase shift, not a way to slice bandwidth. Both channels are 12.5 KHz they are 90 degrees apart respectively on the wave form. You could actually put 4 channels or more in a 360 degree time span, it's done all of the time in the satcom world.


Dan
KA8YPY

ka8ypy wrote:TDMA is a phase shift, not a way to slice bandwidth. Both channels are 12.5 KHz they are 90 degrees apart respectively on the wave form. You could actually put 4 channels or more in a 360 degree time span, it's done all of the time in the satcom world.

Uhh, NO.

Time Division Multiple Access involve sending one channel's data for part of the time, then another channel's data for part of the time, etc. Lather rinse repeat for all channels. Specifically, MotoTRBO is a 2 slot TDMA system, so one channel's data is sent for half the time, and the other channel's data is sent for the other half the time.

What you are describing would be a form of quadrature modulation, except:
1) in QPSK, the phase shift is applied to groups of bits from the data stream (well, technically, it is applied to symbols, which may or may not represent an integer number of bits, but given just how much WRONG was in the parent post trying to explain information coding theory is probably a little too much here).
2) You cannot "put 4 channels into a 360 degree time span", at least not by phase shifting the data by 90 degrees each time without some other coding going on, as there would be no way to differentiate the signals at 0 and 180 degrees from each other.
3) In the satcom world, various forms of QPSK and QAM are used, with various numbers of symbols (various numbers of degrees phase shift between symbols), but again, the symbols themselves represent the same data stream. If the data stream is multiplexed, it usually is either a form of Code Domain Multiple Access (spread spectrum), or a form of, WAIT FOR IT, Time Division Multiple Access, where the data stream is sliced up into time slots and the time slots are then assigned to communication channels.

So, do we call trbo 625 equivalent because of the two time slots derived from a 12.5kc chan, or 12.5kc because of the overall envelope no matter how many time slots it carries?

eta: applying some google fu, it seems Mototrbo satisfies FCC requirements for 625 operation despite the envelope. It is considered 6.25kc equivalent. And if father clear channel is happy, we're happy.

Bill_G wrote:So, do we call trbo 625 equivalent because of the two time slots derived from a 12.5kc chan, or 12.5kc because of the overall envelope no matter how many time slots it carries?

eta: applying some google fu, it seems Mototrbo satisfies FCC requirements for 625 operation despite the envelope. It is considered 6.25kc equivalent. And if father clear channel is happy, we're happy.

Yes, most of the 6.25kHz equiv. protocols are really 12.5kHz wide but 2 or more time slots.

The downside to these approaches is that you have to have a master timing reference so everybody knows where their time slots are - which makes simplex difficult to impossible (usually, these protocols fall back to a 12.5kHz non-TDMA format for simple simplex.)

Ah,

If Wowbagger is stating fact, then what is the format the radios are in when in Digital simplex? As pointed out the time base would not be available to TDMA the modulation.

Would this be then only a FDMA AMBE transmission?

Humm, oh, does anyone have a service monitor with a D-star vocoder in it.

Thinking, Thinking,

CoM

chartofmaryland wrote: Humm, oh, does anyone have a service monitor with a D-star vocoder in it.

There is no such thing as a "D-star vocoder" - there is the DVSI AMBE vocoder, which is used in D-Star, MotoTRBO, APCO-25 F2 (Motorola extension), APCO-25 P2 (APCO standards track), and the G4GUO protocol AOR is pushing as a digital HF mode.

The Aeroflex 3900 has the AMBE vocoder in it for MotoTRBO and P25, but does not have the D-Star protocol stack (although it can generate a D-Star waveform using the ARB functions).

Good info here in the thread.

Regarding TRBO on simplex, my understanding is that it drops to a single timeslot but still goes out in digital. Not sure if I understand correctly though, but I recall reading that the repeater was needed to keep the timing correct for 2 timeslot transmissions.

escomm wrote:Good info here in the thread.

Regarding TRBO on simplex, my understanding is that it drops to a single timeslot but still goes out in digital. Not sure if I understand correctly though, but I recall reading that the repeater was needed to keep the timing correct for 2 timeslot transmissions.

When mototrbo is in simplex it is using the entire 12.5 due to no timing from a repeater. You can imagine that typical Talk Around Operations like we are used to on analog, does not play well in the mototrbo world. But to me, The benifits outway the downfall, just go to a simplex channel that is not your repeaters frequencies. And your in business. It has been awile since I have played with the TRBO. Due to my new job, I only get to play with P25 trunking Miss playing with the TRBO stuff. But hey, when my New APX7000 VHF/800 shows up, I'm sure Ill be happy again

MOEtorola wrote:

"When mototrbo is in simplex it is using the entire 12.5 due to no timing from a repeater."

If that's true, why does the CPS ask me for a time slot when I program a simplex channel?

Tom in D.C. wrote:MOEtorola wrote:

"When mototrbo is in simplex it is using the entire 12.5 due to no timing from a repeater."

If that's true, why does the CPS ask me for a time slot when I program a simplex channel?

You're right, it does ask for a time slot. However, it doesn't matter what time slot you choose, it will be a full 12.5 KHz spacing. Try different radios on both, they will talk to one another.
As I've said earlier, lazy code writers.