ITU‑T (International Telecommunication Union – Telecommunication Standardization Sector) issues technical recommendations – it does not issue certificates. What the industry calls “ITU‑T testing” means conformance testing against ITU‑T standards for protocol and voice quality – certificates are issued by labs. Do not expect an official ITU‑T certificate.
In the automotive field, ITU‑T standards fall into three major groups: H‑series (multimedia communication protocols), P‑series (voice transmission quality), and G‑series (codecs). The test equipment, methods, and judgement criteria are completely different.
1.1 H‑series – multimedia communication
H.324 is the framework for low‑bit‑rate multimedia terminals – covering video‑phone and video‑conferencing codecs, multiplexing, and control protocols.
The base H.324 family includes H.263 video, G.723.1 audio, H.223 multiplexing, and H.245 control. AMR audio and H.264 video are optional extensions – not part of the native mandatory set. Testing is done layer by layer – not a black‑box whole.
In vehicle applications, the CarPlay video‑call architecture early referenced H.324 but is now Apple’s private implementation – not compatible with standard H.324 and does not use the full H.245/H.223 protocol family. Generic H.324 conformance testing cannot substitute for CarPlay‑specific interoperability tests – many mistakenly believe that passing standard H.324 satisfies CarPlay video‑call admission – only to find it was wasted.
1.2 P‑series – voice transmission quality
·P.1100 – narrowband (200–3400 Hz)
·P.1110 – wideband (50–7000 Hz)
·P.1120 – super‑wideband (20–20000 Hz)
The test logic is completely different from H.324 – H.324 cares about correct codec interoperability; P‑series cares about audio fidelity – they are not interchangeable.
1.3 G‑series – voice codecs
AMR, AMR‑WB, EVS – these codec standards belong to G‑series. In vehicle testing, G‑series is usually integrated into H.324 video‑call tests or P‑series voice‑quality tests – not a standalone assessment.
2. H.324 Video‑Phone Test Framework
2.1 Terminal capability and protocol conformance
First, test capability exchange – after the DUT establishes an H.245 control channel with a reference terminal, they exchange codec capabilities – which video formats, which audio codecs, which H.223 annexes for multiplexing – must match the declared spec sheet.
If actual capabilities do not match the declaration – even just one optional codec missing – the lab will mark protocol conformance as failed.
2.2 H.223 multiplexing tests
H.223 multiplexes video, audio, and control data into one serial bitstream – test points: logical‑channel allocation, adaptation‑layer packing/unpacking, error protection, and resynchronisation.
H.223 has annexes – Annex A is basic; Annex B adds resync and error protection – if your product claims Annex B, you must run the full Annex B test cases – including channel‑error injection.
2.3 Video codec interoperability
Run a bidirectional video call between the DUT and a reference terminal – video encoding using H.263 Baseline or H.264 Baseline – compare resolution, frame rate, and bit‑rate control behaviour frame by frame.
2.4 Audio‑video sync
The delay between audio and video in a call – ITU‑T G.114/G.135 give acceptable lip‑sync tolerance of roughly ±80 ms – some OEMs set internal limits as tight as ±40 ms – not an ITU‑T universal limit – follow the admission authority’s specification.
Use standard test sequences – record speaker output and screen display for frame‑by‑frame A/V difference analysis.
3. P‑Series Voice Quality Test Items
3.1 Send and receive frequency response
P.1100, P.1110, and P.1120 each have different frequency‑response templates. Send path – from microphone to network interface – full‑link frequency response. Receive path – from network input to speaker output – full‑link response.
The test source is an artificial mouth – each bandwidth standard has different calibration parameters – do not use one calibration for all three.
3.2 SLR and RLR (Sending/Receiving Loudness Ratings)
SLR and RLR are core subjective evaluation metrics. ITU‑T P.1100/P.1110 reference points: SLR = 8 dB ± 3; RLR = 2 dB ± 3.
These are design targets – not absolute pass/fail thresholds. Exceeding the range does not automatically mean failure – overall MOS and loudness/frequency response are considered together. Some labs mechanically treat out‑of‑range as failure – confirm the judgement logic before starting.
Brands like Apple (CarPlay) have tighter windows – CarPlay requires SLR within 7 ± 2 dB and RLR within 3 ± 2 dB – a report that meets ITU‑T reference but not Apple’s will fail at CarPlay admission.
3.3 Echo cancellation and double‑talk
Echo cancellation is the core of full‑duplex hands‑free – Terminal Echo Loss (TEL) testing simulates echo suppression under different acoustic paths. P.1110 wideband adds more complex multi‑path echo scenarios – acoustic echo plus residual electrical echo superposition.
Double‑talk tests verify that both directions can transmit and receive simultaneously – no clipping or significant attenuation when one party speaks. Test cases include alternate speech, overlapping dialogue, and far‑end voice.
3.4 Noise suppression
Speech intelligibility in background noise is a high‑frequency automotive test. P.1110 noise‑suppression cases are more numerous than P.1100 – background noise types include non‑stationary and impulsive noise – wind noise, engine noise, rain – real driving conditions.
3.5 Delay and packet‑loss compensation
End‑to‑end one‑way voice delay – the 150 ms figure comes from ITU‑T G.114 (general voice transmission) – not from P.1100/P.1110 hands‑free test specifications.
Packet‑loss compensation tests simulate loss rates from 0% to 5% – plot voice quality degradation. Good compensation algorithms show no significant degradation at 3% loss.
4. Test Environment and Equipment
4.1 HATS (Head and Torso Simulator)
Formal P‑series testing typically requires a HATS complying with ITU‑T P.57 – including an artificial mouth meeting P.58 – P.57 covers the full head‑and‑torso model, P.58 specifies acoustic parameters – do not confuse the two.
Customer‑owned phones can only be used for compatibility pre‑testing – not formal testing. HATS dimensions, ear‑canal models, and mouth characteristics are standardised.
4.2 Lab environment control
Background noise – NR15 is common for automotive hands‑free; some brand specs use NR20 – check the customer’s SOW. Ambient temperature: 23 ± 3°C – temperature affects speaker and microphone performance. Typical office noise >45 dB(A) – not suitable for P‑series testing.
4.3 Firmware version lock
Submitted firmware must be final mass‑production stable. P‑series audio scores are directly tied to the firmware version – if you change the version, the score changes. If you modify the audio DSP algorithm after testing, the old report no longer represents the current product – affected test cases must be retested.
5. Tracking ITU‑T Standard Versions
5.1 Current effective versions of P.1100 and P.1110
P.1100 and P.1110 are continuously updated via amendments – there is no unified “edition number”. Recent amendments have updated echo cancellation and full‑duplex double‑talk test models.
5.2 Lab accreditation synchronisation
After an amendment is published, labs must first obtain ISO 17025 scope extension before their new reports are accepted by brands. A single P‑series extension takes ~1 month – full suite (P.1100, P.1110, P.1120 plus HATS calibration) takes 2–3 months.
5.3 When P.1120 super‑wideband is needed
P.1120 (20–20000 Hz) is not mandatory for all automotive projects – CarPlay does not require it – only devices supporting full‑band 48 kHz sampling need it. However, some high‑end overseas OEMs may mandate P.1120 base‑link tests – do not decide solely by hardware bandwidth.
For ITU‑T conformance testing, contact BlueAsia at 13534225140 (King) or king.guo@cblueasia.com.
相关新闻
