Rape complainants are still being asked "humiliating and irrelevant" questions about flirting, dancing, sex toys, clothing and contraception, with judges failing to step in, a new study shows.

Researchers who listened to audio from 40 New Zealand rape trials found women were largely left to fend for themselves during "brutal" cross-examination, despite laws designed to protect them.

Lawyers also used complicated language which many of the complainants - who were often and young and vulnerable - didn't understand; and patronised them for being unable to properly describe their anatomy.

The study's unprecedented access to audio provided both a window into the closed-door rape trial process, and evidence for change, the study's lead author Professor Elisabeth McDonald said.

"These are mostly quite young women, and they're often struggling. They are vulnerable. And then we put them in an unfamiliar environment and expect them to talk about rape and they're told to speak up or talked down to," McDonald said.

"And most other research asks them about the trial experience and they say it was awful, but the strength of this work will be we now have a piece of information that shows what is actually happening in the courtroom."

• Courtroom shake up: New rules in rape cases hailed as decades-overdue
• How thousands of rape cases vanished from police statistics over 20 years
• CCTV footage captures teen's last moments before horrific rape and murder
• Rose McGowan claims Harvey Weinstein ran 'rape factory' for years

The four-year study's initial focus considered how often rape myths - false beliefs such as "if you wore a short skirt you were asking for it" - were relied up in cross examination, and how that added to the trauma of giving evidence.

It was an issue highlighted most recently during the murder trial of British backpacker Grace Millane - where women also assaulted by her killer were told they were liars or attention-seekers, with some breaking down on the stand.

However, the study also discovered other aspects of questioning which the complainants found distressing.

This included small things, like being repeatedly told to speak louder because the microphone wasn't working, to big things, like being unable to understand what the lawyers were saying.

Examples included counsel using terms like "courting" instead of "dating", or "brassiere" and not "bra" with women who were young and uneducated.

One complainant was asked to read her own statement, and became upset when she couldn't. The Crown lawyer hadn't realised she was illiterate before the hearing began.

Other women struggled to describe their rape, or find the correct terms for their genitalia. In one case, a prosecutor appeared to mock his own witness for using the slang "rooting" for "sex".

"It's awful. And it's quite de-personalising. That's the feeling that you get," McDonald said. "Part of that is that the people who work every day in courtrooms forget how unfamiliar and overwhelming it is."

The study found that although the issue in all the trials was consent, the focus of cross-examination was usually the complainant's credibility - and undermining that was where a reliance on rape myths came in.

For example, if a woman had a shower after being assaulted, defence lawyers argued she couldn't have been raped. In another case, when a woman didn't shower, they therefore said she couldn't have been raped.

Similarly, when women didn't call the police straight away or tell anyone about the rape, lawyers said they were lying. In one hearing, a victim was asked six times about why she didn't report her assault immediately. The judge never intervened.

In other cases women were asked about "flirting" with others earlier in the night; their dancing; their clothes; their children and pregnancies, and how many sex toys they owned.

"The inference is, you were not raped, it was consensual sex, later regretted," the study said.

"If you were really raped, you would remember every detail, you would have fought, you would have injuries, you would have gone straight to the police, you would have eschewed any contact with the defendant."

Complainants were also regularly questioned about if they had been in a relationship with someone else during the rape - with lawyers suggesting they'd made up the assault to hide an affair.

Although judges are allowed to step in and rule out irrelevant evidence, or intervene in repetitive or improper questioning, the study found they rarely did so, or were inconsistent when they did.

To address this, the study recommended the judiciary undergo further training to ensure they were following the laws - including any new legislation brought under the sexual violence bill currently working its way through Parliament - properly.

It also suggested the specialist sexual violence pilot court, which formed part of the research, be expanded to the whole country.

In total, it made 55 recommendations, including urging the Government to update the definition of consent in New Zealand law; and changes to the Evidence Act about what is admissible.

There are currently no plans to review the substantive law relating to rape, although it is on Justice Under-Secretary Jan Logie's "longer-term" work programme.

The New Zealand Law Society has said previously that while it supports efforts to reduce retraumatisation, any proposed law changes needed to strike the right balance between complainant welfare and upholding the fundamental right to a fair trial.

Access a copy of the study here.