-
Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Who at Northern.... (Poor quality audio for auto-announcements)
- Thread starter S101
- Start date
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R
RailUK Forums
lokiavivson
Member
The male voice in the first video is Peter Corley, introduced in May 2023.
The female voice in the second video is Trisha Jarman, replaced by Peter Corley from May 2023.
And yes I agree that it didn't need to be changed, Northern just love ruining things for no apparent reason!
318266
Member
ScotRail have what appears like the same TrainFX system on their 156 and HST fleets, and I've always thought that the tinny voice box certainly is quite noticeably cheap, which is a stark contrast to the 1st class seats they have!
Rail2002
Member
I agree. It's terrible and echoey. It sounds as though someone recorded it in their bedroom on a laptop microphone. The TransPennine ones (same owning company) are far more professional.
DownSlow10
Member
That’s it yes! Thank you
much better they were.Rail2002
Member
Not great IMO. I remember the way he'd say "this is the Northern service tuuuhh". Regional accents don't work on trains because people travel from all over the country and it can be difficult to understand. The female announcements regarding delay repay are particularly taxing, especially if other passengers are engaged in chatter. An RP voice works best, such as the lady who does the Chiltern announcements.That’s it yes! Thank you
DownSlow10
Member
I do remember the tuuuh but was changed to standard “to” not long after the removal of his voice, there is a reason as to why his voice was removed from the train announcements onboard, I’ve just done some research and was because he was assaulted onboard a train due to a fare dodger and resulted in him leaving his job.
Class777
Member
It had seemed they were okay until I was on 331105 the other day and the announcements were a mix of tinny and almost like what you get on a cheap neck mic for £10 from Amazon!
Hopefully the processing can be fixed as it doesn’t make the operator look the best, especially given all the changes with the announcement system over the last 18 months or so.
Hopefully the processing can be fixed as it doesn’t make the operator look the best, especially given all the changes with the announcement system over the last 18 months or so.
My main issue isn’t the quality of the announcements, but that all too often the automated announcement system fails (or perhaps is deliberately disabled), and then guards do not make voice announcements to compensate. This really isn’t good enough in 2026!
tfw756rider
On Moderation
Of course, the recording of the announcement files themselves is a factor, but TrainFX has notoriously muffled and unnecessarily bassy audio when played through the trains with the proper 4"/5" speakers (not the tiny little ones on some 150s) - and yes, I know that some of them have unsuitable voice coil inductance for the application and so filter out treble, but still).
It's like they don't know anything about audio equalisation or filtering - they didn't apply a high pass or 'low cut' filter (to filter out the unnecessary, ceiling panel-rattling bass), and a treble boost to increase intelligability, because they didn't know how to (or they didn't even think of it).
So it's up to the operators to apply the high-pass / 'low cut" bass filtering and the treble boost to the files, but of course they can't be expected to do that. I mean, clearly they sometimes can't even get good recordings, so...
It's like they don't know anything about audio equalisation or filtering - they didn't apply a high pass or 'low cut' filter (to filter out the unnecessary, ceiling panel-rattling bass), and a treble boost to increase intelligability, because they didn't know how to (or they didn't even think of it).
So it's up to the operators to apply the high-pass / 'low cut" bass filtering and the treble boost to the files, but of course they can't be expected to do that. I mean, clearly they sometimes can't even get good recordings, so...
Rail2002
Member
Golly, that's awful.
Surely the DfT who are in charge of Northern can arrange a professional recording studio and speaker systems. It's them after all who stipulated similar announcements for buses. For some reason, I always thought that the current voice was a train enthusiast who was offered the opportunity. It's concerning that he's head of communications at Northern and felt such lacklustre and poor quality announcements were fit for purpose. As someone else said, the speakers themselves seem to be fine, it's the quality of recording that's the issue. Unfortunately, Northern SCs have a habit of duplicating announcements but the sound quality is much better, meaning no fault with the train's speakers.
tfw756rider
On Moderation
The trains' public address systems (on pre-1998 stock) are designed for manual announcements only. They either apply the frequency filtering/equalisation to the microphone signal or just rely on the natural frequency range of the microphone used being suitable enough.
Any 'auxiliary' signal (like a future automated announcement system) would need to provide its own frequency filtering/equalisation, but TrainFX fails to do that. That combined with recordings that aren't very good makes for an even worse end result.
I thought the driver/fitters had to test them every day before duty.
Yes, They are knackered!
Tracked
Established Member
Was on a Sheffield-Scarborough train on Sunday, not sure how they managed it when listing the stops, but the voice seemed to be moving further away from the microphone the closer they got to Scarborough.
Incidentally, was on a train yesterday with a couple of Northern staff (guard/driver) sat in the passenger seats, even they were saying how awful the announcements were. Think I remarked on here last Christmas that the announcements on the eastern units in Northern's fleet were uniformly terrible (this is on 150/170/195/331/333) whilst the western ones (195/156) were ok, I'm not giving the location away other than being able to say this is definitely not the case ...
Incidentally, was on a train yesterday with a couple of Northern staff (guard/driver) sat in the passenger seats, even they were saying how awful the announcements were. Think I remarked on here last Christmas that the announcements on the eastern units in Northern's fleet were uniformly terrible (this is on 150/170/195/331/333) whilst the western ones (195/156) were ok, I'm not giving the location away other than being able to say this is definitely not the case ...
Considering how awful the rest of the TrainFX kit appears to be, it wouldn't suprise me if there was only 10MB of storage available and the only audio format it understood forced sub phone call quality audio, meaning even with the best recordings in the world, they'll still sound like garbage when played back (and live announcements would sound fine).
This sort of special codec isn't unusual. I once used to deal with a phone PBX software which mandated mono 10-bit / 4600 sample files, which had to be encoded in their own codec which stripped the audio to about 200Hz - 2kHz. Everything sounded awful - while voice messages were just about usable, the on hold music was simply torture to listen to.
This sort of special codec isn't unusual. I once used to deal with a phone PBX software which mandated mono 10-bit / 4600 sample files, which had to be encoded in their own codec which stripped the audio to about 200Hz - 2kHz. Everything sounded awful - while voice messages were just about usable, the on hold music was simply torture to listen to.
TrainFX on other operators, whilst it has its flaws, is a lot better sounding than Northern’s offering. GWR (Turbos, 150s, 158s); TfW (150s and 158s), ScotRail (153s, 156s) all have relatively clear audio, as did Northern before they replaced Trisha Jarman (or even Peter Corley), subject to the volume being set at the right levels.
But then again, all these voices were either professional voices using professional recording studios, or staff voices with some announcing experience, using professional recording studios.
The current Northern offering has neither, but the awful speakers on units like the 323s don’t help.
Last edited:
Didn't realise GWR Turbos were the same system. They sound miles better. When they were refurbished a few years back I thought it was a new system given how much clearer it was!
Nym
Established Member
TrainFX didn't exist when the turbos were built so it will have been a new system.
Ah which helps explain my experience of manual PAs sounding fine but the automatic announcements being god awful.
Class 323 Fan
Member
Does anyone know why Trisha and Peter were changed?
ic31420
Member
Well.
It hadnt really caused me any annoyance until I saw this thread... Now it really bugs me.
For me the main issue is the big pregnant pauses in the announcements.
Thank for joining this northern service to....... ..... Rochdale. The next stop will be........ Bolton.
Adds an element of apparent suprise to the announcement.
Can the guard turn the system off? On a later service the automatic announcements were replaced by a guard.
It hadnt really caused me any annoyance until I saw this thread... Now it really bugs me.
For me the main issue is the big pregnant pauses in the announcements.
Thank for joining this northern service to....... ..... Rochdale. The next stop will be........ Bolton.
Adds an element of apparent suprise to the announcement.
Can the guard turn the system off? On a later service the automatic announcements were replaced by a guard.
northernbelle
Member
- Joined
- 10 Oct 2018
- Messages
- 779
Agree on both counts and it drives me round the bend. It's like an impression of someone doing an announcement rather than the real deal.
A great deal of care and attention was taken to record and edit the GWR ones to ensure gaps between phrases were natural, volume was normalised and consideration given to how it would sound when all stitched together. It's virtually impossible to get it perfect but I do think the effort has resulted in a reasonably decent (and distinctive to GWR) job.
SuspectUsual
Established Member
- Joined
- 11 Jul 2018
- Messages
- 6,861
I’ve just been on a Northern 170 into Harrogate and every announcement sounds like it was recorded in a blizzard
Class455
Established Member
- Joined
- 19 May 2016
- Messages
- 1,521
I also completely agree that the quality of Northern’s automated announcements are the worst I’ve come across on the entire network. Whenever I catch a 323 in particular, which already have tinny speakers as it is, it’s actually funny how terrible they sound. To me it sounds like they recorded it in a room that isn’t sound proofed where the voiceover artist sounds like he’s just finished a heavy gym session.
This combo with the dreadful TrainFX PIS system surely isn’t DDA compliant and I’m surprised no one hasn’t forced northern to upgrade their PIS. Might be pointless on the 323’s as they’ll be gone within the next year but it’s not much better on their other fleets. I will never understand why they got rid of the Bombardier PIS system on the 170’s for the terrible TrainFX system. They don’t even work in multiple with other fleets that often.
This combo with the dreadful TrainFX PIS system surely isn’t DDA compliant and I’m surprised no one hasn’t forced northern to upgrade their PIS. Might be pointless on the 323’s as they’ll be gone within the next year but it’s not much better on their other fleets. I will never understand why they got rid of the Bombardier PIS system on the 170’s for the terrible TrainFX system. They don’t even work in multiple with other fleets that often.
Purple Train
Established Member
I'm currently on the Heysham Port train, and the announcement is for "Heysham Harbour" (which is fine, as the station has had a long-running identity crisis and both names are used). However, from what I can gather over the horrendous amount of static, the announcer pronounces it HAY-sham instead of the correct HEE-sham, which even the text-to-speech announcements on local buses get right (although the bus announcements mispronounce Lancaster, so I guess it's swings and roundabouts). It's a shame that they appear to be done so cheaply as the announcer has quite a good voice from what I can hear of it.
Transilien
Member
I think it's the same case at Scotrail
318266
Member
ScotRail's new voice is an AI model based upon lines from Vanessa Sloan, who I believe works frontline (ticket office?) in the Paisley area.I think it's the same case at Scotrail