Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: How to transcode high pitch in voice mp3 speech files to opus (Read 4910 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

How to transcode high pitch in voice mp3 speech files to opus

Hi I am Rupesh from India. I have some mp3 files which are speeches recorded by someone at 32 kbps to 64 kbps and sample rate varying from 28 kHz to 40 khz. I think that they have transcoded to mp3 using some batch encoder from high quality original files and they are distributing freely through net.

Some of the files are good at quality I mean I am able to hear what the person is speaking in original files and the transcoded files. When listening to some original and transcoded files I am not able to hear what the person is speakingI mean the pitch is very low. For some source  files itself the pitch is very high and I am not able to hear the person's speech.

At present I am using dbpoweramp 16.2 and in options I have selected opus with 24 kbps vbr and after that I have added dsps channel count set to 1, channel mapping set to 1 ie., mono, resample set to 48000. In channel mapping there is an option to select loudness and I have selected 10.4 decibels.

Upon converting files using above options when I listen to transcoded files I am not able to hear the persons voice properly I mean there is sudden raise in pitch, frequencies.

I have copied all the mp3 files which are having the above properties to a separate directory.Another interesting fact is that these mp3 files are pure speeches without any instruments like guitar, violin, piano etc.,.

I am new to concepts like equalizer, replaygain etc,Can you suggest how to obtain pure voice from the source files and maintaining same volume to the total file and if possible to whole volume.

Are there any options to obtain pure voice without high pitch, noise etc.,.Is aac encoding is better than opus for such issues.


Regards,
Rupesh.

Regards,
Rupesh.

Re: How to transcode high pitch in voice mp3 speech files to opus

Reply #1
What is the exact sampling rate and format of the files that have altered pitch when transcoding?

Re: How to transcode high pitch in voice mp3 speech files to opus

Reply #2
Just to get a better fundamental understanding of the problem: Why do you need or want to transcode the files from mp3 to opus?

Re: How to transcode high pitch in voice mp3 speech files to opus

Reply #3
What is the exact sampling rate and format of the files that have altered pitch when transcoding?


I am providing the audio properties of the source file causing disturbances and also the audio properties of the transcoded files below. I have obtained this information from properties window in Windows explorer.

source file:

Artist A.C. Bhaktivedanta Swami Prabhupada
Title 04 - Kamsa Begins His Persecutions
Album New Krsna Book Dictation Tapes
Track 04
Disc
Genre Meditative
Year 1970
Rating
Composer
Size 11.68 MB  (91% Compressed)
Original Size 128.78 MB
Length 25 minutes 31 seconds
Channels 1  (mono)
Sample Rate 44.1 KHz;
Bit Rate 64 kbps
Encoder FHG (Guess)
Encoder Settings Constant Bit Rate 64 kbps
Audio Quality Low  (Lossy)
Contains Album Art, ID Tag [ID3v2.3 & ID3v1.1]
Channel Mapping Center
File KRSNA_book_dictation-04-Kamsa_Begins_His_Persecutions
Type Winamp media file    [.mp3]


transcoded file with 12 kbps 48 kHz opus:

Artist A.C. Bhaktivedanta Swami Prabhupada
Title 04 - Kamsa Begins His Persecutions
Album New Krsna Book Dictation Tapes
Track 04
Disc
Genre Meditative
Year 1970
Rating
Composer
Size 2.3 MB  (99% Compressed)
Original Size 140.83 MB
Length 25 minutes 38 seconds
Channels 1  (mono)
Sample Rate 48 KHz;
Bit Rate 12 kbps
Encoder libopus 1.1.3
Encoder Settings
Audio Quality
Contains
Channel Mapping
File KRSNA_book_dictation-04-Kamsa_Begins_His_Persecutions
Type VLC media file (.opus)    [.opus]


transcoded file with 24 kbps 48 kHz opus:

Artist A.C. Bhaktivedanta Swami Prabhupada
Title 04 - Kamsa Begins His Persecutions
Album New Krsna Book Dictation Tapes
Track 04
Disc
Genre Meditative
Year 1970
Rating
Composer
Size 4.54 MB  (97% Compressed)
Original Size 140.83 MB
Length 25 minutes 38 seconds
Channels 1  (mono)
Sample Rate 48 KHz;
Bit Rate 24 kbps
Encoder libopus 1.1.3
Encoder Settings
Audio Quality
Contains
Channel Mapping
File KRSNA_book_dictation-04-Kamsa_Begins_His_Persecutions
Type VLC media file (.opus)    [.opus]
Regards,
Rupesh.

Re: How to transcode high pitch in voice mp3 speech files to opus

Reply #4
Just to get a better fundamental understanding of the problem: Why do you need or want to transcode the files from mp3 to opus?

I have a pc running Windows 8 and an android phone with memory card and so I am trying to compress and convert all mp3 files and then copy these converted files to memory card.The speech mp3 files I have downloaded are of size 23.5 gb and upon experimenting and converting to opus format the size is reduced from 23.5 gb to 7.2 gb. Most of the files transcoded or converted are acceptable at quality but for some files I am unable to hear I mean there are noises and also suddenly raise in pitch and frequencies. I want pure voice extract from mp3 files eliminating noises and store in corresponding opus files.
Regards,
Rupesh.

Re: How to transcode high pitch in voice mp3 speech files to opus

Reply #5
I doubt that the files really have "altered pitch" in the sense of frequencies in the opus being distorted from the frequencies in the mp3.

One of your two Opus files had a bitrate of only 12kbps. At such low bitrates Opus can't keep the full bandwidth of the original, so the high frequencies are missing. You should give Opus at least 16kbps so it can reliably maintain quality wideband. If you told dbpoweramp you wanted 24kbps and it gave you 12kbps, dbpoweramp has a problem.

If you're finding that some of the files are playing back with high pitches, it's possible the mp3s you're using have a problem. You say these files are being distributed freely on the 'net - if that isn't a copyright violation you could post a link and someone could look for themselves at what's up with the file. (If the files are copyrighted and aren't supposed to be distributed, don't do that, as it's against this site's terms of service.)

Re: How to transcode high pitch in voice mp3 speech files to opus

Reply #6
I doubt that the files really have "altered pitch" in the sense of frequencies in the opus being distorted from the frequencies in the mp3.

One of your two Opus files had a bitrate of only 12kbps. At such low bitrates Opus can't keep the full bandwidth of the original, so the high frequencies are missing. You should give Opus at least 16kbps so it can reliably maintain quality wideband. If you told dbpoweramp you wanted 24kbps and it gave you 12kbps, dbpoweramp has a problem.

If you're finding that some of the files are playing back with high pitches, it's possible the mp3s you're using have a problem. You say these files are being distributed freely on the 'net - if that isn't a copyright violation you could post a link and someone could look for themselves at what's up with the file. (If the files are copyrighted and aren't supposed to be distributed, don't do that, as it's against this site's terms of service.)
First I have transcoded these files using 12 kbps and the quality was not good and so I have again transcoded using 24 kbps even the quality was not good I mean I am unable to hear the speech.

The source files I have transcoded are contained in the zip files and these files are not copyrighted and so we can download freely. I am providing links below

1. http://www.krishnapath.org/SP-Audio/zip-files/SP-KRSNA_Book-Part1.zip

2. http://www.krishnapath.org/SP-Audio/zip-files/SP-KRSNA_Book-Part2.zip

3. http://www.krishnapath.org/SP-Audio/zip-files/SP-KRSNA_Book-Part3.zip

Can you suggest any dsps available in dbpoweramp batch converter to obtain pure voice and then store in opus files. Can you also suggest minimum bitrate and other settings to obtain quality output opus files.
Regards,
Rupesh.