Skip to main content

Topic: how to compress 93gb speech mp3 files to opus files (Read 3382 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.
Re: how to compress 93gb speech mp3 files to opus files
Reply #25
>dbpowerAMP to convert mp3 to opus and they produced good quality files but they are unable to produce directory structure same as source.

For your naming string set:

[origpath]\[origfilename]

I am sorry to say that it is not possible to type [origpath]\[origfilename] and the only option is to specify output directory and select options already available options like sourcepath/filename.

Re: how to compress 93gb speech mp3 files to opus files
Reply #26
rupeshforu3 you can add "--downmix-mono" to the opus encoder in lamexp, it appears that lameXP don't downmix to mono and it can be a bug. Another option is TAudioConverter it also maintains the folder structure and transfer the tags, is less updated but you can update the encoders and decoders manually.

Is there any other tool or software just to downmix to mono for already transcoded files.

Re: how to compress 93gb speech mp3 files to opus files
Reply #27
rupeshforu3 you can add "--downmix-mono" to the opus encoder in lamexp, it appears that lameXP don't downmix to mono and it can be a bug. Another option is TAudioConverter it also maintains the folder structure and transfer the tags, is less updated but you can update the encoders and decoders manually.

Does TAudioConverter uses ffmpeg because when I open task manager I am able to see processes named ffmpeg. I heard that opus encoder which comes with ffmpeg is of poor quality and I want to use opusenc original which comes with opustools.

TAudioConverter is displaying lot of error messages and it is displaying message as 92 steps of 5000 total steps done so far and so I think it is taking unnecessary time.

It is not writing immediately the transcoded files to destination folder. I think that if lamexp take 4 days to transcode the software you mentioned may take 12 days.
  • Last Edit: 19 June, 2017, 11:03:10 AM by rupeshforu3

  • spoon
  • [*][*][*][*][*]
  • Administrator
Re: how to compress 93gb speech mp3 files to opus files
Reply #28
>dbpowerAMP to convert mp3 to opus and they produced good quality files but they are unable to produce directory structure same as source.

For your naming string set:

[origpath]\[origfilename]

I am sorry to say that it is not possible to type [origpath]\[origfilename] and the only option is to specify output directory and select options already available options like sourcepath/filename.

The registered version 'dBpoweramp Reference' is required to enable dynamic naming, of which [origpath]\[origfilename] is a part of.

Re: how to compress 93gb speech mp3 files to opus files
Reply #29
In the output directory of dbpoweramp I am able to see directory named [origpath] but it it is not maintaining directory structure.

  • spoon
  • [*][*][*][*][*]
  • Administrator
Re: how to compress 93gb speech mp3 files to opus files
Reply #30
Make sure you have dBpoweramp R16.2 installed, open Batch Converter, select the root folder to convert, click convert. On the next page where you select the encoder, where it says 'output to' click the drop box and choose 'preserve source path', choose the location  where you want it to be based (the source path is appended to the path you choose).

Click the 'list / rename' button to verify that the 'Destination Files' are indeed saving to the correct location.

Re: how to compress 93gb speech mp3 files to opus files
Reply #31
Does TAudioConverter uses ffmpeg because when I open task manager I am able to see processes named ffmpeg. I heard that opus encoder which comes with ffmpeg is of poor quality and I want to use opusenc original which comes with opustools.

No it uses opusnenc, ffmpeg is used to decode the original file. Also you can go to the installation folder locate opusenc and overwrite it with the last version to update it.


TAudioConverter is displaying lot of error messages and it is displaying message as 92 steps of 5000 total steps done so far and so I think it is taking unnecessary time.

It is not writing immediately the transcoded files to destination folder. I think that if lamexp take 4 days to transcode the software you mentioned may take 12 days.
Are those really errors or are processing mesages, also per file it does 3 (decode,encode,rename) or 4 (decode,effects,encode,rename) steps, the steps count is for all steps of all files and don't represent the number of files to be processed, and is better to disable all effects filters for fast encoding. LameXP also does the same steps but they are show in a different way. There are not going to be a significant speed difference betwen both unless differences in configuration.
  • Last Edit: 19 June, 2017, 07:21:16 PM by Phanton_13

Re: how to compress 93gb speech mp3 files to opus files
Reply #32
Yes those are processing messages but why it displays those.

Another issue is suppose I add 1000 files to TAudioConverter and when I press start it has divided the work into 4000 sub tasks and it is not performing those steps sequentially. During the process if there is any power failure and after restarting the above process repeat again.

I think that foobar is best suitable for my task and even I have transcoded some files by downgrading to mono and it was done perfectly and even listening quality is also good. The only problem with foobar is maintaining directory structure.

Re: how to compress 93gb speech mp3 files to opus files
Reply #33
Upto yesterday I have converted using lamexp and the resulted files was stereo but phanton_13 suggested to transcode to mono and so I have tried TAudioConverter but it is very slow and so I am trying DbpowerAmp. I have not deleted stereo opus files but made backup.

Is there any real quality gain in mono files than stereo files. Is there any quality loss if we transcode stereo mp3 files to mono opus files.

At present I am able to transcode mp3 files to opus files using DbpowerAmp. In dynamic naming I have selected source path/source file name.

I have selected vbr quality 16 kbps and there is no option to select mono and so I have added dsp options  channel mapping to mono and channel count to 1 ie., mono. I have added another two dsp options resampling and selected 48000 hz after that I have added dsp option bitdepth and selected minimum depth to 24 and maximum to 32.

The convertion process is being done at lightning speeds and so the resulting listening quality is not good.

I think that dbpoweramp is a product of foobar but foobar is producing high quality files at same configuration. Can you suggest how to preserve directory structure in foobar.
  • Last Edit: 21 June, 2017, 12:46:54 AM by rupeshforu3

Re: how to compress 93gb speech mp3 files to opus files
Reply #34
As dbpoweramp is converting at lightning speed can I think that the produced opus files are worse. I listened some of files and they are good to some extent.

Re: how to compress 93gb speech mp3 files to opus files
Reply #35
What does framesize, bit depth mean.

  • spoon
  • [*][*][*][*][*]
  • Administrator
Re: how to compress 93gb speech mp3 files to opus files
Reply #36
Fast speed does not always equal quality loss, a program can be optimised to run 10x faster through CPU intrinsics.

The frequency conversion in opus might not be as fast as dBpoweramps, if opus is doing the conversion, the results of dBpoweramps frequency conversion can be seen here:

https://forum.dbpoweramp.com/showthread.php?36397-Resampler-in-dBpoweramp-is-a-high-quality-resampler-employed


Re: how to compress 93gb speech mp3 files to opus files
Reply #37
Thanks to phanton_13 for giving valuable suggestions and I have followed his suggestions especially sample rate ie., 48 kHz and number of channels ie to encode to mono.

I have used a number of tools and software to convert and compress 93 gb speech files and even I have tried a number of options.

The most efficient tool or software is dbpoweramp which is very fast at converting but it doesn't have options to select some like sample rate and number of channels etc at main window of encoder instead it allows adding those using dsps. It doesn't have option framesize and algorithm complexity --comp neither in main window of encoder or in dsps.

For the most of transcoding I have used vbr quality 16 kbps and 48 kHz sample rate, number of channels to 1 ie., mono, bit depth minimum 24 bit to maximum 32 bit float.

At present I have examined a number of files converted and 80% of the files are at acceptable quality and for remaining 20% I am planning to convert at 48 kbps bitrate.