Upsampling MQA files to original resolution with sox

[granosalis]

(13-05-2020 21:32)Alandbush Wrote:  Given that you have only a few MQA cds, it may be easier/quicker to generate an upsampled version of them yourself, directly using Sox.
You then have the ability to compare the before and after to determine if you hear a difference.

It is not a big difference, but soundstage is more open and the sound is more airy.
But I'm limited at 16/192, because my DAC do not support 384 Khz

[simoncn]

(13-05-2020 23:11)granosalis Wrote:  I tried, but does not work. Screen in attachment.

It looks like /usr/share/ffmpeg is not an ffmpeg executable binary file. The stream.converter property must be set to an executable binary filename, not a folder name.

[Alandbush]

(14-05-2020 05:40)granosalis Wrote:  

(14-05-2020 00:52)Alandbush Wrote:  The earlier stated sox command to upsample MQA (see below) is no longer supported, assuming it was correct initially.

sox -S input.mqa.flac output.flac -vsM 352800

For sox 14.4.2 the command is as follows:-

sox -S input.mqa.flac output.flac rate -v -M 352800

where 352800 is the original sampling frequency (tag ORFS) of the MQA file.

Thanks Alan,
I did it, but the output files is 16bit. Is there any way to produce a 24bit file?
if I put -b 24 the file becomes unlistenable.
Giuseppe

I've tried the following command successfully on a 16bit MQA/CD track downloaded from 2L. It appears 'listenable'.

sox -S input.mqa.flac -b 24 output.flac rate -v -s -M 352800

nb - I have also introduced an extra override (-s 'steep filter'). If only because the command now looks very similar to the originally posted command.

[Cebolla]

Isn't this upsampling with minimum phase steep filter just a substitute for the final MQA rendering step, which produces the 'further unfolds'?

So, there appears to be nothing here that takes care of the initial step, ie, the actual MQA decoding of the MQA source file track to produce the MQA Core signal (aka 'first unfold').


If that has already been taken into account, how do you get the MQA Core signal/first unfold into MinimServer in the first place?

[granosalis]

(14-05-2020 11:39)Alandbush Wrote:  

(14-05-2020 05:40)granosalis Wrote:  

(14-05-2020 00:52)Alandbush Wrote:  The earlier stated sox command to upsample MQA (see below) is no longer supported, assuming it was correct initially.

sox -S input.mqa.flac output.flac -vsM 352800

For sox 14.4.2 the command is as follows:-

sox -S input.mqa.flac output.flac rate -v -M 352800

where 352800 is the original sampling frequency (tag ORFS) of the MQA file.

Thanks Alan,
I did it, but the output files is 16bit. Is there any way to produce a 24bit file?
if I put -b 24 the file becomes unlistenable.
Giuseppe

I've tried the following command successfully on a 16bit MQA/CD track downloaded from 2L. It appears 'listenable'.

sox -S input.mqa.flac -b 24 output.flac rate -v -s -M 352800

nb - I have also introduced an extra override (-s 'steep filter'). If only because the command now looks very similar to the originally posted command.

I have a clip warning and the result of the trasnformation is just noise.

C:\Program Files (x86)\sox-14-4-2>sox -S -b 24 "C:\Users\giuse\home\02 traccia02.wav" C:\Users\giuse\home\Traccia02.flac rate -v -s -M 352.8k
sox WARN wav: User options overriding size read in .wav header

Input File : 'C:\Users\giuse\home\02 Traccia02.wav'
Channels : 2
Sample Rate : 44100
Precision : 24-bit
Duration : 00:02:53.34 = 7644392 samples = 13000.7 CDDA sectors
File Size : 45.9M
Bit Rate : 2.12M
Sample Encoding: 24-bit Signed Integer PCM

In:100% 00:02:53.34 [00:00:00.00] Out:61.2M [ -==| ] Hd:0.0 Clip:7.31M
sox WARN rate: rate clipped 4871455 samples; decrease volume?
sox WARN sox: `C:\Users\giuse\home\Traccia02.flac' output clipped 2437165 samples; decrease volume?
Done.

[granosalis]

(14-05-2020 10:12)simoncn Wrote:  

(13-05-2020 23:11)granosalis Wrote:  I tried, but does not work. Screen in attachment.

It looks like /usr/share/ffmpeg is not an ffmpeg executable binary file. The stream.converter property must be set to an executable binary filename, not a folder name.

It is pointing to the script as suggested in the post. It means the proposed solution by the user dos't work?

[simoncn]

If /usr/share/ffmpeg is a script and not a folder, check that the script has executable permission set.

[granosalis]

(14-05-2020 13:04)simoncn Wrote:  If /usr/share/ffmpeg is a script and not a folder, check that the script has executable permission set.

Yes, I confirm. It is also visible on the screenshoot I have attached in some post before. I'll check it again, by the way. Thanks.

[Alandbush]

(14-05-2020 12:57)granosalis Wrote:  

(14-05-2020 11:39)Alandbush Wrote:  I've tried the following command successfully on a 16bit MQA/CD track downloaded from 2L. It appears 'listenable'.

sox -S input.mqa.flac -b 24 output.flac rate -v -s -M 352800

nb - I have also introduced an extra override (-s 'steep filter'). If only because the command now looks very similar to the originally posted command.

I have a clip warning and the result of the trasnformation is just noise.

C:\Program Files (x86)\sox-14-4-2>sox -S -b 24 "C:\Users\giuse\home\02 traccia02.wav" C:\Users\giuse\home\Traccia02.flac rate -v -s -M 352.8k
sox WARN wav: User options overriding size read in .wav header

Input File : 'C:\Users\giuse\home\02 Traccia02.wav'
Channels : 2
Sample Rate : 44100
Precision : 24-bit
Duration : 00:02:53.34 = 7644392 samples = 13000.7 CDDA sectors
File Size : 45.9M
Bit Rate : 2.12M
Sample Encoding: 24-bit Signed Integer PCM

In:100% 00:02:53.34 [00:00:00.00] Out:61.2M [ -==| ] Hd:0.0 Clip:7.31M
sox WARN rate: rate clipped 4871455 samples; decrease volume?
sox WARN sox: `C:\Users\giuse\home\Traccia02.flac' output clipped 2437165 samples; decrease volume?
Done.

The command I used changed the bit rate of the output file.
Your command changes the bit rate of the input file and presumably results in your reported issue.

[granosalis]

(14-05-2020 14:55)Alandbush Wrote:  The command I used changed the bit rate of the output file.
Your command changes the bit rate of the input file and presumably results in your reported issue.

Got the mistake, thanks a lot!