On 09/07/2015 11:03, Dave Plowman (News) wrote:
In article ,
Graeme Wall wrote:
On 08/07/2015 13:35, Jim Lesurf wrote:
In article , Graeme Wall
wrote:
Probably a silly question to bring up on this group but does anyone have
any opinions on the USB turntables John Lewis are selling?

The reason I ask is I've discovered a box full of 1960s singles that I'd
quite like to stick on my iThings and can't be fagged trying to marry
the hi-fi and computer together to do the job properly as it would
involve major rewiring sessions to get them into the same room.

If you have a laptop available it might be cheaper and give better results
to buy a machine-powered USB ADC like the ones sold by Behringer or
Focusrite. See

http://www.audiomisc.co.uk/HFN/ADC/USBrecording.html

for some more info.


Looks interesting but I don't have a laptop unfortunately!

But if you do have an existing record deck, that could surely be moved to
the computer easily?

It;s the bit between the deck and the computer which is the problem, ie
the proper pre-amp. I've a video A-D converter I can utilise the sound
side of once I've got it to line level. Means I'd have to move the main
amplifier as well which means disconnecting all the wiring for speakers,
tuner etc. Fine if I was going to do the job in one hit but means
derigging/rerigging every time I want to do a batch.

Or you could use a wireless link between the existing
Hi-Fi and computer. That could also be useful for other things.



--
Graeme Wall
This account not read, substitute trains for rail.

On 09/07/2015 12:33, Jim Lesurf wrote:
In article , Dave Plowman (News)
wrote:
In article , Graeme Wall
wrote:
On 08/07/2015 13:35, Jim Lesurf wrote:
In article , Graeme Wall
wrote:
Probably a silly question to bring up on this group but does anyone
have any opinions on the USB turntables John Lewis are selling?

The reason I ask is I've discovered a box full of 1960s singles
that I'd quite like to stick on my iThings and can't be fagged
trying to marry the hi-fi and computer together to do the job
properly as it would involve major rewiring sessions to get them
into the same room.

If you have a laptop available it might be cheaper and give better
results to buy a machine-powered USB ADC like the ones sold by
Behringer or Focusrite. See

http://www.audiomisc.co.uk/HFN/ADC/USBrecording.html

for some more info.


Looks interesting but I don't have a laptop unfortunately!

But if you do have an existing record deck, that could surely be moved
to the computer easily? Or you could use a wireless link between the
existing Hi-Fi and computer. That could also be useful for other things.

Note also that some of the cheaper Behringer ADCs have an inbuilt RIAA
preamp for MM cartridges. I think they're about 25 quid. How good they are
I can't say, but used with care they may be better than a complete cheap
'USB turntable'.


That does sound interesting, thanks.

--
Graeme Wall
This account not read, substitute trains for rail.

On 2015-07-13 10:14:54 +0100, Jim Lesurf said:

In article , Don Pearce
wrote:
Downsampling from 88kHz is no easier than from 96kHz. You have to go
through the same process of interpolation, upsampling, lowpass
filtering, and finally decimation. Never try to do it by skipping
alternate samples - that way lies alias distortion from quantization.

I'm a bit pun surprised that you wrote that. Makes you look like a
mathematician/theoretician rather than an injuneer. 8-]

The actual real computation of a good 96k - 44.1k conversion does tend to
be far more demanding in terms of number of operations / coefficients / etc
than 88.2 - 44.1k. And I'm not talking about simply discarding, but about
getting at least the same level of quality.

So in theory all you have to do is invoke something like a windowed sinc
convolution and write a nice equation or two on a whiteboard. That works
fine for both. The mathematician can then put the pen down and walk away.
But when it comes to the real number crunching the two cases are very
different for similar quality.

e.g. in a TDA method, a power of two ratio downsample means you only need
one set of coefficients and the clocking is trivial along the arrays. Doing
96 - 44.1 requires rather more effort / complexity for the same output
quality even if you throw the same impulse function at them both. This may
matter both in terms of ensuring you've bugfixed and in terms of CPU
loading or the number of devices on the silicon and the power demand.


And of course every audio ADC uses a massively high sampling rate,
followed by the procedure above to arrive even at 44.1kHz sampling.

Erm, we weren't really talking about the ADC but digital-digital
conversions, probably by running software on a general computing system.
But that doesn't change the above point. And TBH even in hardware like an
ADC or DAC running at a high internal rate, simple integer rations make
good results easier to obtain with less number-bashing.

And of course high rate low-bit tends to risk problems like the ones which
can show up in the 'Health Checks' I did a while ago.

Jim

Any program recommendations for converting 96k/24 to 44k1/16 - non real time?
Ideally linux or mac - or any, if command line.

Arthur

--
Arthur Quinn
real-email arthur at bellacat dot com

In article , Arthur Quinn
wrote:
Any program recommendations for converting 96k/24 to 44k1/16 - non real
time? Ideally linux or mac - or any, if command line.

sox.

Highly recommended for any processing of audio.

It does pretty good conversions by default. It also gives you options to
control the method and aspects like dither PDF, etc. The source code is
also available if you wish to delve and confirm.

I think it uses the same code as 'secret rabbit' which you may find on the
web, and ALSA uses for conversions.

ffmpeg will also work, but its really aimed more generally at AV whereas
sox is aimed specifically at audio.

FWIW I also use sox for wave - flac conversions and it has a nice 'stats'
report that lets you detect things like any clipping, etc. Rate conversions
can lead to clipping if the waveforms have peaks too near to maximum.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

On 2015-07-13 15:19:30 +0100, Huge said:

On 2015-07-13, Arthur Quinn wrote:

[44 lines snipped]

Any program recommendations for converting 96k/24 to 44k1/16 - non real time?
Ideally linux or mac - or any, if command line.

http://linux.die.net/man/1/sox

Thanks - that seems to do the job.

Arthur

--
Arthur Quinn
real-email arthur at bellacat dot com

On 2015-07-13 16:31:06 +0100, Jim Lesurf said:

In article , Arthur Quinn
wrote:
Any program recommendations for converting 96k/24 to 44k1/16 - non real
time? Ideally linux or mac - or any, if command line.

sox.

Highly recommended for any processing of audio.

It does pretty good conversions by default. It also gives you options to
control the method and aspects like dither PDF, etc. The source code is
also available if you wish to delve and confirm.

I think it uses the same code as 'secret rabbit' which you may find on the
web, and ALSA uses for conversions.

ffmpeg will also work, but its really aimed more generally at AV whereas
sox is aimed specifically at audio.

FWIW I also use sox for wave - flac conversions and it has a nice 'stats'
report that lets you detect things like any clipping, etc. Rate conversions
can lead to clipping if the waveforms have peaks too near to maximum.

Jim

Thanks; it seems to be the ideal program for the job.

Arthur

--
Arthur Quinn
real-email arthur at bellacat dot com

On 14/07/2015 09:52, Arthur Quinn wrote:
On 2015-07-13 16:31:06 +0100, Jim Lesurf said:

In article , Arthur Quinn
wrote:
Any program recommendations for converting 96k/24 to 44k1/16 - non real
time? Ideally linux or mac - or any, if command line.

sox.

Highly recommended for any processing of audio.

It does pretty good conversions by default. It also gives you options to
control the method and aspects like dither PDF, etc. The source code is
also available if you wish to delve and confirm.

I think it uses the same code as 'secret rabbit' which you may find on
the
web, and ALSA uses for conversions.

ffmpeg will also work, but its really aimed more generally at AV whereas
sox is aimed specifically at audio.

FWIW I also use sox for wave - flac conversions and it has a nice
'stats'
report that lets you detect things like any clipping, etc. Rate
conversions
can lead to clipping if the waveforms have peaks too near to maximum.

Jim

Thanks; it seems to be the ideal program for the job.


I'd find the lack of a GUI difficult. FWIW I've used Max
http://sbooth.org/Max/ for years and don't recall any problems.

--
Cheers, Rob

In article , RJH
wrote:
FWIW I also use sox for wave - flac conversions and it has a nice
'stats' report that lets you detect things like any clipping, etc.
Rate conversions can lead to clipping if the waveforms have peaks too
near to maximum.

Jim

Thanks; it seems to be the ideal program for the job.


I'd find the lack of a GUI difficult. FWIW I've used Max
http://sbooth.org/Max/ for years and don't recall any problems.

Its certainly true that the manual (man page) for sox is pretty daunting.
But for basic tasks like resampling the commands are pretty simple.
Essentially you just add in, say, "-r 44100" if you want the output to be
44.1k. e,g

If you have a 96/24 wave file called 'wibble.wav' then

sox wibble.wav -r 44100 -b 16 wobble.flac

should give you a 44.1k / 16 bit flac file of the audio called
'wobble.flac'. You can easily spot that '-r' means 'rate' and '-b' bits.

I recommend adding 'stats' at the end of the command, so

sox wibble.wav -r 44100 -b 16 wobble.flac stats

does the same, then prints out some stats info on the result so you can see
things like the signals levels, etc.

The manual is long because sox can do so many things and give so much
control. But for tasks like this you can ignore most of that.

TBH for many GUI programs that process audio or AB my first reaction is to
wonder if they are a front-end for sox or ffmpeg. For all I know, 'Max' may
be so.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

On 14/07/2015 16:21, Jim Lesurf wrote:
In article , RJH
wrote:
FWIW I also use sox for wave - flac conversions and it has a nice
'stats' report that lets you detect things like any clipping, etc.
Rate conversions can lead to clipping if the waveforms have peaks too
near to maximum.

Jim

Thanks; it seems to be the ideal program for the job.


I'd find the lack of a GUI difficult. FWIW I've used Max
http://sbooth.org/Max/ for years and don't recall any problems.

Its certainly true that the manual (man page) for sox is pretty daunting.
But for basic tasks like resampling the commands are pretty simple.
Essentially you just add in, say, "-r 44100" if you want the output to be
44.1k. e,g

If you have a 96/24 wave file called 'wibble.wav' then

sox wibble.wav -r 44100 -b 16 wobble.flac

should give you a 44.1k / 16 bit flac file of the audio called
'wobble.flac'. You can easily spot that '-r' means 'rate' and '-b' bits.

I recommend adding 'stats' at the end of the command, so

sox wibble.wav -r 44100 -b 16 wobble.flac stats

does the same, then prints out some stats info on the result so you can see
things like the signals levels, etc.

The manual is long because sox can do so many things and give so much
control. But for tasks like this you can ignore most of that.


Thanks - not that difficult, then. The biggest issue for me is usually
paths.

TBH for many GUI programs that process audio or AB my first reaction is to
wonder if they are a front-end for sox or ffmpeg. For all I know, 'Max' may
be so.


So far as I can tell, it's this:

http://www.mega-nerd.com/libsndfile/

The rest of the components are encoders and a compressor.


--
Cheers, Rob

In article , RJH
wrote:
On 14/07/2015 16:21, Jim Lesurf wrote:
In article , RJH
wrote:


The manual is long because sox can do so many things and give so much
control. But for tasks like this you can ignore most of that.


Thanks - not that difficult, then. The biggest issue for me is usually
paths.

FWIW I tend to sort that by writing simple ROX apps so I can simply drag
the files to the application icon to give them as input to a chosen
command. That makes it easy to, say, convert a batch of files by
drag-and-drop.

But for individual files when using specific less common options/commands I
tend to find that I can find them via stepping around the terminal's
command history or tab completion. (No idea if Windows has equivalents.)


TBH for many GUI programs that process audio or AB my first reaction
is to wonder if they are a front-end for sox or ffmpeg. For all I
know, 'Max' may be so.


So far as I can tell, it's this:

http://www.mega-nerd.com/libsndfile/

The rest of the components are encoders and a compressor.

I'm not sure, but my first impression is that it will have source code
and/or methods in common with sox and ALSA. If nothing else it shares the
same yukky webpage colours, etc, as 'secret rabbit'. 8-]

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html