I've got a head-scratcher in the continuing quest to get my vinyl on
to CDs and hope that someone has a suggestion for a fix
As I understand it, audio CDs use 44.1kHz 16-bit files. If my recording
s/w is set to this though there's an unpleasant ringing on high frequencies,
eg female vocals. At 96kHz PCM 16-bit sampling the sound is good,
so it's possible to create hi-fi .wav files, although the writer s/w rejects
them as CD- incompatible. Hmmm. The file can be converted to
44.1kHz, but the ringing comes back. However, it can now be written
as a .cda track
I've had a look at CD Recorders, eg Philips, Marantz, that use 96kHz
24-bit DSPs. Somewhere in the h/w that sampling rate is knocked
back to 44.1kHz whilst retaining the fidelity of the original. But starting
price for good ones is $2000
It "may" be the sound card, but as the source and 96kHz sampling
sound OK and its settings are at "Best", I'm sceptical. I've tried 3
different programs - Polderbits, Goldwave and Windows Sound
Recorder - with identical results
> I've got a head-scratcher in the continuing quest to get my vinyl on
> to CDs and hope that someone has a suggestion for a fix
>
> As I understand it, audio CDs use 44.1kHz 16-bit files. If my recording
> s/w is set to this though there's an unpleasant ringing on high
frequencies,
> eg female vocals. At 96kHz PCM 16-bit sampling the sound is good,
> so it's possible to create hi-fi .wav files, although the writer s/w
rejects
> them as CD- incompatible. Hmmm. The file can be converted to
> 44.1kHz, but the ringing comes back. However, it can now be written
> as a .cda track
>
> I've had a look at CD Recorders, eg Philips, Marantz, that use 96kHz
> 24-bit DSPs. Somewhere in the h/w that sampling rate is knocked
> back to 44.1kHz whilst retaining the fidelity of the original. But
starting {Quote hidden}
> price for good ones is $2000
>
> It "may" be the sound card, but as the source and 96kHz sampling
> sound OK and its settings are at "Best", I'm sceptical. I've tried 3
> different programs - Polderbits, Goldwave and Windows Sound
> Recorder - with identical results
>
> Any thoughts ? TIA
>
> --
> http://www.piclist.com hint: The list server can filter out subtopics
> (like ads or off topics) for you. See http://www.piclist.com/#topics
I assume the ringing you hear is on your PC and its playback gear.
If it were me I'd go ahead and burn a standard audio CD and then play
it back on a different system, preferably a quality hi-fi system, and
see if the ringing is heard on playback there too.
>I've got a head-scratcher in the continuing quest to get my vinyl on
>to CDs and hope that someone has a suggestion for a fix
>
>As I understand it, audio CDs use 44.1kHz 16-bit files. If my recording
>s/w is set to this though there's an unpleasant ringing on high frequencies,
>eg female vocals. At 96kHz PCM 16-bit sampling the sound is good,
>so it's possible to create hi-fi .wav files, although the writer s/w rejects
>them as CD- incompatible. Hmmm. The file can be converted to
>44.1kHz, but the ringing comes back. However, it can now be written
>as a .cda track
>
>I've had a look at CD Recorders, eg Philips, Marantz, that use 96kHz
>24-bit DSPs. Somewhere in the h/w that sampling rate is knocked
>back to 44.1kHz whilst retaining the fidelity of the original. But starting
>price for good ones is $2000
>
>It "may" be the sound card, but as the source and 96kHz sampling
>sound OK and its settings are at "Best", I'm sceptical. I've tried 3
>different programs - Polderbits, Goldwave and Windows Sound
>Recorder - with identical results
>
>Any thoughts ? TIA
>
>--
>http://www.piclist.com hint: The list server can filter out subtopics
>(like ads or off topics) for you. See http://www.piclist.com/#topics
It sounds (no pun intended) like there is a high frequency component that is being aliased down into the audio frequency range by the sampling process.
You could take a look at the spectrum of one of your 96kHz records to try to determine what the unwanted frequency component is. This might give you a hint as to where it is coming from. Maybe there is a natural resonance in the stylus pickup that you would normally not hear, but that is being digitized and aliased.
> I assume the ringing you hear is on your PC and its playback
> gear. If it were me I'd go ahead and burn a standard audio CD
> and then play it back on a different system, preferably a quality
> hi-fi system, and see if the ringing is heard on playback there too.
It is. The ringing is in the file
One test piece I use is "Solveig's Song" from Peer Gynt. At 44.1kHz
the soprano sounds like she's singing through a cheese grater, but
at 96kHz it's pretty good
> It sounds (no pun intended) like there is a high frequency
> component that is being aliased down into the audio
> frequency range by the sampling process
I've just done a test by recording a piano on FM radio to test the tone
arm/cartridge theory. The ringing is still evident at 44.1kHz and not
at 96kHz. Yours and Bob's aliasing thoughts are duly noted though,
possibly it's happening somewhere else
Jinx wrote:
.
> It is. The ringing is in the file
>
> One test piece I use is "Solveig's Song" from Peer Gynt.
> At 44.1kHz the soprano sounds like she's singing through
> a cheese grater, but at 96kHz it's pretty good.
Jinx,
I recorded "Solveig's Song" with "Cool Edit 2000"
freeware few months ago.
Record options were 32 bit (float) 44100. Then I used
dynamic processing, because of too wide dynamic range:
did you compare levels of quiet "Solveig's Song" with the
crazy "King's Daughters Dance".
"Save" options are numerous. I used 256 Kbps 44100 (mp3)
at Quality Level 9 (high quality, low speed of saving).
Vinyl disk was Helen Donat - soprano, Eshenbach -
conductor. (1985).
PC: Cel 1.7, M/B I845 with Realtek AC97 sound.
All was great.
On 9 May 2003 at 13:06, Jinx wrote:
> As I understand it, audio CDs use 44.1kHz 16-bit files. If my recording
> s/w is set to this though there's an unpleasant ringing on high frequencies,
> eg female vocals. At 96kHz PCM 16-bit sampling the sound is good,
> so it's possible to create hi-fi .wav files, although the writer s/w rejects
> them as CD- incompatible. Hmmm. The file can be converted to
> 44.1kHz, but the ringing comes back. However, it can now be written
> as a .cda track
It's aliasing.
Your sound card probably doesn't have a very good lowpass filter, or
it's programmable and your software driver is not setting it correctly.
Build yourself a good lowpass filter that is -30db or better by 22KHz and
all will be better.
Hm, I did all that (Recording, converting, removing clicks from
scratches, noise, adding+naming+numbering tracks) with an on-board
sound-card and a cheap CD-burner with the included software (Create CD).
No problems at all, worked just fine.
>
> I've got a head-scratcher in the continuing quest to get my vinyl on
> to CDs and hope that someone has a suggestion for a fix
>
> As I understand it, audio CDs use 44.1kHz 16-bit files. If my recording
> s/w is set to this though there's an unpleasant ringing on high frequencies,
> eg female vocals. At 96kHz PCM 16-bit sampling the sound is good,
> so it's possible to create hi-fi .wav files, although the writer s/w rejects
> them as CD- incompatible. Hmmm. The file can be converted to
> 44.1kHz, but the ringing comes back. However, it can now be written
> as a .cda track
>
> I've had a look at CD Recorders, eg Philips, Marantz, that use 96kHz
> 24-bit DSPs. Somewhere in the h/w that sampling rate is knocked
> back to 44.1kHz whilst retaining the fidelity of the original. But starting
> price for good ones is $2000
>
> It "may" be the sound card, but as the source and 96kHz sampling
> sound OK and its settings are at "Best", I'm sceptical. I've tried 3
> different programs - Polderbits, Goldwave and Windows Sound
> Recorder - with identical results
>
> Any thoughts ? TIA
>
> --
> http://www.piclist.com hint: The list server can filter out subtopics
> (like ads or off topics) for you. See http://www.piclist.com/#topics
-- SyWiTec * Systeme für Wissenschaft und Technik
Bamberg & Monsees GbR
Steindamm 22 * D-28719 Bremen
Fon +49-421-646775 * Fax +49-421-646785 http://www.sywitec.de/
Your's and Herbert's success seems to indicate that there's
something in my system that's not performing well. Whether
that's internal or external to the PC won't be known until I try
a few other tests. I'll record a signal generator to find out if
there's actually a problem with the b/w and take it from there
The sampling rate is decreased by a process called decimation. When the
input and the output sampling frequencies are not harmonic the decimation
must be very good so as not to produce extra distortion and artifacts,
especially in signals near the high cutoff of the output sampling
frequency.
In theory when sampling at 44.1 kHz this problem should not happen but
many soundcards cannot do 44.1 kHz exactly in hardware and convert to 44.1
kHz afterwards (using resampling - again - in the driver). So you have the
same problem.
If nothing really works maybe you can try adding high frequency noise to
your 96 kHz track before resampling (noise at 23 kHz or such, Bw = 1kHz
and -40dB or lower). This should dither the higher frequency components
during resampling so the resampling artifacts should be dithered too. It's
worth a try.
I am *sure* that by using a higher end audio card all the problems will go
away. You will likely not be able to duplicate its results by fudging with
the resampling.
> I recorded "Solveig's Song" with "Cool Edit 2000"
> freeware few months ago.
> Record options were 32 bit (float) 44100. Then I used
> Vinyl disk was Helen Donat - soprano, Eshenbach -
> conductor. (1985).
> PC: Cel 1.7, M/B I845 with Realtek AC97 sound.
> All was great.
>
> Mike.
Mine is Adele Stolte, Gewandhaus Orch, cond Neumann, 1972
Even when I think I've made some minor improvement to the
recording settings that doesn't sound too bad with Joe Walsh,
Adele Stolte's broken high notes soon put an end to the happy
face. I'll try some of the processing options you did, see if that
makes a difference. Otherwise I may have a look at another
sound card. It's so close now to be annoying
I didn't read in detail all the thread, but anyone thought about an
eventual CODEC?
Are you using pure WAV format or any of the CODECs that Windows programs
use to compact WAV files' data?
This is, for sure, a source of ringing in some frequencies, and as it is
stated that the problem is only on the file, this is the most probable
cause of ringing, for my experience, of course.
>> I recorded "Solveig's Song" with "Cool Edit 2000"
>>freeware few months ago.
>>
>>
>
>
>
>> Record options were 32 bit (float) 44100. Then I used
>>
>>
>
>
>
>> Vinyl disk was Helen Donat - soprano, Eshenbach -
>>conductor. (1985).
>> PC: Cel 1.7, M/B I845 with Realtek AC97 sound.
>> All was great.
>>
>> Mike.
>>
>>
>
>Mine is Adele Stolte, Gewandhaus Orch, cond Neumann, 1972
>
>Even when I think I've made some minor improvement to the
>recording settings that doesn't sound too bad with Joe Walsh,
>Adele Stolte's broken high notes soon put an end to the happy
>face. I'll try some of the processing options you did, see if that
>makes a difference. Otherwise I may have a look at another
>sound card. It's so close now to be annoying
>
>--
>http://www.piclist.com hint: The PICList is archived three different
>ways. See http://www.piclist.com/#archives for details.
>
>
>
>
Did you try a different recording software ? Maybe NERO just for a
test. If it's still there, it's either the soundcard or the combination
between your pre-amplifier and the sound-card. Can you check the input
signal (to line-in of the soundcard) with an o'scope ? And compare
it to the signal an line-out of the sound-card.
>
> > worked just fine.
> > regards,
> > Reelf
>
> Your's and Herbert's success seems to indicate that there's
> something in my system that's not performing well. Whether
> that's internal or external to the PC won't be known until I try
> a few other tests. I'll record a signal generator to find out if
> there's actually a problem with the b/w and take it from there
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways. See http://www.piclist.com/#archives for details.
-- SyWiTec * Systeme für Wissenschaft und Technik
Bamberg & Monsees GbR
Steindamm 22 * D-28719 Bremen
Fon +49-421-646775 * Fax +49-421-646785 http://www.sywitec.de/
So far only Peter and I have touched on the basics of the issue here -
that you absolutely must not introduce signal to the A/D converter above
half the sample frequency.
All A/D converters must (by international law :) be preceded by a lowpass
filter. Try recording sine waves to see if yours are working.
Don't do it at anything higher than -10 db though, audio equipment rarely
is designed for full-volume 20KHz signals!
> I've made a start on testing the sound system. Here are
> some traces of various frequencies and what happens
> to them after sampling
>
> http://home.clear.net.nz/pages/joecolquitt/gw_samples.html
>
> It appears that modulation gets increasingly worse
Aha! This tells all. Notice the 22.5KHz @ 44.1Ksps wave? There is
one! There shouldn't be anything, because the lowpass filter on your
sound card should have cut off before 22.05KHz.
So either your sound card is crappy, or the driver isn't turning on the
filter.
Get another sound card or build yourself a lowpass filter for the input to
this one.
Cheerful regards,
Bob
P.S. You might give Cool Edit a try if you are looking for a good sound
program for the PC. It won't fix your sound card, but it's a good
program.
> Aha! This tells all. Notice the 22.5KHz @ 44.1Ksps wave? There is
> one! There shouldn't be anything, because the lowpass filter on your
> sound card should have cut off before 22.05KHz.
I'd have expected the 22k5 @ 44.1 to be considerably more chopped
up than that
> So either your sound card is crappy, or the driver isn't turning on the
> filter.
I've looked right through the system and can't find anything to tweak.
Possibly being built onto the m/b it really is just low-end and this is
all I can expect
> Get another sound card or build yourself a lowpass filter for the input to
> this one.
I went out window shopping at the wholesalers this afternoon. I quite
like the look of the Creative Audigy series. There's one that has an
internal 24-bit 96kHz and it's within budget
> Cheerful regards,
>
> Bob
>
> P.S. You might give Cool Edit a try if you are looking for a good
> sound program for the PC. It won't fix your sound card, but it's a
> good program.
Will do. I'll see what's bundled with the sound card though. The SBs
have Creative 3, which I'll look up
On Saturday 10 May 2003 00:09, Jinx scribbled:
> I went out window shopping at the wholesalers this afternoon. I quite
> like the look of the Creative Audigy series. There's one that has an
> internal 24-bit 96kHz and it's within budget
>
> ....
>
> Will do. I'll see what's bundled with the sound card though. The SBs
> have Creative 3, which I'll look up
Check out the SB Live! 5.1 cards. Supposedly very similar technology and audio quality to the SB Audigy and much, much lower-cost.
Price ranged from $60 to "expensive". After carefully reading
spec sheets (and trying not to get boggled by the acronyms,
buzz words and ad-speak) one from the middle would suit me
I actually have never had a problem with even the cheapest soundcard...even
the old
ISA cards. I would suggest you try cool edit pro before you buy a new
soundcard.
There really are a bunch of options for downsampling the audio, I am sure
you will
find one that works. Even if you dont, there are such great filters that
you can
probably fix whatever artifacts you are getting. It is fun buying new
hardware though,
like I said the SB Live 5.1 platinum is nice, but there are pleanty of other
cards out
there that have teriffic sound and features.
At 05:09 PM 5/10/2003 +1200, you wrote:
> > Aha! This tells all. Notice the 22.5KHz @ 44.1Ksps wave? There is
> > one! There shouldn't be anything, because the lowpass filter on your
> > sound card should have cut off before 22.05KHz.
>
>I'd have expected the 22k5 @ 44.1 to be considerably more chopped
>up than that
I think that GoldWave is probably doing some kind of interpolation when
generating highly zoomed-in displays. This could also partly explain the
"modulation" that you see. If your sampling rate is close to twice the
input frequency, then there will be a slow periodic variation of peak
sample magnitudes (a mere consequence of sampling, just like Moire patterns
in scanning). When GoldWave attempts to "connect the dots", it probably
ends up looking like a modulated sine wave, rather than the periodically
varying highly choppy thing you'd expect.
Sean H. Breheny <spamBeGoneshb7spamBeGoneCORNELL.EDU> wrote:
> At 05:09 PM 5/10/2003 +1200, you wrote:
> > I'd have expected the 22k5 @ 44.1 to be considerably more chopped
> > up than that
>
> I think that GoldWave is probably doing some kind of interpolation when
> generating highly zoomed-in displays. This could also partly explain the
> "modulation" that you see. If your sampling rate is close to twice the
> input frequency, then there will be a slow periodic variation of peak
> sample magnitudes (a mere consequence of sampling, just like Moire patterns
> in scanning). When GoldWave attempts to "connect the dots", it probably
> ends up looking like a modulated sine wave, rather than the periodically
> varying highly choppy thing you'd expect.
See: www.circuitcellar.com/library/eq/151/index.asp
Problem #7 (and its answer)
That's a nice example, thanks! What I was trying to say, though, was that
GoldWave, instead of showing the sample points, just makes the waveform
look like a continuous line, regardless of how much you zoom in. If you saw
just points when you got down to the single sample level, I think things
would be less misleading, since you would still notice how few samples
there were per cycle. It would not appear like a well oversampled modulated
sine wave.
So, the core of the problem is a necessary consequence of sampling, but the
interpolation (what I'm calling drawing lines to connect the samples in
GoldWave's display) tends to mask the problem and make it harder to
diagnose at a glance.
>Sean H. Breheny <shb7EraseME.....CORNELL.EDU> wrote:
> > I think that GoldWave is probably doing some kind of interpolation when
> > generating highly zoomed-in displays. This could also partly explain the
> > "modulation" that you see. If your sampling rate is close to twice the
> > input frequency, then there will be a slow periodic variation of peak
> > sample magnitudes (a mere consequence of sampling, just like Moire patterns
> > in scanning). When GoldWave attempts to "connect the dots", it probably
> > ends up looking like a modulated sine wave, rather than the periodically
> > varying highly choppy thing you'd expect.
>
>See: www.circuitcellar.com/library/eq/151/index.asp
>Problem #7 (and its answer)
>
>-- Dave Tweed
>
>--
>http://www.piclist.com#nomail Going offline? Don't AutoReply us!
>email EraseMElistservmitvma.mit.edu with SET PICList DIGEST in the body
Jinx wrote:
> Yeah, I looked at basic SB through Live 5.1 to the Special
> Editions, Platinum Drives (nice) etc
>
> http://www.soundblaster.com/products/audigy/
>
> Price ranged from $60 to "expensive". After carefully reading
> spec sheets (and trying not to get boggled by the acronyms,
> buzz words and ad-speak) one from the middle would suit me
By the way, if your M/B has ISA bus, you may try
old good (really) Sound Blaster AWE 64. It's price
is less then zero now.
> > By the way, if your M/B has ISA bus, you may try
> > old good (really) Sound Blaster AWE 64. It's price
> > is less then zero now.
> >
> > Mike.
>
> It's got everything but ISA ;-(
>
> --
> http://www.piclist.com#nomail Going offline? Don't AutoReply us!
> email EraseMElistservspamspamBeGonemitvma.mit.edu with SET PICList DIGEST in the body
> Have you played the record backwards to see if the
> 'hidden message' is causing the problem?
>
> Pookie
Not going there again - I got into all kinds of strife with
netherworld denizens and "cisuM fO dnuoS ehT". Julie
Andrews so looks like butter wouldn't melt too
CLEAR.NET.NZ>
