Seeking advice
tr...@my-deja.com
I briefly looked at the archives of this newsgroup to see if my question
is asked before. Did not see anything. So here are a few questions for
you experts out there after some background info.
I am about to transfer some old cassettes and LP's to digital media
(CD's). Other than using commercially made CD's and getting exposed to
mp3, Real Audio, etc. on the net, I do not know much about digital
media, especially about quality issues.
The material I will be transferring is rare music. I would like to
retain as much of the quality as possible. However, the tapes that I
have are, in some cases, 30 years old, and the LP's I have probably have
quite a few scratches. In other words, the quality may not be there to
start off with
While I am investigating what format to use (comments on that also would
be appreciated), I have a question about hardware:
* Perhaps the general question is if the quality of the soundcard
matters. I currently have a Creative Sound Blaster PCI 128, which I
consider to be a low-end one.
* If the soundcard matters, is there anyplace on the net that ranks the
quality of the various sound cards? I am curious as to how much it
matters...
* What else matters in terms of the computer equipment I will use?
Another factor here of course, is the quality of the analog equipment I
will be using. I currently have a home stereo, nothing fancy. I have two
Sony portables (TC 153 and 158).
Any comments would be appreciated.
Al
Sent via Deja.com
http://www.deja.com/
Walter Knapp
> I am about to transfer some old cassettes and LP's to digital media
> (CD's). Other than using commercially made CD's and getting exposed to
> mp3, Real Audio, etc. on the net, I do not know much about digital
> media, especially about quality issues.
>
> The material I will be transferring is rare music. I would like to
> retain as much of the quality as possible. However, the tapes that I
> have are, in some cases, 30 years old, and the LP's I have probably have
> quite a few scratches. In other words, the quality may not be there to
> start off with
>
> While I am investigating what format to use (comments on that also would
> be appreciated), I have a question about hardware:
>
> * Perhaps the general question is if the quality of the soundcard
> matters. I currently have a Creative Sound Blaster PCI 128, which I
> consider to be a low-end one.
>
> * If the soundcard matters, is there anyplace on the net that ranks the
> quality of the various sound cards? I am curious as to how much it
> matters...
>
> * What else matters in terms of the computer equipment I will use?
>
> Another factor here of course, is the quality of the analog equipment I
> will be using. I currently have a home stereo, nothing fancy. I have two
> Sony portables (TC 153 and 158).
This is actually quite a common question. Quite a few are doing this,
including me. You can get quite high quality copies if you spend the
time and have good software. It's a bit discouraging at first, and you
will probably end up redoing your first efforts as you learn. Once you
sort it all out, it's not tough, though a bit tedious.
First off, here's a link to a page discussing going from LP to audio CD.
This is the route you will want to head, and the tapes would be handled
much the same as the LP's. Or at least the differences are obvious:
http://homepages.nildram.co.uk/~abcomp/lp-cdr.htm
This page will give you a pretty good idea what you are in for.
Quality of soundcard matters. You need CD quality, 44.1khz, 16bit
sampling. Do some test recording off some good CD's to see if your's is
up to it. You, of course, don't need to get too carried away as the
quality of the originals is only a certain level.
If you don't have any, you need a good set of headphones. You will be
listening to the fine details of the audio, trying to see if the filter
was set right, so on. Good comfortable headphones are a must. The kind
that provide good isolation are best.
You will need a good sound editor, preferably with some good noise
filters, decrackle and declick filters. It can be tedious work if you
want good clean copies, and this is highly dependent on the software. Do
not jump in with the first editor you see, but investigate the process
first to see what has the features that you will need and how good they
are. And once you get started, you will find it takes some practice to
get ideal editing. The automatic filters rarely do a first class job. Or
at least require a lot of understanding of the settings.
Since it sounds like you are using a Windows PC, you will want to bring
in the audio as .wav files. These preserve the full quality of whatever
you can get from the originals. And note, these are uncompressed audio,
they will take lots of disk space, about a gig to bring in and edit one
CD full of stereo. Do not store the audio in any of the compressed
formats until you have made as good a copy as you can on CD. The 640megs
of the CD plus room for scratch files and so on is where the gig goes.
So you need to plan for large storage needs, and high quality archiving,
not just HD. You can use CD-R, but it's not rewriteable, a problem for
the stuff you are actively editing. I use optical disks for this. And
archive to CD-R once I'm done.
I use macs, so my software choices won't help you. CoolEdit Pro is often
recommended on PC's, but I'm not sure how far you can get with it as far
as filtering and noise problems. It will certainly handle getting the
audio into digital form.
As far as the stereo system end, try to get it working as clean as
possible, clean the records carefully, tune up the turntable, so on.
Every bit of noise, hum, clicks and crackles you eliminate this way
really helps. After you try some editing, you will probably go back and
clean and tune some more...
Particularly with the tape, you may want to get set up and record it all
to computer files before starting to edit. If you have lots it could
take some time to get through the editing process and getting it into
digital form will keep from any farther deterioration of the tape. CD-R
is cheap enough you could use those to archive the unedited files. Then
work through them without having to worry that what you have not gotten
to is going bad. The process does tend to break naturally into the two
parts, first, getting it into digital files, and second cleaning them
up. Of course burning to CD at the end is trivia, the work is done by then.
$p@m H8er
difference, but may not be significant enough for what you're doing. I'll
recommending checking out my 'Recording Info' web page (main page at end of
this message).
You'll probably wanna get a good audio editing package to record all this,
as sound recorder isn't robust enough. They're also linked on my web page.
You'll need to have a fast hard drive to make sure you don't lose any of the
audio due to your hard drive not keeping up with your extensive
recording.....
If you need more assistance, e-mail me as directed below, or from my web
page.....
Brian
tr...@my-deja.com wrote:
> I am about to transfer some old cassettes and LP's to digital media
> (CD's). Other than using commercially made CD's and getting exposed to
> mp3, Real Audio, etc. on the net, I do not know much about digital
> media, especially about quality issues.
>
> The material I will be transferring is rare music. I would like to
> retain as much of the quality as possible. However, the tapes that I
> have are, in some cases, 30 years old, and the LP's I have probably have
> quite a few scratches. In other words, the quality may not be there to
> start off with
>
> While I am investigating what format to use (comments on that also would
> be appreciated), I have a question about hardware:
>
> * Perhaps the general question is if the quality of the soundcard
> matters. I currently have a Creative Sound Blaster PCI 128, which I
> consider to be a low-end one.
>
> * If the soundcard matters, is there anyplace on the net that ranks the
> quality of the various sound cards? I am curious as to how much it
> matters...
>
> * What else matters in terms of the computer equipment I will use?
>
> Another factor here of course, is the quality of the analog equipment I
> will be using. I currently have a home stereo, nothing fancy. I have two
> Sony portables (TC 153 and 158).
--
Brian S. Graham Entertainment
Spam-Haters reply to: dj at knoxvillewedding dot com
Visit my DJ/Audio site at http://www.knoxvillewedding.com
The only kind of SPAM I want comes from Hormel, thank you!
Fatman0
www.audioadvisor.com .
Then a good home recording cd deck. I use a $300 Panasonic. I uses only
AUDIO recordable cd's I use
CDR and CDRW. I record to cdrw on my nice home system. Then take the cdrw
to my computer and read it in with Feurio. I use POPFIX, using two passes
(the third pass seems to change the sound a bit) to remove pops and clicks.
I then use Feurios track editor to split the wav files into tracks, if I did
not mark them while recording on the CD recorder deck. Then burn to a
cheaper computer CDR. If I want a very permanent copy, I burn to CDRW,
which should last longer than the dye based cdr's.
Works for me.
I have also tried minidisc. It can't handle wide range music from an lp.
So I stick with CD recordable.
I do this alot, and I try to keep it simple. You can email me if you like.
Fred
<tr...@my-deja.com> wrote in message news:91k5g5$tub$1...@nnrp1.deja.com...
Mark D. Zacharias
players won't read the RW discs.
Additionally there may be cost constraints. Perhaps the fellow can't afford
an extra 200.00 or more for the cleaner, as well as the extra cost of an
audio-only burner in addition to his computer CD-R drive.
Given that cleaning the records the best possible is desirable, many people
use some variation of this method:
1. Plug your turntable into a decent integrated amp or receiver's phono
inputs.
2. Take the Record Out signal from the back of the amp/receiver via an
adaptor converting the RCA cable to 1/8 inch stereo plug, or a complete
cable designed for this,
3. Plug directly into the sound card "Line In" jack.
4. Use the recording software which came with your sound card to record the
tracks. Other recording software is available if you prefer. Download from:
http://www.hitsquad.com/smm/
5. To preserve track spacing, record each record side as one large wave
file, then add track markers and divide using DC Art, Cool Edit or some
other appropriate software. The same software will have pop and click
removal features. The above site will also have demo versions of this
sofware.
6. Burn the disc in Disc-At-Once mode to avoid an extra 2-second spacing
between tracks.
Mark Z.
Fatman0 <fat...@my-deja.com> wrote in message
news:F1B%5.1909$xU2.1...@news1.atl...
Dré
> Hello:
(Changed the subject line so it actually covers the content)
> I briefly looked at the archives of this newsgroup to see if my question
> is asked before. Did not see anything.
(Odd, this question comes around at RAT at least a few times a week).
> So here are a few questions for
> you experts out there after some background info.
> I am about to transfer some old cassettes and LP's to digital media
> (CD's). Other than using commercially made CD's and getting exposed to
> mp3, Real Audio, etc. on the net, I do not know much about digital
> media, especially about quality issues.
> The material I will be transferring is rare music. I would like to
> retain as much of the quality as possible. However, the tapes that I
> have are, in some cases, 30 years old, and the LP's I have probably have
> quite a few scratches. In other words, the quality may not be there to
> start off with
Well.... That's one of the nice things you can do with a computer. Scratches can
actually be manually removed using (for example) the pen funtion in Sound Forge.
Just use the <Find> routine to find the offending glitch (or, use your ears and
a set of good headphones OR speakers).
> While I am investigating what format to use (comments on that also would
> be appreciated), I have a question about hardware:
> * Perhaps the general question is if the quality of the soundcard
> matters. I currently have a Creative Sound Blaster PCI 128, which I
> consider to be a low-end one.
It actually is quite sufficient for the use you have for it. It's A/D converters
(only part of soundcard which is interesting for this purpose) is fairly good
and quite probably outperforms the equipment connected to it (cassette or TT).
> * If the soundcard matters, is there anyplace on the net that ranks the
> quality of the various sound cards? I am curious as to how much it
> matters...
There's always Arny's site: www.pcavtech.com which goes into great detail about
the technical side of soundcards (S/N, various distortion types, freq.
response).
> * What else matters in terms of the computer equipment I will use?
What matters VERY much is the OS used. I have done quite a few (as in several
hundreds of hours thusfar) analog recordings (transfering the Vinyl part of my
(quite substantial) collection to a more portable medium (CD)) and one of the
biggest problems I encountered was missing frames. This is NOT due to the speed
of the harddrive (the 172k stream can hardly be considered "state of the art"
(and every newish harddrive has delayed calibration nowadays)) but it's actually
due to the piss poor interrupt handling of Win9x. This goes for Win95 as well as
for Win98 (no idea about Win ME, lost appetite for these Win versions).
So, as mentioned, one of the first problems I encountered were the missing
frames (parts, chunks) in the middle of the music (pretty much at random, once
every 4 recordings or so (average 5 minutes per recording)). I at that time
decided to audit every piece of recorded audio which was quite tedious. Someone
(don't remember who) pointed me to the fact that Win9x suffers from poor
interrupt handling whereas WinNT doesn't so I switched to WinNT (4). Problems
were GONE (with same harddrive BTW!!!) and never EVER surfaced again (not even
when the harddrive was being defragged at the same time). So I would suggest you
either audit every recording OR go for WinNT (or Win2000 which also works like a
charm (successor of Win NT 4)).
BTW: This problem is also NOT related to the recording software used since it
appears with: Cool Edit 96, Cool Edit Pro, Sound Forge 4.5, Wavelab 2.0 and
Goldwave. This problem CAN be related to the soundcard used. The cards of which
I KNOW they exhibit this problem (under Win9x) are: SB 16, 64, 128 and SBLive!
> Another factor here of course, is the quality of the analog equipment I
> will be using. I currently have a home stereo, nothing fancy. I have two
> Sony portables (TC 153 and 158).
There are a LOT of different things that can turn your effort into a mess. One
of the most often seen problems is groundloops. Someone hooks up his Hi-Fi
system to his computer and everything turns into "hummvile". Reason for this is
that Hi-Fi systems are grounded through the CATV connection. Quick check for
removing "hummvile" is to remove the CATV connection (BOTH video and radio).
Lasting fix is to insert a groundloop isolator in both the radio and TV feed (be
aware of problems when you use your cable for say Internet connection).
Then there is the quality of the RIAA(A) amp used. Quite often, the quality of
the RIAA(A) pre-amp inside a Hi-Fi unit suffices but if they don't suit your
needs, you need to go shopping for an external RIAA(A) amp.
Connections: Keep them as short as possible. Especially in the case of the 1/8"
jack plugs, do insert/remove them at least several dozens of times before
leaving them in the appropriate hole (for their connection (and cleaning action
during insertion) is piss poor).
Recording: Open up the software mixer of your soundcard. Be sure to select the
RECORDING mixer (looking at the playback mixer is quite useless ;-). MUTE
everything except the input which is going to be used for recording.
Now fire up your favorite wav recorder. We're first going to set the level. Put
on a record and play the thing. Now adjust the appropriate slider in the
software mixer (recording panel) until the level peaks during the loudest
passages at -2dB (Cool Edit and several other packages have an internal VU
meter). The 2dB is to allow for some headroom.
You have now set the recording level. No press stop on the TT (while keeping the
needle on the record) and with Cool Edit, record a few seconds. Now select these
few seconds and use Cool Edit's FFT (Fast Fourier) utility to analyse the
content. There should be NO big bumps in the spectrum (especially around 60,
120, 180.... Hz) and the level should be around -100dB (lower is better).
If all that is OK then you have set-up the computer end of your recording
efforts.
Now consider a few other things: Use a FRESH needle. It might be worth
considering to increase the needle pressure to the highest limit as suggested by
the needle's manufacturer (since that will result in the lowest distortion).
Record wear is a non issue here since you're only going to play it once.
Remember to also readjust the anti-skating accordingly.
The simple way is to use the reading on the anti-skating dial, the nice way is
to use a test recording and look with a spectrum analyser (like the one inside
Cool Edit) at the output. Set anti-skating so distortion is lowest in BOTH
channels (too high and left channel will distort, too low and right channel will
distort).
Clean your Vinyl THOROUGHLY. I personally use a carbon brush for that but there
are various other items worth considering as well (just don't go overboard).
Playing wet might be someting to consider IF you know the downsides of wet
playing. I don't, so I don't.
OK, now you're ready to record. After finishing your recordings, listen to them
(for starters because of the Win9x interrupt flaw) for pops and ticks. I
personally use Sound Forge for this part since Sound Forge has a very nice PEN
function. With the pen function you can literally draw a wav file (or apply
corrections to them). If you hear a pop/tick, move the cursor to the offending
part, zoom in (using the zoom function) and, using the pen function, redraw the
offending tick part (quite often only in ONE of the channels) with a shape which
you think MOST resembles what it should have been. You'll be AMAZED with what
you can get away with using this function.
NEVER rely on automated features for pop/tick removal. They'll always degrade
the ENTIRE recording (and often don't discriminate very well between transient
signals and actual ticks/pops, thus resulting in a dull and tediously bad
recording).
If you want to apply noise reduction, the one inside Cool Edit PRO works like a
charm (don't know about CE2000). You DO have to look at the settings first BTW
(and of course get a "fingerprint" from a quiet part).
For CD-burning, I would recommend using Feurio. It's about the greatest audio CD
recording software out there (on a WinTel platform that is).
Well, that's about it I think?
Some adresses:
www.feurio.com
www.syntrillium.com (Cool Edit)
www.goldwave.com
www.steinberg.net (WaveLab)
www.sonicfoundry.com (SoundForge)
--
André Huisman
New Line licht & geluid
hui...@new-line.nl
http://www.new-line.nl
--- pardon my French, I'm Dutch ---
Dave Platt
>> The material I will be transferring is rare music. I would like to
>> retain as much of the quality as possible. However, the tapes that I
>> have are, in some cases, 30 years old, and the LP's I have probably have
>> quite a few scratches. In other words, the quality may not be there to
>> start off with
>
>Well.... That's one of the nice things you can do with a computer. Scratches can
>actually be manually removed using (for example) the pen funtion in Sound Forge.
>Just use the <Find> routine to find the offending glitch (or, use your ears and
>a set of good headphones OR speakers).
I'd strongly encouraging a good, careful _physical_ cleaning of the
album prior to conversion. It's much better to get the gunk out of
the grooves, and eliminate the cause of many of the pops and ticks,
rather than trying to remove them via software.
If you're serious about doing these sorts of transfers from vinyl -
and _especially_ if you're interested in buying used vinyl to complete
your collection - I think you'd find a good wet-wash/vacuum system to
be a very worthwhile investment. I use the Record Doctor II (from
Audio Advisor - it's basically a low-end Nitty Gritty). VPI also has a
good reputation, and I've heard that the Disc Doctor manual cleaning
system (without vacuum) works well also.
Manual glitch-finding-and-removal can produce very good results, but
is extremely labor-intensive. I haven't really liked any of the
automatic glitch- finders I've used - they either miss a fair number
of pops and ticks, or mistakenly "fix" actual musical transients.
Setting the filter / selection sensitivity controls requires a lot of
trial and error.
>> * What else matters in terms of the computer equipment I will use?
>
>What matters VERY much is the OS used. I have done quite a few (as in several
>hundreds of hours thusfar) analog recordings (transfering the Vinyl part of my
>(quite substantial) collection to a more portable medium (CD)) and one of the
>biggest problems I encountered was missing frames. This is NOT due to the speed
>of the harddrive (the 172k stream can hardly be considered "state of the art"
>(and every newish harddrive has delayed calibration nowadays)) but it's actually
>due to the piss poor interrupt handling of Win9x. This goes for Win95 as well as
>for Win98 (no idea about Win ME, lost appetite for these Win versions).
>
>So, as mentioned, one of the first problems I encountered were the missing
>frames (parts, chunks) in the middle of the music (pretty much at random, once
>every 4 recordings or so (average 5 minutes per recording)). I at that time
>decided to audit every piece of recorded audio which was quite tedious. Someone
>(don't remember who) pointed me to the fact that Win9x suffers from poor
>interrupt handling whereas WinNT doesn't so I switched to WinNT (4). Problems
>were GONE (with same harddrive BTW!!!) and never EVER surfaced again (not even
>when the harddrive was being defragged at the same time). So I would suggest you
>either audit every recording OR go for WinNT (or Win2000 which also works like a
>charm (successor of Win NT 4)).
Or Linux (or FreeBSD, I imagine). I've been doing these transfers for
about 2 years, on a lowly Pentium 100 desktop with 16 MB of RAM.
Haven't had any problem at all with dropped data.
Make sure your hard-disk drivers have busmaster DMA enabled. This
greatly reduces the CPU workload, and allows the CPU to run your
recording task even while the drive is writing data.
>Clean your Vinyl THOROUGHLY. I personally use a carbon brush for that but there
>are various other items worth considering as well (just don't go overboard).
>Playing wet might be someting to consider IF you know the downsides of wet
>playing. I don't, so I don't.
Ick. From all I've heard, wet playing is 99% habit-forming - once
you've played an album wet, you'll always have to play it wet, because
the mineral-and-dust residue which builds up in the grooves will cause
it to be incredibly noisy when played dry. The only way back is to
give the album a very thorough wet-wash-and-vaccum treatment.
I've also heard that wet playing actually causes more surface wear and
damage than dry playing, for somewhat-esoteric reasons.
If you buy used vinyl, a simple cleaning with a dry brush probably
won't cut it. Most of the used albums I've bought, really did need a
thorough wet cleaning.
>OK, now you're ready to record. After finishing your recordings, listen to them
>(for starters because of the Win9x interrupt flaw) for pops and ticks. I
>personally use Sound Forge for this part since Sound Forge has a very nice PEN
>function. With the pen function you can literally draw a wav file (or apply
>corrections to them). If you hear a pop/tick, move the cursor to the offending
>part, zoom in (using the zoom function) and, using the pen function, redraw the
>offending tick part (quite often only in ONE of the channels) with a shape which
>you think MOST resembles what it should have been. You'll be AMAZED with what
>you can get away with using this function.
Diamond Cut's DC-ART has a nice "paste/interpolate" feature, which
performs this re-drawing automatically using some sort of
curve-fitting algorithm. It's very effective.
In really severe cases, where you've got a really horrendous pop or
tick or GNRCH or mistracking, it's sometimes possible to do a more
involved repair. You may actually be able to select and copy a few
cycles of the audio waveform from immediately before, or immediately
after the point of damage, "paste" this data over the damaged regions,
and then clean up whatever misalignment occurred at either end of the
paste. This trick can be used to correct for things as nasty as a
bubble in the vinyl (one which causes a really severe THWOMP followed
by some visible tonearm bounce, once per revolution of the LP, for
8-10 seconds of play). It takes a lot of care and effort to get
everything lined up correctly - I must have spent a couple of hours on
this one repair - but if the material you're working with is important
and impossible to replace it may be worth while.
>NEVER rely on automated features for pop/tick removal. They'll always degrade
>the ENTIRE recording (and often don't discriminate very well between transient
>signals and actual ticks/pops, thus resulting in a dull and tediously bad
>recording).
Agreed.
>Some adresses:
>
>www.feurio.com
>www.syntrillium.com (Cool Edit)
>www.goldwave.com
>www.steinberg.net (WaveLab)
>www.sonicfoundry.com (SoundForge)
www.diamondcut.com (DC Art)
--
Dave Platt dpl...@radagast.org
Visit the Jade Warrior home page: http://www.radagast.org/jade-warrior/
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!
Walter Knapp
> Manual glitch-finding-and-removal can produce very good results, but
> is extremely labor-intensive. I haven't really liked any of the
> automatic glitch- finders I've used - they either miss a fair number
> of pops and ticks, or mistakenly "fix" actual musical transients.
> Setting the filter / selection sensitivity controls requires a lot of
> trial and error.
On the mac side, the declick/decrackle filter in Spark XL is pretty
good. Only one I've used I trust for more than just hunting the clicks
for me to edit. One of the advantages it has is that it will preview the
result while you set the settings. And you can play either the cleaned
music, or what it takes out. That latter will actually give you a better
setting. As an aside, Spark XL will allow you to save what you take out,
so you can mix it with some nice clean music to make it sound like it
came from a LP...
> If you buy used vinyl, a simple cleaning with a dry brush probably
> won't cut it. Most of the used albums I've bought, really did need a
> thorough wet cleaning.
Depends on the source, I've gotten some that were really cleaned well.
Often though, if it's rare, you have to take what you find.
> >NEVER rely on automated features for pop/tick removal. They'll always degrade
> >the ENTIRE recording (and often don't discriminate very well between transient
> >signals and actual ticks/pops, thus resulting in a dull and tediously bad
> >recording).
As I noted, in the mac, Spark XL is the only one I'd trust to do it
automated. In a large measure because of the ability to fine tune it's
settings while previewing. Also certainly because they did a good job of
programming there.
Clive Backham
wrote:
>>What matters VERY much is the OS used...... one of the
>>biggest problems I encountered was missing frames. .... it's actually
>>due to the piss poor interrupt handling of Win9x. This goes for Win95 as well as
>>for Win98.
I'd just point out that I've been doing vinyl restoration for several
years now, using Win3.1, then Win95, now Win98SE. (I've not used NT
because my SPDIF I/O card (Zefiro ZA1) doesn't have NT drivers). The
dropouts you describe are usually due to the fact that the Win9x
default configuration is not very conducive to constant data
throughput, but it's perfectly easy to set it up so it works well. I
never have any dropouts on my system. My impression is that the
majority of people who record analogue to hard disk on a PC use Win9x,
and they do so sucessfully.
>I've also heard that wet playing actually causes more surface wear and
>damage than dry playing, for somewhat-esoteric reasons.
The theory as I understand it is this: When dry playing, the friction
of stylus heats up the vinyl, which allows the vinyl to deform
elastically (and thereafter return to its original position after a
few hours). When wet playing, the fluid acts as a coolant, preventing
the elastic deformation of the vinyl. Instead, the stylus shaves off
slivers of the groove wall.
>In really severe cases, where you've got a really horrendous pop or
>tick or GNRCH or mistracking, it's sometimes possible to do a more
>involved repair. You may actually be able to select and copy a few
>cycles of the audio waveform from immediately before, or immediately
>after the point of damage, "paste" this data over the damaged regions,
>and then clean up whatever misalignment occurred at either end of the
>paste. This trick can be used to correct for things as nasty as a
>bubble in the vinyl (one which causes a really severe THWOMP followed
>by some visible tonearm bounce, once per revolution of the LP, for
>8-10 seconds of play). It takes a lot of care and effort to get
>everything lined up correctly - I must have spent a couple of hours on
>this one repair - but if the material you're working with is important
>and impossible to replace it may be worth while.
At the risk of being accused to advertising, I'll just point out that
Wave Repair's "block overlay" feature makes this "paste in a
replacement from nearby" fairly easy.
Clive Backham
email: cl...@capita.nildram.co.uk
LP->CDR notes: http://homepages.nildram.co.uk/~abcomp/lp-cdr.htm
Dré
> I'd just point out that I've been doing vinyl restoration for several
> years now, using Win3.1, then Win95, now Win98SE. (I've not used NT
> because my SPDIF I/O card (Zefiro ZA1) doesn't have NT drivers). The
> dropouts you describe are usually due to the fact that the Win9x
> default configuration is not very conducive to constant data
> throughput, but it's perfectly easy to set it up so it works well.
If it's perfectly easy to set it up so it works well then it might be worth
considering to actually post the procedure, needed to set it up so it works
well! I consider myself by no means a WinTel illiterate YET I was unable (with
the mentioned soundcards) to set up Win9x so it actually did NOT behave like
mentioned earlier.
What I forgot to mention BTW was that these problems increased as Windows got
"stale". On a fresh install, the problem above never surfaced. After a few
months of use (no, not while having added numerous programs that run in the
background), the problem started to surface (and yes, (U)DMA drivers were used
for the harddisks).
One of the features I found in Win9x was that, as Win9x "grew older", the
processor load increased (while doing nothing BTW). Several programs that
explain which background process was/were responsible for this behavour could
not tell me what happened (since it was the Windows KERNEL that posed the "idle
load").
And it might also be very true that it's a combination of soundcard/OS which
gives problems. I have only tested a variety of Creative products.
Anyway, all these problems disappeared after installing WinNT (or Win2000).
So I'd be more than interested to know how the problems could be solved on a
Win9x configuration.
Arny Krueger
"Dré" <see_pos...@il.ok> wrote in message
news:91qcl3$96s$1...@news1.xs4all.nl...
>
> One of the features I found in Win9x was that, as Win9x "grew
older", the
> processor load increased (while doing nothing BTW). Several
programs that
> explain which background process was/were responsible for this
> not tell me what happened (since it was the Windows KERNEL that
posed the "idle
> load").
>
> And it might also be very true that it's a combination of
soundcard/OS which
> gives problems. I have only tested a variety of Creative products.
>
> Anyway, all these problems disappeared after installing WinNT (or
Win2000).
>
> So I'd be more than interested to know how the problems could be
solved on a
> Win9x configuration.
Ahh, "Install Fatigue". I know it well.
Here is what I know about it:
(1) Hard drives get slower the fuller they get.
(2) Adding software changes the registry and some of the changes
don't seem to get backed out even if you "uninstall" the software.
(3) Permanent swap files are a good idea because they "nail down"
prime space (i.e., near the beginning of the disk).
(4) MSConfig.exe in the win98 \windows\systems\ folder is your
friend because it allows you to minimize software that loads
automatically when win9x starts. It works with other flavors of
windows than win98.
(5) Defragging and keeping the disk as empty as possible is very
good.
(6) Keeping the \windows\temp folder empty is always a good idea.
(7) keep an eye on driver updates, particularly those for hard drive
and system devices.
(8) Detail-level power management is bad. Turn it off any way you can
and keep it off. Spinning down hard drives and the monitor after
about an hour or more of inactivity is OK unless you do audio stuff
that takes over half an hour per realtime operation.
(9) Regclean.exe from Microsoft can sometimes "bring back" systems
that have acute cases of install fatigue. Scanreg can work wonders on
win95 systems.
I suspect that many of these "tricks" are pretty well known, but this
is my list. Enjoy! ;-)
(5) Regclean.exe
Jamie Hanrahan
wrote:
>
>"Dré" <see_pos...@il.ok> wrote in message
>news:91qcl3$96s$1...@news1.xs4all.nl...
>>
>> One of the features I found in Win9x was that, as Win9x "grew
>older", the
>> processor load increased (while doing nothing BTW). Several
>programs that
>> explain which background process was/were responsible for this
>behavior could
>> not tell me what happened (since it was the Windows KERNEL that
>posed the "idle
>> load").
>>
>> And it might also be very true that it's a combination of
>soundcard/OS which
>> gives problems. I have only tested a variety of Creative products.
>>
>> Anyway, all these problems disappeared after installing WinNT (or
>Win2000).
>>
>> So I'd be more than interested to know how the problems could be
>solved on a
>> Win9x configuration.
>
>Ahh, "Install Fatigue". I know it well.
>
>Here is what I know about it:
>
>(1) Hard drives get slower the fuller they get.
Huh?
You're in my field now. Expressed the way it is here (simplistically),
your claim simply isn't true. Clarification, please?
--- Jamie Hanrahan, Kernel Mode Systems, San Diego CA
Windows NT/2000 driver consulting and training
http://www.kernel-mode.com/
I'd prefer replies, followups, questions, etc., in news, not via e-mail.
Clive Backham
wrote:
>If it's perfectly easy to set it up so it works well then it might be worth
>considering to actually post the procedure, needed to set it up so it works
>well! I consider myself by no means a WinTel illiterate YET I was unable (with
>the mentioned soundcards) to set up Win9x so it actually did NOT behave like
>mentioned earlier.
Here's the common stuff that I know about:
http://homepages.nildram.co.uk/~abcomp/pcaudiofaq.htm#dropouts
I think you've been unlucky, and let's face it when you perm all the
possible combinations of OS, motherboard, BIOS & soundcard drivers out
there, there are bound to be a few that can't be made to work. But
there are thousands of people who are happily working with audio on
Win9x.
Arny Krueger
"Jamie Hanrahan" <j...@cmkrnl.com> wrote in message
news:3a42d7f2...@nntp.cts.com...
> Huh?
Sure hard drives run in constant angular velocity mode. The outer
tracks are up to 3 times longer than the inner tracks. That means
that at constant linear density they can hold 3 times as many bits.
Drive manufacturers exploit this by "zoning" the drive, where each
zone has the same number of sectors per track. Internal logic
translates commands from the computer into commands that match the
zones. So from the outside the sectors per track remains the same
across the drive. But on the inside there are 5 or more zones, and
the outer zones have a higher DTR and transfer more data per
revolution.
http://www.voodooextreme.com/hw/Features/hdexplained/
"One problem with this is that there is usually more sectors on the
outer tracks than the inner tracks, because they are longer. Drives
which employ this technique are said to use Zoned Bit Recording, or
ZBR, and encompasses almost all new drives. The problem with this is
that throughput varies drastically depending on the track used to
record the data on. The reason it is called zoned bit recording has
to do with the way tracks are arranged. Since there are many tracks
on a drive with little space in between them, the efficiency loss is
marginal if instead of optimizing the number of sectors for each
track, which is difficult, to group together a certain number of
tracks and then, for each track in this grouping, or zone, use a set
number of sectors per track. The number of zones may vary from drive
to drive. The Maxtor DiamondMax Plus 40 has 16 zones."
http://www.super7.net/system/hdd/hddpart2.htm
"In older hard drives, the number of sectors per track was fixed.
The number of sectors that would fit on the innermost track
constrained the number of sectors per track for the entire platter,
because the outside tracks on a platter are longer than the inside
tracks, some of the space on the outside tracks was wasted. Many of
today's more advanced drives use a formatting technique called
Multiple Zone Recording to fit more data onto the surface of the
disk. Multiple Zone Recording enables the number of sectors per track
to be adjusted so that more sectors are stored on the larger, outside
tracks. By dividing the outer tracks into more sectors, data can be
recorded uniformly across the surface of a platter. This, more
efficient use of the disk surface means that higher capacities can be
achieved with fewer platters. The number of sectors per track on an
average 3.5-inch disk ranges from 60 to 120 using Multiple Zone
Recording. Not only is effective storage capacity increased by as
much as 25 percent with Multiple Zone Recording, but disk-to-buffer
transfer rates are also increased. With more bytes per track, data in
the outer zones is read at a faster rate."
Jamie Hanrahan
wrote:
>On Wed, 20 Dec 2000 16:03:32 GMT, j...@cmkrnl.com (Jamie Hanrahan) wrote:
>
>>On Wed, 20 Dec 2000 14:06:08 GMT, "Arny Krueger" <ar...@pop3free.com>
>>wrote:
>>>[...]
>>>Here is what I know about it:
>>>
>>>(1) Hard drives get slower the fuller they get.
>>
>>Huh?
>>
>>You're in my field now. Expressed the way it is here (simplistically),
>>your claim simply isn't true. Clarification, please?
>
>Sure hard drives run in constant angular velocity mode. The outer
>tracks are up to 3 times longer than the inner tracks. That means
>that at constant linear density they can hold 3 times as many bits.
>Drive manufacturers exploit this by "zoning" the drive, where each
>zone has the same number of sectors per track. Internal logic
>translates commands from the computer into commands that match the
>zones. So from the outside the sectors per track remains the same
>across the drive. But on the inside there are 5 or more zones, and
>the outer zones have a higher DTR and transfer more data per
>revolution.
Oh sure, that... however, there is no particular reason to believe that
"hard drive getting full" corresponds to "free space is all on the inner
tracks". Unless your HD is defragmented with all the free space ALSO
defragmented AND all "collected" toward the inner tracks, you'll never see
this effect in real life.
Note that even if the defragmenter attempts to make such an optimization,
it is partly a futile effort BECAUSE of ZBR. What the OS (and the
defragger!) see as monotonically increasing sector addresses do not
necessarily correspond to adjacent sectors. As a result the defragger
might end up doing exactly the wrong thing.
Anyway, the fact is that most HDs spend most of their time seeking, not
transferring data. While it sounds impressive to speak of transfer rates
of 100 MB/sec at the bus interface, for most applications you will see
very little difference in throughput vs., say, ATA33 drives. Especially
since, for example, CD-grade audio sampling only needs a lousy 176
KILObyte/second to be recorded.
Now even though most HDs spend most of their time seeking, and a lot of
that seek time is often attributed to fragmentation of files and of free
space, I don't generally believe in wholesale defragging of HDs... for a
variety of reasons having to do with access patterns of typical file
collections on typical disks. (This is not an issue I'm going to pursue
here, so please, everyone, do not regale me with glowing reports of vastly
increased performance since you defragged your HD.) But that aside,
depending on the defragger you use, the "dropouts during recording to HD"
problem might or might not be solved anyway; many defraggers (Diskkeeper,
for one) don't consider defragging free space to be important!
Defraggers can also be a "test to destruction" of your HD, putting more
cycles on it in a few hours than it might otherwise see in weeks.
What you can do that is MUCH less time consuming, and much less stressful
for your HD, is partition it. Set up a GB or two partition just for ".wav
capture", or whatever your high-speed writing application is, and keep it
empty... except for when you use it as the target of a recording task.
When the recording is done, copy the .wav files elsewhere and work with
them, but leave your "incoming" partition free. Note that since your C
partition that you usually boot and run from is the first partition on the
hard drive, this new one will be an inner partition. That won't matter;
performance will still be far faster than you need for this job.
Having this dedicated partition on a *second* HD, other than the one your
OS and apps and paging file are on, makes this an even better idea. If
you're on IDE, make this the "secondary master" of your IDE controller
rather than the "primary slave". You now have a second set of controller
channel, access arms, and media, all dedicated to your recording. If the
OS needs to page in the odd block of code here or there, it can do so
without moving the heads you're using for the recording. You can use the
rest of the HD for large bulk storage, the tmp directory, your working and
archived .wav file projects, whatever... anything you won't be touching
while recording.
--- jeh
Dré
> > Anyway, all these problems disappeared after installing WinNT (or
> Win2000).
> Ahh, "Install Fatigue". I know it well.
> Here is what I know about it:
> (1) Hard drives get slower the fuller they get.
Well.... Since the drive in question was a DEDICATED drive (dedicated to audio
only) and totally empty at the beginning of most recordings, this doesn't apply.
I hate using a partition which is also used for the OS. Using seperate
harddrives makes sure the continuous throughput is warranted.
> (2) Adding software changes the registry and some of the changes
> don't seem to get backed out even if you "uninstall" the software.
Like I said: NO installs of any kind (this includes install/removal). I am
"fairly" familiar with registry problems. That's why (in the "end" days of my
Win9x quests) I even made a back-up after a fresh install and (booting DOS)
replaced the "stale" registry with the backup every now and then. It certainly
helped in keeping Win9x a bit more "tacktile" (wrong word probably?) but it
didn't help in removing the strange "no processes running yet substantial
processor load" problem.
> (3) Permanent swap files are a good idea because they "nail down"
> prime space (i.e., near the beginning of the disk).
I have ALWAYS had a permanent swapfile (even when I used Win 3.1). I even went
as far as to give it a substantial minimum size so it didn't either increase or
reduce in size during recording. Swapfile was monitored during recording, posed
NO problems (constant in size, no swapping going on either).
> (4) MSConfig.exe in the win98 \windows\systems\ folder is your
> friend because it allows you to minimize software that loads
> automatically when win9x starts. It works with other flavors of
> windows than win98.
I (back then) used a few shareware tools for that. These tools FAAAR exceeded
the capabilities of the config util you mentioned. The ONLY background processes
were the kernel and the recording package. NO FF, no OSA, no silly mouse driver
icon utility, no display caller, no Getright, no Atomtime, no scan button
manager, no nothing. I always check the registry (even today) for those hidden
"start on start-up" thingies and remove them when they aren't very usefull.
> (5) Defragging and keeping the disk as empty as possible is very
> good.
I think that using a dedicated (and empty) harddrive completely covers the
slightest chance of this being a problem. The drive's sustained throughput was
also measured (10MB/s at that time) and that wasn't the bottleneck either.
Neither was the load, posed by the stream to harddrive (less than 1% increase in
proc. load). The drive was perfectly able to sustain a 10MB/s throughput for a
continuous 10 minutes (so delayed thermal calibration also wasn't the culprit).
Never tested it for longer periods BTW.
Besides: The 56MB cache I used should have sufficed in buffering the soundcard
for at least 300 seconds. Let's hope that no harddriver EVER needs that much
time to recalibrate itself! ;-) To go even further, the ENTIRE recording could
snuggly be placed inside said cache (thereby totally rendering any suspicions
about the drive being the culprit completely void IMO).
> (6) Keeping the \windows\temp folder empty is always a good idea.
I know. In the old days, a simple entry in the autoexec.bat sufficed. Nowadays I
just check it once a day or so.This wasn't the culprit either.
> (7) keep an eye on driver updates, particularly those for hard drive
> and system devices.
I always used the newest Microsoft (DMA) drivers. I have also tried (back then)
the Intel (DMA) drivers but they also did NOT solve the problem (Intel drivers
were optimized for use with benchmark tools (whereas I myself used the computer
for actual work most of the time)). Win98 came with it's own DMA drivers (as
does Win2000), on WinNT4 and Win95, I had to install them myself (and turn them
on in the registry). Didn't fix the problem on any of the 4 machines I tested it
on.
> (8) Detail-level power management is bad. Turn it off any way you can
> and keep it off. Spinning down hard drives and the monitor after
> about an hour or more of inactivity is OK unless you do audio stuff
> that takes over half an hour per realtime operation.
Tried that too (on Win98 of course, Win95 didn't offer very much in terms of
power management). Turned off everything (right down to the BIOS), didn't help
either.
> (9) Regclean.exe from Microsoft can sometimes "bring back" systems
> that have acute cases of install fatigue. Scanreg can work wonders on
> win95 systems.
Used Regclean quite often. Didn't help either.
> I suspect that many of these "tricks" are pretty well known, but this
> is my list. Enjoy! ;-)
I lost appetite for OS'es that can't even cope with a 172kB per second stream.
Still, it's strange! Win9x had absolutely NO problem with my Ultraplex40,
pumping some 7MB/sec to the harddrive at the end of a CD yet couldn't perfectly
cope with the 172kB/sec stream from the soundcard. Go figure!
Greetings,
tr...@my-deja.com
"Dré" <see_pos...@il.ok> wrote:
So, as mentioned, one of the first problems I encountered were
the missing
> frames (parts, chunks) in the middle of the music (pretty much at
random, once
> every 4 recordings or so (average 5 minutes per recording)). I at
that time
> decided to audit every piece of recorded audio which was quite
tedious. Someone
> (don't remember who) pointed me to the fact that Win9x suffers
from poor
> interrupt handling whereas WinNT doesn't so I switched to
WinNT (4). Problems
I am one lucky guy I guess! I have no shortages when it comes to
operating systems: 95, 98, NT, MAC OS 7.6, 8.5, 9.0, Linux,
SunOS. and more... Most through work of course.
I was hoping to do everything on my 98 at home, but I will
reconsider. I have a CD burner at work on an NT, so it may not be
that difficult to do.
Does the same problem exist on Linux or MacOS 9?
Also I am not clear as to when these missing frames occur:
When recording to hard disk from LP/Cassette, r wen recording
the wav files to CD from hardisk?
The reason I am asking about this, I can perhaps do the creation
of the wav files at home. Put them on zip disks --hope this is
practical-- and burn the CD at work. That way I do not carry and set
up my cassette deck, etc. to work...
Thanks,
Al
Walter Knapp
> Does the same problem exist on Linux or MacOS 9?
Not in Mac OS 9.0.4, and not in quite a few earlier ones. Never had this
problem through three generations of macs, several OS changes, several
pieces of software. Countless hours recording. I do record to aiff, but
that would make no difference, I could set .wav if I wished.
> The reason I am asking about this, I can perhaps do the creation
> of the wav files at home. Put them on zip disks --hope this is
> practical-- and burn the CD at work. That way I do not carry and set
> up my cassette deck, etc. to work...
You will need a few zip disks, at least a CD's worth.
Dré
> I am one lucky guy I guess! I have no shortages when it comes to
> operating systems: 95, 98, NT, MAC OS 7.6, 8.5, 9.0, Linux,
> SunOS. and more... Most through work of course.
Too bad Win2000 isn't on the list. It's actually a pretty descent OS for those
that use their computer for work (instead of the Linux guys who, generally, just
look at how dandy, stable and fast the thing is (just kidding here!!!)). WinNT4
is also very good BUT you'll miss out on some of the DirectX plug-ins (directX
support under NT4 is very limited). Win2000 has adequate DirectX support (unless
you're a DJ in which case you'll find out that the 0.1% pitch increments as used
in DJ software mixers DON'T work under Win2000!).
All in all I MUCH prefer Win2000 over Win98. Where in the old days I had roughly
1-2 blue's per day on Win98, this has dramatically reduced to 1 per few months
under Win2000 (24/7).
> I was hoping to do everything on my 98 at home, but I will
> reconsider. I have a CD burner at work on an NT, so it may not be
> that difficult to do.
Transporting a CD on other formats DOES take some room. You can expect to need
some 748MB for a full size CD, once it is transferred to WAV file(s).
> Does the same problem exist on Linux or MacOS 9?
Don't know for sure but I haven't heard of it.
> Also I am not clear as to when these missing frames occur:
> When recording to hard disk from LP/Cassette, r wen recording
> the wav files to CD from hardisk?
Wav to CD is a simple transferal of DATA. If an OS would mess that up, it
wouldn't work at all! It's the readout of the soundcards buffer (so during
analog recording) that is the culprit here. If Windows "forgets" to read the
buffer, it simply overflows (and Windows is too lazy to even check the overflow
flag!!!) and thus a part of your audio is lost (forever). Professional
soundcards (those $$$ things) seem to have tackled this by adressing the
soundcard differently (so they don't rely on the OS to read-out the buffer).
I have TRIED to get the same behavour with Win2000/NT. I have gone as far as
defragging the drive I was recording to while extracting an audio CD (at 30+
speed using an Ultraplex40) while doing the mentioned recording AND running a
rendering process at the same time. Proc. Load was 100% all the time, harddisk
was VERY busy yet NOT a single frame of audio was lost.
> The reason I am asking about this, I can perhaps do the creation
> of the wav files at home. Put them on zip disks --hope this is
> practical-- and burn the CD at work. That way I do not carry and set
> up my cassette deck, etc. to work...
You'll need 8 of those 100MB zip drives. Don't count on compression to work
(doesn't do all that well on audio).
I myself nowadays don't have enough room for a good TT setup. That's why the
analog recording is done somewhere else, the wav files are written to 2 CD-RW
discs (as DATA) and then copied to my computer. Works like a charm. I could also
write the CD-RW discs as audio discs but writing them as data provides an extra
layer of error correction (which, in the case of CD-RW discs, probably isn't a
bad idea).
<Snipped redundant text>
Julian
Fatman0 wrote in message ...
>
>I have also tried minidisc. It can't handle wide range music from an lp.
>So I stick with CD recordable.
>
Minidisc can handle the exact same dynamic range as CD. Possibly you had
a problem with a particular piece of equipment or a setup problem.
Arny Krueger
> >>Huh?
Other than the fact that many people disk defragmenters early and
often. This is especially true since all forms of Win9x and Win2K
come with disk defragmenters. I defrag every machine I build. I
generally defrag every machine with a reported performance problem.
>Unless your HD is defragmented with all the free space ALSO
> defragmented AND all "collected" toward the inner tracks, you'll
never see
> this effect in real life.
Isn't this like saying: "If you don't brush your teeth well enough to
get the placque off, you won't see the cavities?"
> Note that even if the defragmenter attempts to make such an
optimization,
> it is partly a futile effort BECAUSE of ZBR.
I don't know about that. The worst that the common kind of defragger
can do is put the free space in the slowest part of the drive.
>What the OS (and the
> defragger!) see as monotonically increasing sector addresses do not
> necessarily correspond to adjacent sectors.
First off there are two dimensions to "adjacent" - radially and
longitudinally. Secondly, "close" and near together always seems to
be better than "scattered all over hell's half acre".
> As a result the defragged might end up doing exactly the wrong
thing.
I think that if you attempt to provide a detailed scenario
illustrating this that is probable, you will have some difficulty.
> Anyway, the fact is that most HDs spend most of their time seeking,
not transferring data.
Actually, most hard drives spend most of their time doing nothing at
all or at most spinning idly. When they are active, they have three
(not two) possible states:
(1) Seeking
(2) Latency
(3) Transferring data
> While it sounds impressive to speak of transfer rates
> of 100 MB/sec at the bus interface, for most applications you will
see
> very little difference in throughput vs., say, ATA33 drives.
I agree with this but I see it as some mixture of how things are and
forever have to be, and a consequence of suboptimal operation and
procedures.
> Especially since, for example, CD-grade audio sampling only needs
a lousy 176 KILObyte/second to be recorded.
This hasn't been a problem on properly-engineered and operational PC
systems for at last 5 years.
> Now even though most HDs spend most of their time seeking, and a
lot of
> that seek time is often attributed to fragmentation of files and of
free
> space, I don't generally believe in wholesale defragging of HDs...
for a
> variety of reasons having to do with access patterns of typical
file
> collections on typical disks.
You'd have to argue long and hard (you've done neither) to convince
me that periodically defragging HD's is a bad idea.
> (This is not an issue I'm going to pursue
> here, so please, everyone, do not regale me with glowing reports of
vastly
> increased performance since you defragged your HD.)
That's called dismissing a critical issue with a wave of the hand.
>But that aside,
> depending on the defragger you use, the "dropouts during recording
to HD"
> problem might or might not be solved anyway; many defraggers
(Diskkeeper,
> for one) don't consider defragging free space to be important!
I'm most familiar with DiskKeeper in a NTFS environment. Defragging
NTFS is a far more complex issue than defragging FAT volumes. Since
most of the readers of the NG's this post appears on are running FAT
volumes what DiskKeeper does or does not do on NTFS volumes seems to
have limited relevance.
> Defraggers can also be a "test to destruction" of your HD, putting
more
> cycles on it in a few hours than it might otherwise see in weeks.
Agreed. OTOH, hard drives are supposed to be able to handle workloads
like this.
> What you can do that is MUCH less time consuming, and much less
stressful
> for your HD, is partition it.
Fine advice as far as it goes, but generally worthless and even
dysfunctional for the general PC user. The needs of the general PC
user are generally best met by having his disk space in a single
partition.
> Set up a GB or two partition just for ".wav
> capture", or whatever your high-speed writing application is, and
keep it
> empty... except for when you use it as the target of a recording
task.
Possibly fine advice for technical specialists. However I favor plan
B: have a disk subsystem that can just get the job done without
burdening the user with managing disk space.
> When the recording is done, copy the .wav files elsewhere and work
with
> them, but leave your "incoming" partition free. Note that since
your C
> partition that you usually boot and run from is the first partition
on the
> hard drive, this new one will be an inner partition. That won't
matter;
> performance will still be far faster than you need for this job.
IMO any PC that is so poorly configured that you have to optimize it
to this extent to get reliable DAE is unsuited for the purpose.
> Having this dedicated partition on a *second* HD, other than the
one your
> OS and apps and paging file are on, makes this an even better idea.
If
> you're on IDE, make this the "secondary master" of your IDE
controller
> rather than the "primary slave".
Often not a good idea or note even possible as the secondary IDE
channel is usually tied up with CDROM and other non-disk equipment.
>You now have a second set of controller
> channel, access arms, and media, all dedicated to your recording.
If the
> OS needs to page in the odd block of code here or there, it can do
so
> without moving the heads you're using for the recording. You can
use the
> rest of the HD for large bulk storage, the tmp directory, your
working and
> archived .wav file projects, whatever... anything you won't be
touching
> while recording.
My preferences in this area have shifted with technology.
(1) I think that having two hard drives generally yields better
performance than 2. That hasn't changed.
(2) The swap file is a high activity file and without a lot of
analysis simply moving it to the second volume is an easy thing to do
that generally yields better performance. W/r/t ripping, if you are
ripping and swapping at the same time there must be something wrong,
like a severe shortage of RAM or too much multitasking. That hasn't
changed
(3) With PIO drives, data transfer time was an issue, and separate
controllers seemed to provide an stronger advantage. Now that DMA 33,
DMA 66, and DMA 100 drives along with the controllers to exploit them
are more generally available, separate controllers seems to be far
less of an issue. I generally don't worry much about putting two IDE
drives on the same controller. That has changed.
To some degree partitioning can be like a religion for some people.
It can work, but it generally presumes a lot about the technical
interest and skills of the people working with the machine and
maintaining it.
IME people want computers to just work and not spend a lot of time
managing them.
I know for sure that a well-engineered computer that is defragged on
a reasonable schedule can rip reliably for months or years at a time
without further attention. And when I say "reliably" I mean "bit
perfect".
Jamie Hanrahan
wrote:
>> defragmented AND all "collected" toward the inner tracks, you'll
>never see
>> this effect in real life.
>
>Isn't this like saying: "If you don't brush your teeth well enough to
>get the placque off, you won't see the cavities?"
Uh....
>> Note that even if the defragmenter attempts to make such an
>optimization,
>> it is partly a futile effort BECAUSE of ZBR.
>
>I don't know about that. The worst that the common kind of defragger
>can do is put the free space in the slowest part of the drive.
Which is what they generally do -- if they bother to defrag the free space
at all.
>>What the OS (and the
>> defragger!) see as monotonically increasing sector addresses do not
>> necessarily correspond to adjacent sectors.
>
>First off there are two dimensions to "adjacent" - radially and
>longitudinally. Secondly, "close" and near together always seems to
>be better than "scattered all over hell's half acre".
>
>> As a result the defragged might end up doing exactly the wrong
>thing.
>
>I think that if you attempt to provide a detailed scenario
>illustrating this that is probable, you will have some difficulty.
I didn't say it was "probable", only that it "might" happen.
>> Anyway, the fact is that most HDs spend most of their time seeking,
>not transferring data.
>
>Actually, most hard drives spend most of their time doing nothing at
>all or at most spinning idly. When they are active, they have three
>(not two) possible states:
>
>(1) Seeking
>
>(2) Latency
>
>(3) Transferring data
Ok, if you want to get really precise about it... In a way we could say
that there is never a time when it is ONLY seeking, since the disk is
spinning all the time. However it of course isn't really looking for the
sector you want to show up until the head is over the right track.
From the driver writer's perspective, I regard "seeking" and "latency" as
part of the same thing -- time from the time I give the drive a command to
the time I get the data. At the controller interface, you tell the drive
to go get a particular range of sectors, starting with head/track/sector
number, and eventually it does. You don't tell the drive "seek to track
n; now wait for sector m to show up under head q; now read." Granted this
isn't exactly what is happening in the drive, but since the h/t/s
addresses we see these days are largely myths anyway...
I have done some database optimization work wherein, in order to get the
absolute best performance (for benchmarks), we did quite a bit of research
into where to put things on the disk. Part of the fun (given ZBR) is
figuring out just where things really are! Commonly-accessed stuff wants
to go in the MIDDLE of the drive, cutting the typical seek time from it to
anything else, or vice versa, in half.
Anyway, the point is still valid -- of the total "time to do an IO", most
of it is not transfer time; it's seeking + latency. IOW you could drop
the transfer time to zero (infinite transfer rate) and not see an
appreciable gain in performance on most systems, with most IO loads.
>> Especially since, for example, CD-grade audio sampling only needs
>a lousy 176 KILObyte/second to be recorded.
>
>This hasn't been a problem on properly-engineered and operational PC
>systems for at last 5 years.
You wouldn't think so, but people keep having problems...
>> Now even though most HDs spend most of their time seeking, and a
>lot of
>> that seek time is often attributed to fragmentation of files and of
>free
>> space, I don't generally believe in wholesale defragging of HDs...
>for a
>> variety of reasons having to do with access patterns of typical
>file
>> collections on typical disks.
>
>You'd have to argue long and hard (you've done neither) to convince
>me that periodically defragging HD's is a bad idea.
I'm not trying to convince you of anything, and I never said I thought it
was a "bad idea". I simply said that I didn't believe in it.
>> (This is not an issue I'm going to pursue
>> here, so please, everyone, do not regale me with glowing reports of
>vastly
>> increased performance since you defragged your HD.)
>
>That's called dismissing a critical issue with a wave of the hand.
No, it is simply saying "I don't want to hear anecdotal evidence".
>>But that aside,
>> depending on the defragger you use, the "dropouts during recording
>to HD"
>> problem might or might not be solved anyway; many defraggers
>(Diskkeeper,
>> for one) don't consider defragging free space to be important!
>
>I'm most familiar with DiskKeeper in a NTFS environment. Defragging
>NTFS is a far more complex issue than defragging FAT volumes. Since
>most of the readers of the NG's this post appears on are running FAT
>volumes what DiskKeeper does or does not do on NTFS volumes seems to
>have limited relevance.
It's true of DK on FAT partitions also, and in both Win9x/ME and NT/2K.
>> Defraggers can also be a "test to destruction" of your HD, putting
>more
>> cycles on it in a few hours than it might otherwise see in weeks.
>
>Agreed. OTOH, hard drives are supposed to be able to handle workloads
>like this.
The workload in terms of operations per second is not what I'm concerned
about. I'm concerned about simply "using up" the available lifetime of
the drive. In effect, every HD has a limit on the number of operations it
can perform before it dies. Not a precise number, of course, but a limit
nonetheless. And defraggers "use up" the HD at a rate far greater than
most normal use of a PC.
>> What you can do that is MUCH less time consuming, and much less
>stressful
>> for your HD, is partition it.
>
>Fine advice as far as it goes, but generally worthless and even
>dysfunctional for the general PC user. The needs of the general PC
>user are generally best met by having his disk space in a single
>partition.
Your opinion only, sir. I know many, many "general PC users" who find it
extremely convenient and natural to partition their HD space and to use
multiple HDs.
>> Set up a GB or two partition just for ".wav
>> capture", or whatever your high-speed writing application is, and
>keep it
>> empty... except for when you use it as the target of a recording
>task.
>
>Possibly fine advice for technical specialists. However I favor plan
>B: have a disk subsystem that can just get the job done without
>burdening the user with managing disk space.
So instead you have to burden the user (and the HD) with defragging it
every so often to maintain that performance level.
>> When the recording is done, copy the .wav files elsewhere and work
>with
>> them, but leave your "incoming" partition free. Note that since
>your C
>> partition that you usually boot and run from is the first partition
>on the
>> hard drive, this new one will be an inner partition. That won't
>matter;
>> performance will still be far faster than you need for this job.
>
>IMO any PC that is so poorly configured that you have to optimize it
>to this extent to get reliable DAE is unsuited for the purpose.
Well, let's see. I'm looking at a relatively new, middle-line HD bought a
few months ago at Fry's, a Maxtor 93073U6. 30GB, UDMA66, 5400 rpm, < 9
msec average access time. This is by no means top-end, but it is typical
-- maybe even a little better -- than the drives typically found in
pre-built PCs. A more technically aware person will probably have a 7200
rpm drive; a more technically aware and rich one, 10,000 rpm. :-)
5400 rpm gives an average latency (half a rev) of 5.5 msec, so let's say
14 msec average "seek + latency" time (which I'm just going to call
"seek", sorry) per transfer. We can do about 71 of those per second; if
we're really unlucky and all seeks are full-length and all latencies are
full-rotation, about 35 per second.
Now FAT32 will set up a 30 GB partition with a 16K cluster size. At 16
Kbytes per write, then to keep up with 176 KBytes/sec, we have to do only
about 11 writes per second to the disk. Sounds like we ought to be able
to keep up with that even with worst-case seeks for EVERY transfer, and if
the free space is contiguous and the program is smart enough to write
larger chunks, we can cut the load on the disk even more.
However, it is not that simple. Writing more and more chunks to the file
does not JUST involve writing to that file. We also have to update the
FAT every time the file is extended, and although the FAT is kept in
memory for performance, it does get flushed to disk every so often.
Still, we should be nowhere near the limit of the disk's performance or
the system's ability to drive it.
Yes, I admit -- my preference for spreading the IO load around comes from
environments with far heavier IO loads. EVERYONE in such environments
knows that it's a good idea to distribute heavy IO loads among multiple
HDs and controllers.
>> Having this dedicated partition on a *second* HD, other than the
>one your
>> OS and apps and paging file are on, makes this an even better idea.
>If
>> you're on IDE, make this the "secondary master" of your IDE
>controller
>> rather than the "primary slave".
>
>Often not a good idea or note even possible as the secondary IDE
>channel is usually tied up with CDROM and other non-disk equipment.
Oh come now -- you can make the CDROM the slave on either channel.
>[...]
>My preferences in this area have shifted with technology.
>
>(1) I think that having two hard drives generally yields better
>performance than 2. That hasn't changed.
I assume you meant "than 1".
>(2) The swap file is a high activity file and without a lot of
>analysis simply moving it to the second volume is an easy thing to do
>that generally yields better performance. W/r/t ripping, if you are
>ripping and swapping at the same time there must be something wrong,
>like a severe shortage of RAM or too much multitasking. That hasn't
>changed
True. IMO simply adding enough RAM to avoid writing to the swapfile
(paging file, really) is more cost-effective these days than adding
another HD -- if that's the only reason you're adding the other HD.
>(3) With PIO drives, data transfer time was an issue, and separate
>controllers seemed to provide an stronger advantage. Now that DMA 33,
>DMA 66, and DMA 100 drives along with the controllers to exploit them
>are more generally available, separate controllers seems to be far
>less of an issue. I generally don't worry much about putting two IDE
>drives on the same controller. That has changed.
I think you should reconsider. With two IDE drives on one controller,
only one can be doing ANYTHING (other than idling) at a time. Put them on
separate controllers (primary vs. secondary) and you get overlapped
operations. For servers with many drives that are busy all the time, this
is a big reason that SCSI is still preferable.
>To some degree partitioning can be like a religion for some people.
>It can work, but it generally presumes a lot about the technical
>interest and skills of the people working with the machine and
>maintaining it.
Agreed. We probably differ in our perception of just how much skill or
knowledge is required to take advantage of partitioning.
>IME people want computers to just work and not spend a lot of time
>managing them.
A nice goal, but we're still a ways from that. :-)
>I know for sure that a well-engineered computer that is defragged on
>a reasonable schedule can rip reliably for months or years at a time
>without further attention. And when I say "reliably" I mean "bit
>perfect".
I don't defrag and I get the same good results. Even on systems where I
don't bother with separate partitions, etc. Are you certain that
defragging is an essential part of your recipe?
--- jeh
tr...@my-deja.com
message about transfering music from LPs and cassette.
Just as I was getting ready to set up my computer and system, I came
across a TEAC standalone CD copier. It is built as an audio component.
Has two caddies for CD's. One I assume is just a player, the other is a
burner.
I did not make a note of model etc., but searching via mySimon I came
across a "TEAC RW-D250 Dual-Tray CD-RW Recorder/Player with Remote".
Retails for $299.
It claimed on the label that one can also use it to record from
cassette and records.
I see some advantages:
* I am thinking that if I were to get a CD burner for home, it would
simply be for music (I have one at work for data, etc.). And this
looks like it is made for that specific purpose.
* I will not have to overburden my computer, which is far more
expensive.
* It could be put with my stereo system, which is downstairs.
Disatavantages:
* Takes considerably more room than an internal CD burner, and I am
running out of room in my stereo cabinet.
* Probably won't be able to edit the music to get rid of pops, etc.
(this may not be actually be bad: I'll probably get the job done
faster. I assume I can always do the editing from the CD's later.)
Not having read any of the specs on this device, I do not know how the
quality would compare to that of what has been described by you experts
here.
Any comments? Anybody has used this or a similar device and compared
the quality of the CD to one that was produced via a computer? (Assume
no editing...)
Thanks!
Al
In article <91k5g5$tub$1...@nnrp1.deja.com>,
CLSNOWYOWL
burning CD's at first...
First of all, it did not allow for me to have the proper control when
converting analog sources to digital signal. The modulation (and the control of
it) was too far off for my tastes.
Also, there are different formats of disks on the market. These stand alone
devices can only accept one or two different disk formats or be compatible to
them. I recall there being at least 5 format types of CD's you can use to burn
data or audio onto. There are not many stand alone devices that can handle at
least 3 or more of the formats.
In addition, stand alone writers are very slow (they take about 40 minutes to
an hour to dub one CD to another). The computer controlled writers can
duplicate a CD in 20 minutes and do it more accurately.
Since these stand alone devices are very technically limited (they will not
allow you to change sampling rate, etc..., control the input or output signal
well, or process the audio accurately when itt's absolutely critical), I
recommend actually doing your burns from an internal CD writer on a computer.
Stand alone devices are really designed for making a copy of a known good CD or
series of CD's, not really meant for analog to digital conversion,
reproduction, etc...
You can actually have a wide range of CD-R and CD-RW formats, write speeds,
sampling rates, signal and modulation controls (particularly for analog to
digital conversion of the signal) by actually having a CD writer on hand with
the appropriate computer system and the software.
Therefore, for all of the above reasons, i found that computer writers are more
accurate, give you better control, and save you time... The other bonus, is
that with some software, like adaptech Easy CD creator, you can actually make
and print your own covers for your own CD jewelcases and even labels for your
CD's.
Jamie Hanrahan
>Thanks for all those who participated in the discussion on my previous
>message about transfering music from LPs and cassette.
>
>Just as I was getting ready to set up my computer and system, I came
>across a TEAC standalone CD copier. It is built as an audio component.
>Has two caddies for CD's. One I assume is just a player, the other is a
>burner.
>
>I did not make a note of model etc., but searching via mySimon I came
>across a "TEAC RW-D250 Dual-Tray CD-RW Recorder/Player with Remote".
>Retails for $299.
>
>It claimed on the label that one can also use it to record from
>cassette and records.
>
>I see some advantages:
>
>* I am thinking that if I were to get a CD burner for home, it would
>simply be for music (I have one at work for data, etc.). And this
>looks like it is made for that specific purpose.
>
>* I will not have to overburden my computer, which is far more
>expensive.
THere isn't a lot of "burden" placed on the computer when you add a CD-R
drive. This is a non-issue.
>
>* It could be put with my stereo system, which is downstairs.
If you're planning on copying from LP or cassette to CD, and you have no
way to play LP or cassette that's near your computer, this is a definite
issue.
>Disatavantages:
>
>* Takes considerably more room than an internal CD burner, and I am
>running out of room in my stereo cabinet.
>
>* Probably won't be able to edit the music to get rid of pops, etc.
Definitely not. The best you'll be able to do is clean the LPs
thoroughly.
In the heydey of LPs several mfrs offered single-ended noise reduction
boxes that could go in your tape loop. Essentially these are fast-acting
automatic high-pass filters and noise gates. You can usually find these
available used. Some prefer the results with one of these devices in the
path; some do not.
Control over the placement of track markers is also MUCH more crude with
the standalone unit.
>(this may not be actually be bad: I'll probably get the job done
>faster. I assume I can always do the editing from the CD's later.)
True, you could put one of your CD-Rs in your computer CD-ROM drive, rip
the music tracks to .wav files, edit those... then dump them back to CD-R.
The latter step requires you to have a CD-R burner on your computer, or
else move the audio CD-R drive next to your computer and feed it from the
sound card analog out. Without a really good sound card, that last step
will add some noise.
>Not having read any of the specs on this device, I do not know how the
>quality would compare to that of what has been described by you experts
>here.
>
>Any comments? Anybody has used this or a similar device and compared
>the quality of the CD to one that was produced via a computer? (Assume
>no editing...)
Quality should be about the same.
You missed two big disadvantages:
1. Buying the standalone unit does not get you a CD-R drive on your
computer. I personally find them indispensable.
2. The standalone unit requires the use of special CD-R blanks that cost
more than regular CD-R blanks, are harder to find, and aren't available in
very many different brands or models. The extra cost includes a copyright
fee, which supposedly makes up for the money you are supposedly stealing
from the artist, publisher, producer, distributor, etc., when you make a
copy of a recording you already own. Having less selection in the blanks
is a problem because compatibility between CD-Rs and music CD players can
be a problem -- in my experieince it's important to be able to try a
variety of blanks to find the ones that work best with your burner and
your players. If the "music CD-R" blanks your store happens to carry
aren't liked by one of your CD players, you're kinda outta luck.
--- jeh
Jamie Hanrahan
Attributions clarified:
On Sun, 24 Dec 2000 13:04:02 -0800, Jamie Hanrahan <j...@cmkrnl.com> wrote:
>On Sat, 23 Dec 2000 15:18:54 GMT, "Arny Krueger" <ar...@pop3free.com>
>wrote:
>>"Jamie Hanrahan" <j...@cmkrnl.com> wrote in message
>>news:3a492ead...@nntp.cts.com...
jeh >>> Especially since, for example, CD-grade audio sampling only
jeh >> needs a lousy 176 KILObyte/second to be recorded.
arny >>This hasn't been a problem on properly-engineered and
arny >> operational PC systems for at last 5 years.
jeh >You wouldn't think so, but people keep having problems...
arny >>IME people want computers to just work and not spend a lot of time
arny >>managing them.
Very true. The problem with this is, briefly, market pressure -- from the
marketeers, not the market. Vendors keep wanting to sell more software
and more hardware. As the capabilities of the PC increase, software
makers continue to use up those capabilities and keep the machine's
"unused reserve capacity" at just about the same point, year after year.
(It is a truism that all HDs, regardless of size, will be about 80% full.)
They do this not only by selling us the same old apps in bigger and more
resource-hungry versions, but also new types of applications we've never
been able to consider using before.
We also get (whether we want to or not) operating system changes no one I
know of ever asked for... but which require faster and bigger PCs to keep
up with.
So, sure, a typical "modern" PC, even with a heavily fragmented HD, can
easily record Red Book audio. Or provide Red Book audio to a CD burner.
But the big thing this year is personal *video* editing. By the time most
folks have upgraded their machines to the point where they can do *that*
without "managing" them much any more, there will be ANOTHER gosh-wow
gee-whiz app that everyone will want to do and which will strain the
resources of our 2 GHz processors, multi-GB RAM, and hundreds-of-GB HDs.
And so you and I will *continue* to get calls from our non-techie friends,
asking, even if not in these exact words, "why doesn't this 'just work'?"
:-)
--- jeh
Steve Raymar
<tr...@my-deja.com> wrote in message news:92c3um$nsr$1...@nnrp1.deja.com...
> Thanks for all those who participated in the discussion on my previous
> message about transfering music from LPs and cassette.
>
> Just as I was getting ready to set up my computer and system, I came
> across a TEAC standalone CD copier. It is built as an audio component.
> Has two caddies for CD's. One I assume is just a player, the other is a
> burner.
>
> I did not make a note of model etc., but searching via mySimon I came
> across a "TEAC RW-D250 Dual-Tray CD-RW Recorder/Player with Remote".
> Retails for $299.
>
Why not have both? I popped about $500 for a Philips internal CD-R
burner five years ago and it gave out after about three years. I was using
it
as a CD-ROM player also and I think that rushed it's demise so I got smart
and bought a seperate CD-ROM player to do the grunt work and a cheaper
but better internal CD burner just for making CDs and CDRs, both for less
than the Philips.
In a couple of weeks I'm going to get an external CD-RW recorder/player
as well. The reason being is that I'm an electronic musician and I can't run
a sequencer, soft synth and record to my hard drive at the same time, so
I'll record the output to an external CD burner. Then I can put that CD into
my CD-ROM player to send the music to my hard drive for any editing I
want to do. My internal CD burner will be mainly for data back up for the
hard drive. We live in amazing and spoiled times! I remember when 8MB
of RAM was over $300 and five years ago I paid $300 for my 2.1G hard
drive(I remember people thinking I was a memory glutten for that!).