You would most likely encounter a severe hi freq. roll off doing this for two reasons I can think of. One, running the cassette at high speed would take 15K and throw it up to 30K if the cassete heads could handle it, and the same thing on the digital side. anything above 10K and above being filtered by the Nyquist filtering at 20K. Might be ok for spoken word but any kind of music with hi freq content would undergo a huge high end loss. ... Kurt

Shouldn't the effect that you're describing happen to the cassette that the audio is being copied to also? Exactly how are the manufacturers able to get around this problem when hi-speed dubbing from one cassette to another?

MANTIK,
To some degree this does occur. A lot of info over 10k gets lost in high speed dubbing of cassettes. But most cassette duplication is done on "Bin Loop Mastering' recorders where the tape is taken off a pancake, recorded and then loaded into the cassette. This eliminates a lot of the problems that are inherent in cassettes such as pressure pads and tape modulation noise. Also the machines have heads designed specifically to transfer at high speed. But anyone who has experience will tell you that a cassette recorded at normal speed, on a quality machine with good tape will have better quality than any mass produced high speed dub.

If you want to try what you have in mind for yourself, just take a normal high speed dubbing deck and make a high speed dub. While you are making the dub run the audio into your workstation. Let me know what the results are. Kurt

Originally posted by MANTIK:
I'm trying to increase my productivity. People give me 120 minute cassettes to transfer to CD and I spend two hours twiddling my thumbs waiting for the audio to get recorded on my PC before I can do any work on it.

:) Hey MANTIK; Is the system tied up while doing this Xfer? The high frequency loss is one thing, there also is a speed variation problem that might come into play. Is there another way to digitize the cassette while your system is available for other work? Like with the purchase of a stand alone burner, or software/hardware addition to your existing system.

It seems a modification or addition, would be money better spent than a custom high speed cassette player. Or you could get hold of a small porta cassette multi-tracker that has 3 & 3/4 ips speed already. I think 2 x is the max you could go. It would probably loose all high end above 10K after going to digital, but if most is vox, it won't be too bad.

Just a suggestion,
--Rick

I’m not sure about other DAW systems, but SADiE has a “background recording” function for just the reason you describe. You can work on one project while another loads in…

Mantik,
Sonic Solutions also has a load in feature. You cannot dump & load at the same time, but you can load & work on another project at the same time. Maybe this is better than the HF loss situation.
Sometimes used Sonic systems can be purchased at good prices.

Thanks for the replies guys. I wrote this out of frustration. I just had a job where a TV syndicator gave me 60 120 minute cassette tapes to transfer to CD. These tapes were a sort of audio diary of his attemps during the Carter Administration to broker a deal between Anwar Sadat and MR. Begin to create a made for TV movie about their peace making process attempts. I had to sit through 120 hours of listening to just conversation of interviews. That means I lost 120 hours of work time.

I didn't originally think of the solution that while I'm working on one job the other could be loading in the background. That would mean that I would just have to figure what to do with myself for just two hours.

I just built a R.A.I.D. array for my PC. I wonder if I can use a variation of that to create the ability to load one project while I'm working on another?

Originally posted by MANTIK:

I just built a R.A.I.D. array for my PC. I wonder if I can use a variation of that to create the ability to load one project while I'm working on another?

:) If you can network in a smaller computer with an MPEG encoder, it might do the trick. It can be a small "no frills" computer.

The encoder computer should be able to create mp3 files of your long docs, using the array for storage. I say mpeg because of the nature of your sound source, cassette.

The encoder software should have it's own light editing capability. You should not be limited in file size, depending on your storage capacity, you may be able to load several docs, organize them, and access them without effecting other tasks from your main rig.

Just something to think about,

--Rick

This is where a Masterlink would come in handy.
Just plug the cassette deck into the ML and let it rip. There's enough space on a large drive to dump alot of time. I'm not sure what the max size HD you can put into a ML but you can change them yourself.
What's the largest drive anybody has ever installed in a Masterlink?

The RAID array, depending on which RAID mode you were using, would allow you either faster data transfer by splitting the data across the spindles of multiple disks, or data backup by recording the same thing to more than one disk. You will still be limited by the speed of the cassette.

Two tracks (assuming stereo) really wouldn't take up much recording time or much disk time. You should be able to continue to do other work while recording those tracks depending on what the secondary task was. The CPU and O/S speed as well as audio hardware capabilities would be more of an issue. If using multiple disks, particularly in a true SCSI array, you would be able to record a multi-track project at the same time without any troubles. Of course you may run into CPU time issues if your RAID array isn't built separately as a true file server.

Wait, something that Kurt said at the beginning that is wrong. If you double the play in speed to your DAW, and then slow it down with the DAW, you will not have a HF loss. It will be just the opposite. You will have a HF boost. Increasing IPS always boost top end, and casues the low end to drop off. Decreasing IPS increases low-end and causes the highs to drop off.

Another thing. If you double time it going in, you will add a mechanical noise, and it will be very noticable. You can't just pitch shift, as that will not compensate for the speed, so you will have to time expand as well.

When you time expand in your DAW. Your recording will not have the detail that it would have had in real time, as it is play in double time. Remember that the computer is recording most HF material at this point.

Originally posted by lowdbrent:
Wait, something that Kurt said at the beginning that is wrong. If you double the play in speed to your DAW, and then slow it down with the DAW, you will not have a HF loss. It will be just the opposite. You will have a HF boost

:) Hey Brent! The original idea for the poster was to expedite the transfer of long program material into his DAW rig.

His idea was to speed up the play of the external device by x amount of speed, we will say 2x for this example. When the DAW digitizes, unless extreme higher resolutions were used, and space availablity, these are large programs, the nyquist cutoff filters would limit the high frequency capability of the original doubled signal, once cut off, gone. The low frequency simply doubles as well, and this can be restored with no ill effect, but to go to CD, the highs would not be present do to the effects of the cut-off filter, and will reduce a portion of the extended response. The other part I can't comment on. Perhaps if you could explain why there would be more HF in another way.

--Rick

Brent,
The original post said nothing about running the DAW at double speed. In that scenario, what I said was correct. Kurt

I didn't say anything about the DAW recording speed. I realize that he has to record at normal speed, and playback at half speed.

My concern was that playing a cassette in at 2x, and then slowing the DAW down to half that speed to play back, will not sound the same as recording in real time.

Because the cassette deck is playing at twice the speed, you will not have all of the spectrum transfered, as it is beyond the f-response of the heads/circuitry, and there will be HF boosts emphasized in a way, that when you slow it down in the DAW, it will not sound natural. That boost will be recorded and will not go away.

:) Hey Brent, heavy stuff here. Yes, there will be non linearity's of all types, some speed variations as well. Point of the whole thing is to not dedicate the use of his entire system to upload these long files, and not allow any other work to be done in the meantime.

Me, I would edit real time and get it over with. Let it roll, thumb the pause control, mark the file, continue etc. Go back, tie it all together, organize and burn it. Frankly, it would be an advantage to the client to get a burner, perhaps DVD audio, or even a small mini dv cam, using audio tracks. Firewire the stuff right in.

--Rick

Rick, is that you in the avatar? You look like Johnny Carson.

Originally posted by lowdbrent:
Rick, is that you in the avatar? You look like Johnny Carson.

LOL, it's me, but I don't know if that's good or bad... hee hee. I told Chris I was backstage after receiving my 3rd Grammy....not!

--Rick

Thanks for the extra replies guys. True to my nature , I have a problem when I keep hearing the phrase "It can't be done!". I believe that was the response that was given to someone who had the absolute nerve to ask this question: " Hey, do you think we could ever take a trip to the moon?" With that in mind, I sent a brief recap of all your comments to this guy (Graham Slee - Audio Electronics Engineer)in England.

He has a company called Bespoke Audio Design. Their claim: "We have the experience to be able to design and develop virtually any audio device, product or system you can think of, to your chosen quality standard." I don't know who does this kind of stuff in the USA. I'm sure we have companies that do it but I don't know who they are.

I ran across this guy when I was looking for components to connect to my PC to help create the superb sound that I get when I do my LP to CD transfers. I got from this guy a phono preamp called the Era Gold $800 in the USA at http://www.lpgear.com.
In his following reply about this situation he mentions something about sending in a recording into my PC at 80rpm. Let me explain. I did an experiment where I took a LP and played it at 80 rpm while taking the audio and sending it into my PC. I then resampled it to normal sounding play speed. It seemed to me to sound OK but Graham disagrees (why did I do this experiment anyway? For the same reason. I hate waiting X amount of hours before I can actually get to do any work on the audio material, anyway here's his response:

Hi,

I'm afraid these guys are right. The problem is the "brick wall" filter necessary to prevent the sampling frequency aliasing the signal when you go into the digital domain. You can't get rid of this filter, or, put differently, if you did, you'd get mega distortion and transmit RF everywhere. Now if you could sample (and hence filter) at a much higher rate that would be better (but probably not easily done).

Let me explain. Lets take the highest frequency of speech, which is commonly considered 3kHz. But being ex-broadcast, let's use the AM radio standard of 4kHz. If you speed things up by 5, that frequency will become 20kHz, just about where the brick wall filter starts to cut. So up to 4kHz you're safe, above that all will be cut.

Conclusion. You can go at 5x speed without loosing quality on speech. Or can you?

One thing we forgot here is the tape EQ which has a time constant of 50 micro seconds (it varies if you're using reel to reel). In other words you have to apply 6dB/octave boost on replay of tape below 3kHz.

So on replay of the 5x dub you would experience bass loss. Solution? You would be best removing the EQ (if you could), dubbing and digitising at 5x speed, slowing back down in the digital domain, and then applying the 50 micro second EQ.

However, none of this is any good for music. When you did your 80 rpm transfer you lost everything above 8.4kHz, but as most musical notes are lower than this, it didn't sound bad. But you lost the harmonics which create the feel of the performance. OK, on CD you never really hear these subtleties all that well, so such a bandwidth may be acceptable.

But, did you consider the record replay EQ (RIAA). OK, I know it's not as noticeable because the curve is pretty constant compared to tape, but there is a break (knee) right in the middle at 1kHz, and in your 80 rpm recording that knee would have come in at 400Hz on your results. Therefore the tonal balance was actually changed. So for high quality record to CD transfers you need to do it in real time (and your deck speed needs to be bang-on).

Now, I'm sorry I'm not that helpful here, but high speed dubbing has always been a problem in the industry, even when it was analog to analog. And it's all down to high frequency performance. Good music cassettes could be dubbed at 4x speed because the limit of tape heads was about 40kHz and the cassette bandwidth was about 10kHz, plus the filters were gentle so there were still some harmonics that got through. In digital it's different because that brick wall filter has to get rid of everything in one octave (blame Nyquist). However, the digital brigade can boast superiority in that, once in the digital domain, they can manipulate virtually anything and in no time at all.

Just out of interest: Would you think that dubbing 50 minutes of high quality sound, including harmonics in excess of 50kHz, in just 1 second is fantastic? It is, and that's exactly what a record pressing plant does. Long live the king!

My best regards and thanks for enjoying the Era Gold,
Graham

PS. I'm always willing to help - discuss - suggest. Never hesitate to write/reply

GSP Audio
http://www.gspaudio.co.uk
0(044)1226 244908

MANTIK,

Don't over-analyze this - there's an easy solution:

If someone gave me sixty 2-hour cassette tapes to copy to CD, I'd install SoundForge on my old Pentium 120 and record to that so I don't tie up my real computer. If all that's needed is a straight dump to CD with only light editing, I'd buy a $40 CD burner for that computer and do the burning there too.

Surely you're charging enough to cover the cost of a cheap used computer and CD burner! :D

--Ethan

Check some garage sales. I bought 3 machines for $10 each. If you have any experience with O/S es you can install Linux for free and tune it to use a small footprint for your requirements. If not, use Win95. I ran Win95 for years on a P100 with 32MB mem and everyone thought it was a much faster machine. Win98 is a resource hog for an old machine.

Your only question is what to do about storage. If your RAID array is actually part of a file server then that is taken care of.