Conversion of tapes to digital - Best Methods

[SrinathK]

Dear all,

I would like to take up a topic on the various ways we've been converting our old cassette tapes into digital. I have also been doing the same with my old cassette collection (and it's still a long way from being done) in my own way, but I'd like to know all the ways by which we rasikas go to preserve our precious collections. We can even discuss if our methods need improving in any way now that storage capacities are so much higher today.

In another discussion we were talking about the best of 3 formats (.mp3, .aac (MP4) and .ogg (Vorbis)) and one of the reasons for that was that now I'm looking for a quality upgrade in my tape conversions.

[Rsachi]

Srinath,
Let me give a baseline. Cassette player with audio 3.5 mm output. Computer with Audacity with the correct input level and record settings.
Remove noise etc. and adjust level. Save in MP3.
44 kHz 128/160 kbps works well for me. Most Parvathi tracks were done like that.

[rajeshnat]

WE discussed this few years back in the below thread
http://www.rasikas.org/forums/viewtopic.php?f=2&t=5513

Mods
Merge this topic to that.

[cacm]

Dear RS,
1) .WAV IS BEST BUT SAMPLING AT 192 IS NOT THAT BAD IN MP3.All the three formats are o.k. if used OPTIMALLY. For example Parvathi tracks will sound BETTER at 192.
2) If you record at DVD AUDIO (sampling at 48khxz as opposed to 44.1 khz)you will be better off.
3) REG. using Audacity etc for NOISE REDUCTION its o.k. for removing clicks, hiss, major banging on mridangam etc but like Second law of thermodynamics S/N CANNOT BE IMPROVEMENT & EVEN DOLBY N.R. SUFFERS this lae's prowess.
4) Now a days MOST amplifiers are o.k. but the WEAKNESS IN CONVERSION IS THE SOUND CARD. In many P.C.S ETC the sound card can actually make the sound worse than the original. VVS & I used professional sound cards& recorded in .wav only & only later on tried MP3 conversion etc. If QUALITY& FIDELITY are SUPREME this is the way to go.
5) THE REAL WEAKEST LINK IS THE LOUD SPEAKER. If you can get a good one things will SOUND DRAMATICALLY BETTER. IF you cannot get one a good HEAD PHONE CAN WORK TOO......AN EX-AUDIOPHILE, VKV

[Nick H]

Save in FLAC.

Lossy formats (MP3, ogg, etc), however good, are intended for convenience of listening, for instance cramming more onto available storage. If one is talking about archiving then the idea of using a format that throws away a part of the music is really contradictory.

Whatever the raving of audiophiles, high-bit-rate MP3 is probably indistinguishable from lossless for listening, but for editing, enhancing, etc etc, one must have lossless.

Please consider that we do not know what techniques may be developed in the future for getting more out of old recordings.

....also an ex-audiophile
(well, a recovering audiophile!)

[Rsachi]

Sure, agree about higher rates, hz, Wav etc. -
Ex X (unknown variable)

[harimau]

4) Now a days MOST amplifiers are o.k. but the WEAKNESS IN CONVERSION IS THE SOUND CARD. In many P.C.S ETC the sound card can actually make the sound worse than the original. VVS & I used professional sound cards& recorded in .wav only & only later on tried MP3 conversion etc. If QUALITY& FIDELITY are SUPREME this is the way to go.

AN EX-AUDIOPHILE, VKV

Most of the digital recorders you get in the market today have very good A-D converters. If you buy a $500 recorder, I would venture to say that it would be better than a $500 sound card for the PC as the recorder doesn't need the circuits necessary to talk to the PC's bus; the digital signal gets written to a SD card. I guess one could get better components in the digital recorder.

I have heard it said that in analog recording, one should record as hot as possible and in digital recording one should record as cold as possible. That is, analog recordings should be at maximum signal strength without going into saturation and digital recordings should be at weak signal strengths. I have had no difficulty in amplifying recordings at -48 dB sufficiently so that the signal strength after amplification is close to -3 dB. Of course with digital recordings, there is no noise added in the amplification process. (Please feel free to disagree with me.)

I tend to record at 44.1/wav format to be consistent with CD standards but I guess I will up the sampling rate to 48 kHz in future recordings.

One can go to a website such as http://www.taperssection.com and read all about the dozens of recorders available and their merits and demerits.

[cacm]

4) Now a days MOST amplifiers are o.k. but the WEAKNESS IN CONVERSION IS THE SOUND CARD. In many P.C.S ETC the sound card can actually make the sound worse than the original. VVS & I used professional sound cards& recorded in .wav only & only later on tried MP3 conversion etc. If QUALITY& FIDELITY are SUPREME this is the way to go.
AN EX-AUDIOPHILE, VKV

"Most of the digital recorders you get in the market today have very good A-D converters. If you buy a $500 recorder, I would venture to say that it would be better than a $500 sound card for the PC as the recorder doesn't need the circuits necessary to talk to the PC's bus; the digital signal gets written to a SD card. I guess one could get better components in the digital recorder." Harimau

MY POINTS ARE MORE FUNDAMENTAL. It has to with ANALOG vs Discrete digital sampling etc. In Photography for example DIGITAL PHOTOS TO come anywhere near Analog pics unless you spend TONS OF MONEY THE PICS ARE INFERIOR. In SOUND the problem is similar. It has to do with analog vs pixels , sampling theorem, power spectrum, information (sound or light) as perceived by interaction with Brain etc.
Simply put still NO ONE has been able to come up with a digital equivalent to the TUBE MCINTOSH AMPLIFIER. Same applies to Digital Photographs.
Of course the REASON digital things BEAT the analog tings has to do with COST. The sad case of BETA vs VHS is the MOST GLARING EXAMPLE.It is indeed a revolution that the Panoramic Camera I was involved in the design for moon landings etc which cost close to half a million dollars ( in FIFTIES DOLLARS)my son can using digital techniques come CLOSE to at 500 dollars. DIGITAL THINGS ARE "TOO CLEAN" so many Noise reduction programs esp. commercially available ones had to include a TUBE equivalent! DISCRETENESS with all advantages esp. cost & just relying in LOWER END of SPECTRUM OF FREQUENCIES WHICH IS THE CASE FOR OUR MUSIC is SATISFACTORY ENOUGH.
I was trying to address the problem at HARIMAU LEVEL! Still the FACT REMAINS THE AVAILABLE BEST RECORDINGS ARE NO WHERE CLOSE TO hearing M.M.I. & M.S.S. IN PERSON WITHIN THE RANGE OF THEIR NATURAL VOICE! This bothered S.B. no end as I had worked with him on Contact Mics & other means to get near the purity & found it IMPOSSIBLE. Without misrepresenting his views RAVI KIRAN I feel PREFERRED Chitra Veena over Saraswathi Veena for such reasons. The ONE THANTHI APPROACH OF M.S.G. SOLVED THE PROBLEM AT LEAST FOR VIOLIN.
However it requires a M.S.G. of which Genre I have met only one & thats M.S.G.! ..... As I said: Ex-audiophile with failing EAR-BRAIN TRANSDUCER!
VKV

I have heard it said that in analog recording, one should record as hot as possible and in digital recording one should record as cold as possible. That is, analog recordings should be at maximum signal strength without going into saturation and digital recordings should be at weak signal strengths. I have had no difficulty in amplifying recordings at -48 dB sufficiently so that the signal strength after amplification is close to -3 dB. Of course with digital recordings, there is no noise added in the amplification process. (Please feel free to disagree with me.)
I tend to record at 44.1/wav format to be consistent with CD standards but I guess I will up the sampling rate to 48 kHz in future recordings.
One can go to a website such as http://www.taperssection.com and read all about the dozens of recorders available and their merits and demerits.---Harimau

TRY 48KHZ. With good recordings you will be surprised as I was. Counter INTUTIVE! VKV

[harimau]

Of course the REASON digital things BEAT the analog tings has to do with COST. The sad case of BETA vs VHS is the MOST GLARING EXAMPLE.It is indeed a revolution that the Panoramic Camera I was involved in the design for moon landings etc which cost close to half a million dollars ( in FIFTIES DOLLARS)my son can using digital techniques come CLOSE to at 500 dollars. DIGITAL THINGS ARE "TOO CLEAN" so many Noise reduction programs esp. commercially available ones had to include a TUBE equivalent! DISCRETENESS with all advantages esp. cost & just relying in LOWER END of SPECTRUM OF FREQUENCIES WHICH IS THE CASE FOR OUR MUSIC is SATISFACTORY ENOUGH.
........
TRY 48KHZ. With good recordings you will be surprised as I was. Counter INTUTIVE! VKV

What if I recorded at 24-bit sampling, 96 kHz instead of 16-bit, 48kHz?

Would I get better quality?

There is a different recording technique that uses the SADC mode available only on certain recorders... Something about Super Audio....and supposedly it enables one to recover music no matter what the future recording modes maybe (CD.....etc).

Should I say damn the storage costs and do it that way?

[SrinathK]

I've been using Total Recorder for years now -- it can stream audio from any software source and the sound board option will record from the input mic -- the option I use for tape. What I intended was to describe my recording method in a little more detail (which I will do over the weekend). An addon pack gives a very easy to use noise reduction feature that even allows you to hear the filtered out noise to make sure you aren't losing tones along with the noise (I've found out that very aggressive noise reduction will eliminate musical tones also along with the noise, so I have to take care not to overdo it). As such noise reduction is something I have done only for about 4-5 recordings

Few things also to keep in mind : 1) When recording from a cassette, unless you know what is the effect of each and every knob of the equalizer, DO NOT use it while recording, either on the stereo, or the recording software. Total recorder for one gives me an option to use the equalizer for playback or recording. I keep it disabled while recording. Also check out if Windows mic settings have an equalizer too.

2) The mic volume should not be so high that it saturates the speakers (you can see those 2 speaker bars fluctuating while playing. Ideally they should never exceed 60-70% or move anywhere close to the red zone. Otherwise the recording will turn extremely harsh and then there's nothing that can be done to correct it. A test record would reveal how loud the cassette is and whether the mic volume has to be adjusted.

3) And if anyone is trying to put a mic in front of the stereo speaker and trying to record it to the PC -- STOP ! Always use a stereo pin cable to connect the headphone of the stereo to the mic input of the PC otherwise every background noise will also be picked up -- the quality of the recording is far inferior too.

4) NEVER use a PC's / Laptop's built in mic, ever! Virtually every laptop speaker and mic is an abomination in terms of quality. Always use a stereo pin cable. This holds even if you are trying to record your own music.

5) Use higher bit rates. Min 192 (I recommend 320) kbps for MP3, 192 for AAC MP4 and 224-320 kbps VBR for VORBIS. It adds up to around 2-2.5 MB per min of playback time. If you have the space for it, go for FLAC or other lossless. Especially use the highest settings if recording a streaming music since the original is already lossly compressed.

[cacm]

[quote="harimau"
TRY 48KHZ. With good recordings you will be surprised as I was. Counter INTUTIVE! VKV
Harimau:
What if I recorded at 24-bit sampling, 96 kHz instead of 16-bit, 48kHz?
Would I get better quality?....IT IS CONCEIVABLE it can be better.. I THINK IT DEPENDS ON 1) The type of music in terms of frequency content. 2) MY statements are RESTRICTED to the Experiments I did with respect to GNB CONCERTS as he was prone to vary his Sruti at different points in his career for different concerts.. Actually I readjusted things by making persons who have heard the ACTUAL CONCERTS IN QUESTION to judge the recordings.
Harimau:
There is a different recording technique that uses the SADC mode available only on certain recorders... Something about Super Audio....and supposedly it enables one to recover music no matter what the future recording modes maybe (CD.....etc).
Should I say damn the storage costs and do it that way?... With data costs literally dirt cheap I WOULD SAY: YES. VKV

[cacm]

Dear Srinath, To me long & short of it is any tinkering(Noise Reduction)made after DIGITISING is poorer compared to tinkering made to an ANALOG RECORDING because of DISCRETE SAMPLING OF FREQUENCIES. VKV

[eesha]

Dear Srinath, To me long & short of it is any tinkering(Noise Reduction)made after DIGITISING is poorer compared to tinkering made to an ANALOG RECORDING because of DISCRETE SAMPLING OF FREQUENCIES. VKV

If the signal is sufficiently over sampled , then it makes no difference whether noise reduction is done to analog recording or post digitisation. Beyond that point, it is all placebo effect. If someone says tinkering was done to the analog signal and if am an "analog guy", then I will feel as if it is sounds better than tinkering done to a digitised version.

Similarly with digital vs analog photography, while the pixel values of an analog photo can take all the infinite values in real number space , there is an inherent limitation in human eyes, so we dont need infinite quantization levels.

even in digital photos, when we compare compressed with un compressed, we can see that the top 20 - 30 % eigen vectors are able to reconstruct an image sufficiently well. And discrete cosine transform and other techniques are able to do a lossy compression without any noticeable difference to the normal eye. This being the case, need of analog signals (sound/image) is only an idealistic goal and not required for human eye / ear if digitisation is done at many times the nyqyist rate.

It is again placebo effect and conditioning of mind beyond a point.

[cacm]

Dear Srinath, To me long & short of it is any tinkering(Noise Reduction)made after DIGITISING is poorer compared to tinkering made to an ANALOG RECORDING because of DISCRETE SAMPLING OF FREQUENCIES. VKV

If the signal is sufficiently over sampled , then it makes no difference whether noise reduction is done to analog recording or post digitisation. Beyond that point, it is all placebo effect. If someone says tinkering was done to the analog signal and if am an "analog guy", then I will feel as if it is sounds better than tinkering done to a digitised version.
Similarly with digital vs analog photography, while the pixel values of an analog photo can take all the infinite values in real number space , there is an inherent limitation in human eyes, so we dont need infinite quantization levels.
even in digital photos, when we compare compressed with un compressed, we can see that the top 20 - 30 % eigen vectors are able to reconstruct an image sufficiently well. And discrete cosine transform and other techniques are able to do a lossy compression without any noticeable difference to the normal eye. This being the case, need of analog signals (sound/image) is only an idealistic goal and not required for human eye / ear if digitisation is done at many times the nyqyist rate.
It is again placebo effect and conditioning of mind beyond a point.

Derar E,
WHILE what you have written is generally true in REALITY what you have written IS NOT EXACTLY CORRECT. There are areas called"APODISATION" & "SUPER RESOLUTION"- explored in my Ph.D. Thesis & COMMONLY USED in military applications today at least in USA-- that have been used in using the PHASE IN solving MAXWELL'S equation VECTORIALLY as opposed to Scalar solutions which can result in Resolution & contrast values at specific spatial frequencies in ELECTRONICS& OPTICAL APPLICATIONS greater than the scalar classical values that can be derived BASED on SCALAR approaches described in UNCASSIFIED ARTICLES , BOOKS& JOURNALS. My SPECIALITY IS OPTICS. Believe me these are what are called FAR Field applications & in Laser propagation Sampling theorem is DISOBEYED.....VKV

[varsha]

Best Methods
1.Run the tape on a spare player , up and down .Just to loosen it for optimal performance while converting.
2.If not available , stick a sketch pens cap portion into the hub and rotate in a circular movement,back and forth , by hand.Works well as a meditation tool too.
3.Dispose the tape after conversion.Else you will have a problem of disposing a larger bulk.
Like I have with 5000 tapes now:)
4.Clean the head after each recording.
5.Keep checking samples of wave for smudged waves in one of the channels.Dustdoes not always sit equally on spaces that pick up the two tracks.Very important since tapes are not likely to have been used for years.
6.Keep looking out for interesting bits at the end.One may expect to find remnants of the underlying track on which the concert was recorded.And something more interesting than the concert you thought you were having.
7.Check the audioplayer for short citcuits that may be causing a hum.a loose screw touching a metal piece here etc.
8.Do the playing through a walkman.preferably on dc.always better than ac,though i have never understood why.
9.I have been buying cheap walkmans for 400bucks and retiring them , every few months
10.Link up with collectors and see if the work is already done by someone else.Chances are yes.Unless you were a neighbour to Brindamma and were experimenting with your toy of a walkman in the vicinity.
Life is too short to savour what is on mp3 already.Sincerely hope you have more time than me.
11.After 20 years of converting , I regret the time I wasted in trying to get at the most outstanding quality in terms of bit ratings or conversion algorithms.My settling of the tradeoff has been leaning towards getting on with the listening.
12.Have one original plus two backups.
13 Share with a promiscuous zeal of a casanova.You will never know the kind and and quality of friends you can make.
I have over 300 , on every continent of the world save Antartica.
It is a nice feeling to know that one is connected so widely.Try and see if any of your tracks has a comment Carnatic classical concert archives.If you find one , it came through me.
14.Nadaswaram tracks are most difficult to master because of the tavil.
15.Watch out for one of the members of Alathur duo for his whispered grunts.
16 Every concert where Lalgudi has accompanied , look out for his grunts of appreciation when the main performer is on .

It is a wonderful world ......

[cacm]

Dear Varsha,
While what you have written is VERY FUNNY there is a lot of TRUTH in what you have written. Having set up the VVS Recording system with four full time employees and having spent close to five years PERFECTING the recording techniques I can write more STUFF but I will spare the readers! I can however say honestly that at the end I have satisfactory results & am currently HAPPILY a RETIRED SCIENTIST AS WELL AS EX-AUDIOPHILE...REGS, VKV

[SrinathK]

Dear sir, I wish your MMI archival project all the very best. Since you intend to record in HD, have you considered 192 kbps AAC or Vorbis VBR audio formats?

[cacm]

Dear sir, I wish your MMI archival project all the very best. Since you intend to record in HD, have you considered 192 kbps AAC or Vorbis VBR audio formats?

I do not know whether your remarks are addressed to me. Assuming it is our attempts started & were implemented 15-10 years back. At the time we decided to get the professional level equipment as cost was not a consideration. Next we ran the tape to determine various gross defects & signal levels etc & other usual defects. Then we did not attempt any NOISE REDUCTION as I was fully aware & had written alogrithams for the SMITHSONIAN efforts to
restore the Jazz& Blues recordings of close to hundred years. We just trained four persons to record the tapes & cassettes which were either originals or first generation copies into cd's. Even THE hd's came later & we digitally copied it into them .Of course all the other CONVERSIONS ( mp3 & allied schemes inc noiseless etc and varying parameters etc have been made but I was of the opinion that .wav was the best & while many professionals like M.Chandrasekaran, S.Rajam, Dr.S.A.K.Durga have publicly said in meetings where I played them that it was exactly like the sound they heard in MMI'S Concerts 50 years ago & complemented us. I STILL FEEL NONE OF THE RCEORDINGS ARE AS GOOD AS MMI & MSS HEARD IN PERSON IN THEIR VOICE DISTANCE RANGES....VKV

[Nick H]

I STILL FEEL NONE OF THE RCEORDINGS ARE AS GOOD AS MMI & MSS HEARD IN PERSON IN THEIR VOICE DISTANCE RANGES....VKV

In a conversation on another forum, with audiophiles, I commented that if my hifi system sounded like most of the live, amplified, concerts that I go to, I'd throw it away!

However, apart from a handful of halls which I don't think are even worth setting foot in, the experience of live, on the spot music, combined with interaction with the musicians and with other audience members, is truly a live experience, and the reproduction of recorded music can never be expected to compare. Of course, recorded music has other advantages

[eesha]

Derar E,
WHILE what you have written is generally true in REALITY what you have written IS NOT EXACTLY CORRECT. There are areas called"APODISATION" & "SUPER RESOLUTION"- explored in my Ph.D. Thesis & COMMONLY USED in military applications today at least in USA-- that have been used in using the PHASE IN solving MAXWELL'S equation VECTORIALLY as opposed to Scalar solutions which can result in Resolution & contrast values at specific spatial frequencies in ELECTRONICS& OPTICAL APPLICATIONS greater than the scalar classical values that can be derived BASED on SCALAR approaches described in UNCASSIFIED ARTICLES , BOOKS& JOURNALS. My SPECIALITY IS OPTICS. Believe me these are what are called FAR Field applications & in Laser propagation Sampling theorem is DISOBEYED.....VKV

sir

listening to music is not a millitary application , so I guess we need not solve any Maxwells equations vectorially or worry about any laser propagation where sampling theorem is defeated.

what matters is whether a sharp listener / audiophile is able to discern any difference between 44.1 khz vs 48 khz vs 96 khz vs whatever higher sampling rates that are available/ will be available. And the answer is , no human ear can make out anything difference while sampling beyond 41.khz. Ultimately music is for human ears, so from the subject being discussed (which is "Conversion of tapes to digital - Best Methods"), 44.1 khz would suffice. Anything beyond that is purely academic.

[cacm]

DEAR E, YES WE DO! DON'T HAVE TIME TO ELABORATE HERE. PHYSICS, MATHEMATICS ARE THE SAME. IT IS NOW PROVEN beyond doubt 48khz is SUPERIOR to 44.1KHZ... I am retiring from this due to senility & lack of energy....VKV

[SrinathK]

@NickH It's true. What we are referring to is mic-less voices, heard directly by the ears in very good acoustic environments with no background noises -- especially no fans, a background silent enough to hear the seconds hand tick like a hammer echoing in your ear canal. I remember the sheer number of classroom rehearsals we had done for school annual day programs and we never had problems hearing each other either. But put a mic in and suddenly it feels like you are drowned out and you have to listen to the speaker for feedback, which by the time it reaches your brain is already too late -- this I feel is the main problem behind those out of tune notes we hear often. The problem is that it even affects your tone -- what can you do about tonal modulations or shruti when you're barely managing to be audible to yourself?

Back on topic, my intention was that we should probably make a sticky post on best recording methods for today when Terabyte storage is affordable, that from now on we should push for better quality and higher bit rates and also explore various formats like AAC Vorbis or audio-embedded videos as an option as I have found their overall compression to be better than Mp3 for comparable bit rates (at least my hardware seems to make them sound better. Yes Lossless FLAC too for direct TB harddrive storage, but it's not going to be download friendly until TB internet plans arrive at today's prices with a proportionate speed increase -- but the day may not be too far off when lossless will be affordable & portable enough.

I'd recommend AAC, Vorbis, FLAC and the highest bit rate MP3 in the case of recording audio online as compressing from an already lossy MP3 or whatever results in further loss of quality. In this AAC and Vorbis fare noticeably better while recording from streaming audio.

Being rasikas, almost all of us having vast archives collected painstakingly over decades and generations, I think while many of us are good at digitizing audio, the forum could put up a guide for converting tapes to digital in a sticky post which could be easily understood and be useful to anyone interested. Many of us who have earlier been using lower bit rates can also upgrade to better kbps and better formats if possible.

Lastly, for safety I have so far always kept 2 backups of my collections (except the 20GB stash I accumulated in the last 2 months and which urgently needs a backup) on 2 PCs and external hard drives. That I think is an investment we have to make.

[cacm]

Dear S.K., YOU DESRVE S.K. TITLE! GOOD IDEA TO BACK UP! VKV

[Nick H]

...And always keep a recent backup copy away from your home.

Another hint:

Don't keep your at-home backup external drive attached to your PC. If a power surge fries your PC, you don't want to loose your backup at the same time.

Some of these lessons I have learnt the hard way!

[rahm221]

The DELL laptop I have has no line in port or a mic input. How can I get the sound from the cassette player into the computer?

[Nick H]

You have to use an external audio interface ("sound card").

They range in price from moderate to phew! but, among the lower priced units, There is one from Behringer that is considered usable and good value for money. I'll try to edit in the model number.

...

Here you go, Behringer UCA202 at Amazon.in.

Not exactly a tiny amount of money if we think in idlis or doasi (when I was a child I used to calculate value of large sums in chocolate bars!) but it will also allow you to play out to your hifi, or anything with line-in, as well as record.

You should connect a line-out to the unit's line-in --- but headphone-out will work fine (with the right cable). Just start with the tape machines volume at zero and turn it up slowly until you get the desired level.

[rahm221]

Thanks Nick