I recently upgraded my crappy ION Pro record player to a slightly less crappy Numark one with pitch control.
That got me thinking about how I could manipulate pitch of digital audio through hardware in order to avoid digital artifacts. And this led me to the idea of getting a cassette recorder.

Now the reason I want to manipulate pitch is that I make a lot of chilled, some times, sample based hip hop/rnb and like to pitch down 3 or 5 semitones. 90% of the time pitching down 3 the algorithms sounds decent, the other 10% if the source has a lot of odd harmonics they sound awful. Source content that sounds good at -5 is even harder to come by.

My plan is to get one of the Tascam recorders either a portastudio or one of the rackmount ones, I already have a loopback set on my interface that allows me to record computers playback into my DAW, and my monitor controller which has a "tape out" that bypasses the volume controls. In summary, record for example spotify to tape, change the pitch/speed of the playback and record into Studio One.

Now my questions for y'all:
1) I know tape has a lot of hiss, but I've never used a "professional" grade machine with Dolby A/B/C noise reduction and modern tape, only one of those aux to cassette adapters which suck lol. Would the noise floor be bearable for my application?

2) The Tascam units I've seen have +/- 12% pitch control (which I'm assuming just speeds up and down the motor), can you record at +12% then playback at -12%? This I believe works out to -4.7-ish semitones.

3) Do you have any recommendations for cassette recorders?

4) Any other ideas for how to achieve a higher quality pitch shifting?

 

The basic idee of Dobby NR is:

• EQ UP the Highs when u record
• EQ DOWN the Highs when u play

(more complex systems involve companding which won't be discussed here)

Assuming you have an EQ, you can sorta accomplish this. Hope it helps.

 

Yes I believe type B is up to 15db boost & cut and type C is 20dB, and the amount is dependent on the source level.

but would the noise floor after this type of processing be acceptable by todays standards? or will I have to spend a lot of time in RX doing noise reduction and replacing pitch shifting artifacts with ones from noise reduction?

Also factoring in that the higher end tascams could record at "hi-speed" which also reduces the noise floor.

This is kinda hard for me to determine on my own as all this gear I'd have to buy used and therefore can't try it and return it, so I have to rely on anecdotal facts.

 

Don't know how the Dolby is gonna react to changes in pitch as it will move the frequencies being compressed while recorded, then shifted, then expanded during playback. The compressed frequencies will have been moved so when it expanded, it won't be the same band that was compressed originally.
Experimentation may disprove this theory, but I think not likely.
I'd worry about the tempo straying during playback too. Maybe some kind member here has a Porta One and can do this pitch shifting experiment for you and post the result before you invest.
I believe the noise floor may present an issue, but you may be able to process the noise out afterward using something like RX. RX would be relatively easy as you could record a blank section to sample no signal noise as the filter it applies. My experience with noise removal has been process lightly, and do several passes as needed. If you hear that you went to far with it, just undo to the last good pass.
If you are getting artifacts using a traditional algorithmic plugin, have you tried something different? Perhaps Elastique Pitch. I've used it and it's a pretty good one.

 

Last edited: Aug 24, 2017

Are you asking for multitrack? If not, a Nakamichi, something like CR 7A or a ZX 9 or an MR-1. However, a problem with them was that a tape made on Nak only sounded good on another Nak. But their sound was amazing.

 

@tonyg0499, yes, the Nakamichi's sounded pretty awesome. I think they could play the high bias tapes which helps signal to noise-wise if I remember correctly.

 

I think your point on the frequency shifting is valid and something that I hadnt considered. Assuming the EQ frequency is at 1.5kHz after -3 semitones, you'd be restoring the equivalent of 1.8kHz.

I'm okay with a little bit of noise, as long as I don't trade pitch shifting artifacts for noise reduction ones. If Waves W43 can do the job its okay with me, If it requires RX, that would add an additional step to my work flow.

More or less every DAW uses Elastique. Studio One has 3 separate algorithms licensed from them, I've tried all, I've tried FL's implementation of it, and Melodyne's. The algorithms sound good on solo instruments, but not so much on an mix - and I even sample off Tidal at 44.1khz. On a side note I had done a little experiment comparing -5 of the same song one from spotify (320 ogg) and Tidal 44.1k and the differences were pretty obvious, but still not good enough for me.

 

I'd only need 2-track.

I did a quick search for Nakamichi and wow their cassette decks are going for $400-$1000 on ebay. Unfortunately the ones I've seen don't have variable pitch which is the whole reason I'm considering this.

This CR-7 Sold for $2600
http://www.ebay.com/itm/NAKAMICHI-C...ar-Warranty-Near-Mint-Condition-/131636163409

 

Yeah, I'd think uncompressed (wav, flac, etc.) would fare better than something coming down from a stream or mp3 or such. If you have much choice about the source that is.
You might could source a decent reel to reel as well. I'd think that the quality/noise would be better, especially with a higher tape speed. The question there would be if it had variable speed, and how much it could be adjusted.
Just trying to throw ideas out.

 

What?!! $2600!!! I wonder what the 1000ZXL would cost today.
Prices for analog vintage gear is out of control. MR1 had pitch and the BX300s. I used to record my radio mixshows only on Naks. I remember recording from vinyl to the Naks to preserve those rares 12 inch test pressing copies that sounded better than the retail records.The Naks had a sound that was hard to find on other decks. I use the pitch control to mix in and out of songs being played on the Naks.

As way of explanation to the newer generation,allow me to say that the better sound quality of test pressings resulted from the test pressings being the first made with the stampers which wore out as more copies were being pressed. There was something special about a test pressing mastered by Francois Kekorvian being played through a Richard Long sound-system.

 

Last edited by a moderator: Aug 24, 2017

So I found a relatively modern Sony Tape deck for 80$ the TCWE675
+/-30% pitch, records in double speed, every version of Dolby noise reduction.
I heard recording of its playback and there was next to no hiss. (the specs say -70dB noisefloor with Dolby C)

Only problem is that it has a ton of additional controls and functions, like multiple memory spots and what not.
Im gonna contact the seller and pick it up tomorrow if its still available.

 

All the research I've done tonight has been very interesting.
I've found that in the same way there are vinyl die hards (I personally enjoy vinyl for the experience, artwork and merch factor not the sonics), there are cassette aficionados. I always thought cassettes were the bastard middle child of audio formats and that eventually history would skip over that period and go from straight from vinyl to CD lol. Turns out that is not the case.

But it makes sense, cassettes are like vinyls that you can record to.

 

The "Nakamichi Sound" was the result of the Naks heads being equalized differently than other manufacturers'. The Naks have a non-standard IEC record/playback EQ curve. It gives better sound with less distortion if played back on the same deck but "not quite right (usually kinda dull) when played on a machine with the stock IEC playback EQ. What is perhaps responsible for making the Nak sound more identifiable is that they were famous for being able to record up to 20khz as well as recording at higher levels with low distortion, when few other machines could come close to that.

Check out http://www.naks.com/

 

Some of the Naks ( The Dragon) had quartz lock and double direct drive. The The ZX-9 is direct drive on the take up capstan with the supply capstan being belt driven off the take up, like most other direct drive transports.

 

Great info, @tonyg0499!

 

It brings back lots of memories to talk about them Naks. Every major studio, mastering house and club in the 70's and 80's had at least one Nak.

 

These old tape decks usually need a thorough refurbishing, capstan rollers and other moving rubber parts like the drive belt might need replacing. There are parts available on the net and there are folks who love to restore them, so be prepared to invest a bit more for upkeep and overhaul.

 

You brought up an interesting point. AFIK Nakamichi repairs are very expensive. There was a guy named Willy Hermann that was the best when it came to repairing the Naks. I have no idea if he is still around.

 

Why not forget cassette which are the worst of all magnetic media besides 8-tracks. Go with Reel to Reel. The speeds are faster for better quality and less noise, and they have better Dolby electronics if it has them.

 

If I get you right the reason you want to use tape is that you want to avoid artifacts from pitch shifting algorithms that preserve time.

Actually you don't need any hardware to do that. All you need to do is to playback your audio with a different sample rate and resample it. It's the same effect as playing back your tape recordings at different speed.

The more complex time/pitch shifting algos enable you to preserve time and formants, if you just need a simple pich&time shift sample rate manipulation is the way to go.