I used to routinely transfer 2 track music masters backwards. The 
results were noticeably better than a transfer made forwards. The tapes 
were non-Dolby encoded (I was never a fan of noise reduction for music 
recording). Azimuth is absolutely critical. It has to be spot on as well 
as the playback EQ calibration. This process was always done on the same 
machine that recorded the master tape. Azimuth and playback EQ are 
calibrated with the tape playing forward and then the tones are played 
in reverse, recorded and observed. If there is any difference in the 
recorded level of the source tones on the reverse copy, then the 
playback alignment has to be re-checked and the culprit is usually 
azimuth. I always adjust azimuth with a dual trace scope and overlap the 
channels to insure absolute phase although there are a few ways to 
calibrate azimuth and get it right. When it comes to the absolute 
polarity of the copy, it was never an issue because the phase 
relationship remains the same if all is adjusted properly, even though 
absolute phase is reversed. Did many A-B listening tests with everyone 
concerned and an overwhelming majority preferred the backwards transfer. 
Those who weren't sure could usually not tell the difference. Then, of 
course, there were those nervous producers who were afraid of anything 
outside the box.

I have never tried this with 1/4 track or 4 channel formats and Richard 
Hess makes a valid point about the difference in 4 channel heads vs. 1/4 
track. I did try the process on a 2" 24 track tape and the results were 
not great and I have to reason that it was an azimuth issue because 
multi-track heads are never perfect. The 2" transfer was tried on an 
AMPEX MM1200 which are fixed azimuth machines.


Corey Bailey Audio Engineering

On 2/7/2016 8:39 AM, Richard L. Hess wrote:
> Hi, Tom,
> This is something I would love to hear Jay McKnight's answer to, but I 
> think Ted and John Chester succinctly nailed it. The below was mostly 
> written prior to seeing the other two great replies.
> The first comment: you will also need to reverse the absolute polarity 
> of all tracks. I had this confirmed when I did a two-track 
> letter-from-Vietnam tape recently. I did it in one pass (nice thing is 
> that the azimuth alignment works the same for both sides if they were 
> recorded with the same azimuth). The system had a very asymmetrical 
> voice waveform and it was VERY obvious that the two sides were in 
> opposite polarity.
> As to better, it's hard to say. Why don't you take a good tape and try 
> it? See what you think. Of course, to remove absolute polarity bias, 
> you will need to invert the polarity on the reverse-play copy 
> (assuming the forward play polarity is correct). If you like the 
> reverse play uncorrected better, then flip the polarity of the forward 
> play version.
> The best thing would be to sort-of align all 8 tracks in the DAW and 
> then gang the two sets so you can solo one set quickly to do rapid 
> switching A-B comparisons as well as long-form listening.
> There are two thoughts about this:
> (1) since analog filters have group delay, running it backwards 
> compensates for that to some extent
> (2) since analog filters have group delay and it's expected, running 
> it backwards messes things up compared to how we are used to hearing 
> tapes sound
> An interesting thought was what Studer published about their new 
> attempt at tape equalization and phase compensation. This throws 
> another variable into the equation. I always wondered how this 
> affected interchange with machines that weren't configured this way. 
> This was initially published upon the introduction of the A810. Since 
> the A820 uses many of the same electronics cards, one might think that 
> this is also applicable to the A820. I am not certain. Goran? I 
> suggest reading the papers, however. I have them on my server, but I 
> just checked and they are still on the Studer ftp server.
> and
> If you would prefer to click on the individual files, point your 
> browser to:
> SwissSound_A810_Phase_Compensation.pdf
> PI_1982-01_E_A810_New_Developments.pdf
> As the two other responses said, there are other networks in addition 
> to the R/P "standard" EQ, including the head itself. Based on my 
> experience in doing this, whatever the difference is, it is small. I 
> find that often absolute polarity is a bigger difference than 
> direction of playback. But, I generally do reverse playback on voice 
> grade tapes and most of the voice grade tapes I get are poorly 
> recorded but they have Grandma's voice on them.
> As an aside on this entire process, be careful in transferring 
> quarter-track stereo as inline heads have measurably poorer crosstalk 
> performance than the normal quarter-track stereo heads.
> I had originally thought to modify an A80 with four playback 
> electronics channels--it's not THAT hard and I have a spare bucket, 
> but when I thought about the crosstalk, I simply mounted a 
> quarter-track stereo head and for those quarter-track stereo tapes I 
> choose to do on the A80, I do each side in a separate pass.
> Cheers,
> Richard
> On 2/7/2016 8:48 AM, Tom Fine wrote:
>> Hi All:
>> I am about to transfer a small pile of Quad reel tapes, and wanted to
>> revisit this question -- which also applies to full-track and 2-track
>> tapes -- will I get better results playing the tapes tails-to-heads (in
>> reverse) and then reversing the digital file (back-to-front), and of
>> course assigning the correct tracks to the correct channels (the tape
>> would be upside-down if played backwards, so 1=4, 2=3, 3=2 and 4=1). I
>> have read commentary that playing a reel backwards allows for sharper
>> wave fronts and thus crisper dynamics. It seems like it wouldn't _hurt_
>> anything to play the tapes backwards, but I would like more info from
>> the tape-playback experts.
>> Important to note - I know this can't be done with NR-encoded tapes,
>> that those must be played forward (the regular way) for the NR decoder
>> to work properly. But what about if I transferred the tape backward and
>> then send the digital audio out to the decoder, is there any reason that
>> would work? (I don't think so, but wanted to ask the experts).
>> Thanks in advance for facts/discussion about this topic.
>> -- Tom Fine