[Brent Hutto]

I know that with video being digitized from movies on film it can sometimes be tougher or require more bit rate to properly capture grainy and/or dirty old prints than a fine-grained, pristine modern print. Does any of that apply to old audio recordings?

I'm thinking of some of the CD's I have of old jazz recordings from the late 50's/early 60's where the fidelity is quite good but there's a certain amount of tape hiss even on the remastered CD and certain analog (tape) artifacts that we were all used to hearing routinely back in the day. Most of this stuff was mono but they remastered it into stereo. Or it's sometimes not remastered but just fake stereo.

So for me a modern jazz ensemble recording at something like 256k or 320k AAC is indistinguishable from a WAV file (I've done plenty of ABX tests to find out). Would there be any reason to think a cleaned-up, remastered, stereo-ized version of a noisy old tape master would cause less fidelity at those bit rates?

I will say in ripping some of this stuff to ALAC it seems to (losslessly) compress a bit more than similar modern recordings. So I think the answer is no, it is not harder to compress.

 

[Satellite_6]

I would think that it would just be harder to tell lossy from the original.  . .

 

[xnor]

In my FLAC music collection:
 
average bitrate for old recordings (70's, 80's) or recordings with little compression: 878 kbps
average bitrate for new recordings or loudness war casualties: 1044 kbps
 
all files had the same format 44.1 kHz, 16 bit
ymmv.

 

[xnor]

Yep, did another test. I took a progressive rock track with great dynamic range and compressed it with FLAC level 8, bitrate: 804 kbps. (upper waveform)
Then I DRC'd it with a multiband compressor with a broadcasting preset and compressed it again with FLAC level 8, bitrate: 963 kbps. (lower waveform)
 

 

 

[Brent Hutto]

Very interesting trial you did there.

I have a few CD's stored in Apple Lossless (ALAC) format. Some of them are old mono converted to fake stereo albums and others are modern recordings in the same jazz and classical genres. I can't see any correlation between resulting lossless-compressed file sizes and being old and hissy versus modern and clean.

But none of them are going to have any noticeable compression done to them. So that seems to be the one thing that can jack up the difficulty of shrinking files. OTOH, maybe the compressed ones are actually easier to fool the ear with using lossy (MP3) compression. Hard to say.

 

[xnor]

Hmm I encoded both with lame mp3 VBR V0 and again the dynamic range compressed file has a higher bitrate, not much though (241 vs. 248 kbps).
 
Can't say much about fake stereo, old hissy or modern clean recordings.
 
 
edit: Actually, there is some noise on the recording I used so I removed that and now the bitrate is even lower (FLAC, level 8): 780 kbps.
 
I guess noise / random signals are harder to compress, would make sense anyway.
 
edit 2: FLAC compressed pink noise results in 1239 kbps, white noise in 1375 kbps.
 

 

[khaos974]

And don't forget that the louder the signal, the less compression you''ll be able to get simply because the values of the waveform are larger, I think that's the main reason why lossless codecs are less efficient with brickwalled/*modern* music.