changing file speed with Audacity (audible quality losss ?)

Bomber, why do you want to speed up the audio then slow it down. It seems to me like digging a hole then filling it up. I do not like to fiddle with my music so I can't fathom why you would want to do it unless you are just testing audacity.

 

I thought I said this earlier...
Have lots of old recordings of live stuff, when I get transfers from master cassette recordings they can often be at wrong play speed, so I correct the play speed as best as I can, and disregard the old cassette or whatever once I've burned it to CD or DVD-A...
I have a lot of stuff in the first place, so keeping backup of everything takes up too much room.
But sometimes when listening, can be months later I feel I might not have got the speed correction right the first time, so I have to adjust the speed on something I have already previously adjusted the speed on...
Obviously if I care enough about the music to correct it to the proper play speed then I care enough to not want to do anything which will be causing audible quality loss.
*OK I did a little test, I took a nice crystal clear live recording of a track I'm very familiar with, I changed the speed of it back and forth 9 times (which is ludicrous !), and at least with my computer speakers I'm having trouble deciding whether I can really hear a difference or not, sometimes I think there is a very subtle difference, but I'm not even sure if my mind is just fooling me, although I know I have a fairly good ear. So if that's the result I get from 9 changes I guess I don't really have to worry about 2 changes !

 

Can you hear the difference between a good quality wav file and the 128kbps mp3 of the
same file ?

 

Yes I can hear an obvious difference between a lossless WAVE and a lossy 128kbps MP3 file of the same source for any song.
In the case you just asked about I can notice the difference after listening for just 2-3 seconds.
Why do you ask ?

 

I haven't tried the Audacity stretch myself, I will give it a try.

Some people are more sensitive to 128kbps mp3 than others. Sometimes, to me, they sound
terrible, and other times it's passable. I think it depends on the material
and on whether the file gets any benefit from using joint stereo.

On the face of it, your statement that a 9 times stretch resulted in negligible
degradation sounds surprising - but I would be wrong to judge it without trying it.

 

Yeah I'm very surprised I can't notice a difference after 9 times, the whole point of me starting this topic was that I was worried there would be an audible difference after just 2-3 times!
And let me tell you I gave it a solid work out!
The original song was 7min9secs, and a very clear clean sounding recording.
I changed the speed percentage +7, -5, +3, -8, +4, -0.3, +0.1, +0.05, -0.03
Back to the time of the original file within .01 of a second.
I'm really struggling to convince myself I can hear a difference.
But yeah the original WAVE vs 128kbps MP3 the difference to me is very clear.
*NOTE - I just tried converting to 128kbps MP3 using dBpoweramp and the difference between it and the original lossless file is nowhere near as clear as when I converted to 128kbps MP3 using Real Player - seems dBpoweramp did a much better job converting to MP3.

I've been meaning to ask, with the Audacity speed change, would you expect audible quality loss to be more obvious when changing a file that was originally very clear and perfect sounding ?, or one that was originally a little imperfect with some noise like hiss etc ?

OK if it is of any interest to you I have uploaded at MegaUpload the exact file I used to experiment with using Audacity speed change, it's the original FLAC from my CD, you can download it here :
http://www.megaupload.com/?d=BTQA0LFW

 

OK Bomber07, I'll give your file a try in Audacity and I'll try it in Sound Forge 8
using their routines. In SF8, there are a bunch of settings, for example, "voice",
"music", 'guitar" - each leaves a slight flavor on the stretch which tales you away
from the original.

SF8 is 5 years old. Perhaps Audacity has found an improvement in the process?
I don't expect to get to his until tomorrow now (because my son is using my box).

 

No worries, I'm just curious as to how my ear compares to a more 'expert' opinion (re : how the quality of the Audacity speed changed version reverted back to original speed compares to the original unedited file).

 

Dave,
All 128s are not created equally. Some use advanced compression before the truncation process. These can sound good if the music lacks acoustic instruments. You also have the young who have the ability to hear the high pitch but have not developed the ability to listen carefully.

I am less surprised that you can't hear a difference after 9 times stretch. Lossless is what, 1300 BR and in a double blind test even great listeners can't tell 190 CBR from lossless. You have a massive buffer to ruin before you can hear the difference.

Thanks for the more detailed clarification of why you are doing this. Still, you might consider backing up a dozen or so tapes to a DVD using Flac or Ape at the highest compression before you work on them. The compressions are surprisingly efficient. Then when you are satisfied with your twiddling, note the time length then take the back up and process it only once. That will preserve the most quality. Are you using the pitch to determine the proper speed?

 

Have to agree with davexnet, depends on the material and of course the encoder used..
That said, when making statements like this,
"Yes I can hear an obvious difference between a lossless WAVE and a lossy 128kbps MP3 file of the same source for any song" one wonders how you went about testing..ABX?
Just wondering is all, and as davexnet pointed out, we'd be wrong to judge..
Now over @ the HA forums, statements like that will get you a smack-down w/o backing it up...And they will demand your ABX results..

 

Bomber07 I'm looking at Audacity. What control are you using?
In "effects", I see change speed, pitch or tempo.

I assume you're changing speed or pitch, since incorrect pitch is the main
problem when cassette is played too fast / slow.

I must admit, my testing with "change speed" got good results, does indeed seem
superior to Sound Forge 8.

I came across this problem once myself a few years ago. However, I fixed it in the
analog realm. I decided the cassette deck itself was out, so I opened it
up and adjusted the little screw on the capstan motor to slow it down a little.
It's been fine ever since.

 

Did you not see I then updated / edited my post to say :
*NOTE - I just tried converting to 128kbps MP3 using dBpoweramp and the difference between it and the original lossless file is nowhere near as clear as when I converted to 128kbps MP3 using Real Player - seems dBpoweramp did a much better job converting to MP3.
eg. Real Player conversion to 128kbps MP3 produced very audible loss of quality, dBpoweramp & Adobe Audition conversion to 128kbps is quite hard to tell from the original un-edited file.

 

Hello,
I'm using the "Change Speed" button.
I can't always fix the speed on a cassette deck, as I have quite a few live casettes transfered to digital which were friends master casettes etc., I often only have access to the digital transfer from master cassette.
I enjoy my live music!, but sometimes you can find a song which should correctly play at 3:15 minutes plays at something like 3:00min. Which sounds clearly out of pitch / off speed, creeping towards sounding like chipmunks! haha
I'd be interested to know if you can tell the difference between the original un-edited file, especially after 2-3 speed changes.

 

You either converted it or you didn't.. which is it?..

Either you can or cannot tell the difference..Which is it?
Sorry if it seems I'm nitpicking..

 

Sorry when I said I 'tried' converting the lossless to 128kbps MP3 I mean I 'did'.
The 128kbps MP3 created with Real Player was awful sounding, clearly degenerated, absolutely no need for any kind of blind test.
I then created another 128kbps (CBR) MP3 using the dBpoweramp programme(I think Lame 3.98r encoder) which I found much harder to tell the difference between it and the lossless original.
I ended up mixing up the 128kbps (CBR) MP3 and the lossless original, so I didn't know which ones they were. I listened to one part of the song for 5 seconds (repeated a few times) and decided which I felt was the lower quality sounding one. I mixed them up 10 different times and listened to a different part of the song each of the 10 times (about 3-4 listens lasting 5secs to each part) and 9/10 times I correctly chose the MP3 as the file which sounded lower quality (the only time I got it wrong I was feeling distracted by the noise of kids outside yelling and bouncing a ball).
I could tell by listening to the high pitched part of vocals, guitar, or listening to the hi-hat.
So yeah when I sit down and concentrate on it I do feel quite confident I can tell a 128kbps MP3 created with decent programme apart from a lossless version.
***EDIT*** I just did the same test using the original lossless file and file speed changed with Audacity 9 times, it was much more difficult for me to decide which was the higher quality version, it took me 2-3 times longer to decide each time, and I only got 5/10 correct.
So it really seems Audacity speed change is doing a good job at keeping the quality. I am very surprised...
I used the song I posted a link to in one of my previous posts.
I'm not sure why I was asked in the first place ?, why do you want to know if I can tell the difference between lossless and a 128kbps MP3?

 

Just trying to keep you all honest...

I have a real test you, if you have the album "Don't Shoot Me I'm Only the Piano Player" by Elton John, see if you can tell the difference between a 128 and lossless. As far as I can tell, all the extreme highs are missing on both the CD and vinyl. Any acappella will also completely fall within the range of a 128 CBR, probably within a 30 CBR.

I truly doubt that dbPowerAmp was at fault for the lower performance 128. I am not defending it but pointing out it was probably your selection of an encoder and the settings. As I mentioned earlier, some encoders use advanced compression before they truncate others just truncate. At 128 this is very noticeable. AAC, Realmedia and WMA all use the same basic encoder that they have modified, which uses psychoacoustic compression before they truncate the highs. In case you do not know or forgot, that compression is extremely effective at removing data you will not hear. It is lossy but not hear able lossy used in VBRs. Since you could have chosen any of those encoders using dbPowerAmp and did not, I think the fault was yours.

 

Keep honest ?
Why would I lie about what I can and can't hear ?
Davexnet asked me the question "can I hear a difference between the lossless and a 128kbps MP3?", maybe Davexnet should specify exactly which encoder/settings etc. he meant because this seems to have gone to the extreme, and maybe getting a little off track from the reason I started this topic about wheter using "Speed Change" in Audacity causes audible quality loss, unless the 2 somehow relate to each other in a way which hasn't been clearly explained to me !?!?!

I did some tests using the file I posted a MegaUpload link to earlier in the thread.
OK when converting lossless to 128kbps using dBpoweramp I just changed the original ENCODER Lame 3.98r CBR (which you suggested does not do a high grade job), and I changed the dBpoweramp settings and CONVERTED to Windows Media 10 Audio, using Windows Media 10 Audio Professional CODEC, the TARGET was Bit Rate CBR
and a ticked 2 Pass Encoding, SETTINGS were 128kbps 44kHz 2channel 16bit CBR.
I believe they are the settings you are suggesting are superior ?
Using those settings I just did a test by mixing up the 128 with lossless, I correctly chose 10/10 times the 128kbps file to be the lower sounding quality file, it was quite easy when listening to the hi-hat, to my ears it didn't sound any better than the original CODEC Lame 3.98r CBR MP3.

I then changed the CODEC to Windows Media Audio 9.2, Quality VBR, 2 pass, 128kbps, 44kHz, stereo VBR and once again I correctly identified the 128kbps file 10/10 times, however I don't understand why the file says it's 143kbps when I chose 128kbps, is it really 143kbps?

Then a 3rd time using CODEC Windows Media Audio 9.2, Quality VBR (2 pass unticked), VBR quality 75, Stereo VBR, the output file says it's 149kbps, I correctly identified the lower quality file 10/10 times.
To be honest I don't see how "variable bit rate" is supposedly higer quality, as I heard some really awful patches where the sound was quite obviously lower quality without even having to compare to the lossless file, and this was supposedly 149kbps.

Oddly the original MP3 encoder I used (which you suggested I was at fault for using) seems to me like it did the best job, and has been recommended by Hydrogenaudio after extensive tests as being the most reliable MP3 encoder......
http://wiki.hydrogenaudio.org/index.php?title=Lame

This seems to have gone off track as I can't see how it relates to if anyone can hear the difference berween a lossless file and the same file Speed Changed with Audacity 2-3 times.

 

bomber07, don't take it so darn personal..AFAIC, you're the one that's now off-topic..
Seems like you've already determined, based on your particular test method(s) that you can hear the difference between a lossy file with different encoders/settings etc..Fine, as was stated on the other forum(same topic) that certainly is plausible..
Who are we to dispute, since none of us share ears, eq, listening environment, yada, yada..However, as you well know, one can claim anything on the net..Which is why when claiming test results of say, 10/10 correct, you will be asked to provide double-blind ABX test results..Not so much in this forum but certainly over at HA, which they have, btw..FYI, this nothing new..Many folks can tell the difference between a lossless file and a lame encoded 128kbps VBR..Which is why if you want a transparent setting use a higher one e.g the recommended settings..But it's a good idea for you to conduct a proper ABX test to determine your sweet-spot..
I highly doubt, our good friend Mez, was/is calling you a liar..Just trying to keep us all honest..
Oh, and btw, I love ' The Cure'..

 

I know he wasn't really calling anyone a liar, I was simply responding to his comment about keeping everyone honest, by remarking why would I lie !?, I'm not taking it serious at all, I was joking just as much as he was... ;-)
Yeah I know a lot of people can tell the difference at 128kbps, A LOT !, that's why I'm wondering why I've ended up trying all these different settings/codecs etc., I'm not sure how it all relates to the reason I originally started this thread, which was - does the Audacity "Speed Change" function create audible quality loss...
I've given my own results of a blind type test, I did not hear any difference (though I can with a 128kbps MP3), so I guess there's not much anyone else can add, unless people have their own experiences with the programme/function...
I'm just not sure what relevance the whole MP3 discussion has?, I didn't start it, because I have nothing to do with MP3's, I never ever under any circumstane ever convert anything to MP3 !
Cheers
- Though I am now quite interested to know some statistics on how many people can detect quality loss by ear getting up to 192-320kbps etc, and what the best encoders/settings are etc.

 

B/C for many it is transparent or close to the original but much, much smaller in size..Even with today's large capacity storage device(s), my first guess would be, portability..
FWIW, my personal transparent settings(with most samples) is lame -V2 ~192kbps VBR or AAC/m4a quicktime TrueVBR~256kbps, for my portable(s)..Not saying it's the "best encoders/settings", only what works and is transparent from the original(s) to my ears..
And now we really are going off-topic..
Cheers!