Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: ShrinkMyTunes: Good idea, or bull? (Read 19216 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

ShrinkMyTunes: Good idea, or bull?

I recently saw an article on Wired.com about a product called ShrinkMyTunes (http://www.shrinkmytunes.com/en_gb/) that is designed to shrink MP3 files so that you can fit more on your portable device. They claim to use "content sensitive heuristic optimization algorithms" developed by NASA.  Based on the example files they have hosted, it looks like they downsample and reduce the bitrate to something ungodly (the first example is 59kbps VBR/24kHz).

Personally, I think it's a crock of shit, but what do you guys think? Is it worth using that low of a sample and bit rate?

Edit: Their "original files" are 128, 160, 192 and 256kbps CBR and 44kHz, but all the "shrunk files" are ranged from 55-59kbps VBR and 24kHz. That doesn't seem very up-and-up, at least in presentation.

ShrinkMyTunes: Good idea, or bull?

Reply #1 –
probably bullshit.

ShrinkMyTunes: Good idea, or bull?

Reply #2 –
24kHz. There's your answer. Goodbye high-frequency content!

ShrinkMyTunes: Good idea, or bull?

Reply #3 –
Well Wired.com says it seems to work in their tests:

http://www.wired.com/software/softwarerevi...9/shrinkmytunes

We need someone to buy this puppy and give it a thorough test. Their web site says 88% of people can't tell the difference between their squeezed MP3 and the original MP3! Let's see what the audiophiles on HA have to say about those results...

Quote from Wired article:
"ShrinkMyTunes uses a patented algorithm to analyze each song, studying the different sonic elements to see where it can cut corners by removing redundant information or slightly smoothing out the finest details in the sound waves.

The result is a file that plays back in any MP3 player and sounds only the slightest bit less punchy, but is much smaller in size. To the trained ear, the converted MP3s sound like they have a medium amount of variable-bitrate (VBR) compression applied. But the quality-to-file-size ratio is much higher than what you get with built-in VBR rippers from iTunes or Windows Media Player. Even LAME's best settings can't beat ShrinkMyTunes at these file sizes."

ShrinkMyTunes: Good idea, or bull?

Reply #4 –
Seeing a reader's comment on Wired, I confirmed that the samples on ShrinkMyTune's website have a LAME header, with a version field "3.97". 



It seems that the Best Compression mode resamples to 24KHz and encodes with a VBR bitrate ~60Kbps, while Best Quality resamples to 32KHz and encodes ~100Kbps.  So much for their purported "proprietary patented adaptive variable bit rate compression encoding", especially if they're using LAME!  I wonder if anyone could re-create the exact settings they use, just for fun...

If true... then they are essentially selling LAME and a frontend.  o.O  For $40 a pop.  Wow.  That's audacious.

ShrinkMyTunes: Good idea, or bull?

Reply #5 –
Quote from: Canar on
24kHz. There's your answer. Goodbye high-frequency content!


At those kind of bitrates, they're probably lowppassing pretty low, too...

ShrinkMyTunes: Good idea, or bull?

Reply #6 –
Theres a detailed review & test samples at: PlayerBites.com
To my ears, too much is lost.

ShrinkMyTunes: Good idea, or bull?

Reply #7 –
Quote from: Funkdude on
At those kind of bitrates, they're probably lowppassing pretty low, too...

~10.0 kHz lowpass @ "Best Compression"
~14.7 kHz lowpass @ "Best Quality"

ShrinkMyTunes: Good idea, or bull?

Reply #8 –
I ABX'd the test files on their site:

Quote
foo_abx 1.3.1 report
foobar2000 v0.9.4.3
2007/09/07 22:55:56

File A: outpatients_uncle_harry.mp3
File B: outpatients_uncle_harry_shrunk.mp3

22:55:56 : Test started.
22:56:09 : 01/01  50.0%
22:56:15 : 02/02  25.0%
22:56:24 : 03/03  12.5%
22:56:31 : 04/04  6.3%
22:56:39 : 05/05  3.1%
22:56:45 : 06/06  1.6%
22:56:52 : 07/07  0.8%
22:56:58 : 08/08  0.4%
22:57:08 : 09/09  0.2%
22:57:13 : 10/10  0.1%
22:57:21 : 11/11  0.0%
22:57:27 : 12/12  0.0%
22:57:35 : 13/13  0.0%
22:57:42 : 14/14  0.0%
22:57:48 : 15/15  0.0%
22:57:51 : Test finished.

----------
Total: 15/15 (0.0%)


The "shrunk" version is not just easy to distinguish it is absolutely unlistenable.

I noticed that they used LAME to encode the "shrunk" MP3s but I saw no offer for the source code of LAME on their site and I bet they don't distribute the LGPL with it either. I have requested the source code for LAME from them.

ShrinkMyTunes: Good idea, or bull?

Reply #10 –
40 bucks for a transcoding application that is actually freeware ... it's time to take these people out of business.
The name was Plex The Ripper, not Jack The Ripper

ShrinkMyTunes: Good idea, or bull?

Reply #11 –
Quote from: Zarggg on
They claim to use "content sensitive heuristic optimization algorithms" developed by NASA.

They made b.s. claims, so they've passed "development" over to someone else, unfortunately NASA. 

Quote from: rjamorim on
Haha, reminds me of Mafuka!
http://www.rjamorim.com/rrw/mafuka.html


Had it not been for the little tidbit of info you put at the bottom of the page I would've sworn it was the slang for what's quoted below.
I'm just glad my last name isn't that! 

Quote
An interesting detail is that "Mafuka" is believed to actually mean an english slang, "mother f**ker".

ShrinkMyTunes: Good idea, or bull?

Reply #12 –
Quote from: Borbus on
I noticed that they used LAME to encode the "shrunk" MP3s but I saw no offer for the source code of LAME on their site and I bet they don't distribute the LGPL with it either. I have requested the source code for LAME from them.


Amazon are selling this however looking at their terms and conditions you can't post a review repeating a criminal accusation. I suspect they'd use this to stop a review informing people of the rip off this software appears to be even though contract law != criminal law.

I wonder how you inform them of this kind of thing...

ShrinkMyTunes: Good idea, or bull?

Reply #13 –
Someone on iLounge asked me about this program a while back.  I am glad I told them not to use it as it looked like BS.  Seriously, someone needs to stop this company from making these outrageous claims.  If the software was free then I don't really see a need to complain too much but people are putting down $40 for something that foobar2000 can do for free.

I was easily able to ABX my original mp3 file with a mp3 file I made using their settings.  The source came from myself that was encoded from a lossless ripped directly from the CD.  I used Lame 3.97 at -V 0 --vbr-new so the resulting file had a bitrate of around 259kbps VBR.  The file produced with their settings (transcoded from the -V 0 mp3) ended up having a overall average bitrate of 99kbps VBR.

I used the stock Apple earbuds that come with iPods and I could still ABX a difference.  That is just sad.  Someone needs to contact this company (or someone who can stop them) for trying to get people to spend money on their junk software.  They say it is better than LAME but they use LAME in the application itself.

ShrinkMyTunes: Good idea, or bull?

Reply #14 –
This is hilarious:

Quote
But can't I do this anyway by re-encoding at a different bit rate or using VBR encoding?

As part of the shrinking process, the song is analysed by ShrinkMyTunes' proprietary content-sensitive heuristic systems. ShrinkMyTunes converts your MP3.s  and WAV files to a variable bit rate (VBR) format if it is not already in that format, which only counts for small part of the file size reduction. The majority of the size reduction is achieved using the ShrinkMyTunes patented adaptive variable bit rate audio compression encoding. The system decides the best bit rate and sample rate the file can be encoded at to allow the song to maintain as much as possible the current song quality.

So they have actually patentend LAME's VBR mode?


Quote
Does shrinking my tunes affect the sound quality of my music?

The sophisticated heuristic algorithms used in ShrinkMyTunes cause minor audible change to yourMP3s most of which cannot be detected using most standard MP3 players. In research conducted prior to releasing ShrinkMyTunes, 88% of research participants could not correctly select the shrunk MP3 from the original. 
If you are concerned about sound quality of your music collection, we suggest saving your shrunk tunes into a new folder when the software prompts you. That way, you can keep your original MP3s in their current format and have a new folder of shrunk MP3s.

To read an independent review of ShrinkMyTunes song quality please click here

At least they more or less warn their customers and advise them to keep the originals when in doubt.

ShrinkMyTunes: Good idea, or bull?

Reply #15 –
Quote from: chelgrian on
I wonder how you inform them of this kind of thing...


At risk of replying to myself this is a Z-Group (http://www.zgroupplc.com) who are well known for very dodgy marketing claims. It's the same company that is behind On-Speed which essentially just reconfigures your web browser to use their compressing and caching proxy then claims "broadband speeds" from a 56k modem.

They also have the audacity to sell a "mobile" version possibly they are counting on people not knowing that Opera Mini (which does exactly the same compressing proxy trick but is open about what it is doing and wha the limitations are) exists.

ShrinkMyTunes: Good idea, or bull?

Reply #16 –
Note that LAME is developed under the LGPL. This is slightly different from the GPL. As long as they do not modify LAME itself, but simply dynamically link their software to LAME, their software can remain proprietary. However, they still must openly offer source code for LAME and also distribute the full text of the GNU LGPL with their product.

If they have actually made a derivative work of LAME then they must provide the source code for the whole piece of software (for download, since the software itself can be downloaded) and, obviously, distribute a copy of the LGPL with their software.

I wonder if someone can take a look at the software if they can get hold of it and see if they actually are only dynamically linking to the LAME library or if they have created a derivative.

ShrinkMyTunes: Good idea, or bull?

Reply #17 –
Quote from: Borbus on
I wonder if someone can take a look at the software if they can get hold of it and see if they actually are only dynamically linking to the LAME library or if they have created a derivative.

The existence of lame_enc.dll would satisfy the first part of your query, while a binary comparison would determine the second. While there exist compilers galore, not many would be used to compile LAME, so few comparisons should be needed. Note that statically linking (bundling) LAME into the main binary or a wrapper is considered a derivative work (by the FSF) and would be in violation of the LGPL unless the software also provides linkable object files (or source code); the LGPL requires that relinking to a compatible library remain possible. Note that this requirement assists with possible infringement cases as the replacement LAME must run--else the interface has changed and is thus a derivative work.

ShrinkMyTunes: Good idea, or bull?

Reply #18 –
Quote from: Moguta on
It seems that the Best Compression mode resamples to 24KHz and encodes with a VBR bitrate ~60Kbps, while Best Quality resamples to 32KHz and encodes ~100Kbps.  So much for their purported "proprietary patented adaptive variable bit rate compression encoding", especially if they're using LAME!  I wonder if anyone could re-create the exact settings they use, just for fun...


Looks like regular old -V9 for 'Best Compression' and maybe -V7 for 'Best Quality' to me...

ShrinkMyTunes: Good idea, or bull?

Reply #19 –
Quote from: Whelkman on
Quote from: Borbus on
I wonder if someone can take a look at the software if they can get hold of it and see if they actually are only dynamically linking to the LAME library or if they have created a derivative.

The existence of lame_enc.dll would satisfy the first part of your query, while a binary comparison would determine the second. While there exist compilers galore, not many would be used to compile LAME, so few comparisons should be needed. Note that statically linking (bundling) LAME into the main binary or a wrapper is considered a derivative work (by the FSF) and would be in violation of the LGPL unless the software also provides linkable object files (or source code); the LGPL requires that relinking to a compatible library remain possible. Note that this requirement assists with possible infringement cases as the replacement LAME must run--else the interface has changed and is thus a derivative work.

Well we know that this application is a single executable file so that would suggest that it is, in fact, a derivative work. However, could they have used something like PECompact so that it is really a proprietary executable and lame_enc.dll compressed into one executable? I think that would actually be considered dynamically linking, right?

Either way, the company MUST freely offer source code for LAME on their site and distribute the LGPL with their software. They don't offer source code and I highly doubt that they distribute the LGPL with their software.

ShrinkMyTunes: Good idea, or bull?

Reply #20 –
Quote from: Borbus on
However, could they have used something like PECompact so that it is really a proprietary executable and lame_enc.dll compressed into one executable? I think that would actually be considered dynamically linking, right?

Yes, though I believe they would need to make the [de]compression method known so that relinking would be possible. Otherwise it's an obfuscation wrapper, which is forbidden.

Quote from: Borbus on
Either way, the company MUST freely offer source code for LAME on their site and distribute the LGPL with their software. They don't offer source code and I highly doubt that they distribute the LGPL with their software.

To comply they would only need to distribute source to the people who received compiled code (i.e. customers) and only upon request, but, yes, LGPL text would be a requirement for inclusion with the distributed binary. Removing copyright notices is illegal regardless of license (with the exception of public domain release text).

ShrinkMyTunes: Good idea, or bull?

Reply #21 –
Quote from: joeshrubbery on
Quote from: Moguta on

It seems that the Best Compression mode resamples to 24KHz and encodes with a VBR bitrate ~60Kbps, while Best Quality resamples to 32KHz and encodes ~100Kbps.  So much for their purported "proprietary patented adaptive variable bit rate compression encoding", especially if they're using LAME!  I wonder if anyone could re-create the exact settings they use, just for fun...


Looks like regular old -V9 for 'Best Compression' and maybe -V7 for 'Best Quality' to me...


When I heard about this software it annoyed me. So today I whipped up a free replacement. The interface could do with some work, and there may be some small bugs, but it should work fine.

Select files. Shrink them. Files will be in a "Shrunk" folder in the folder of whatever files you selected.
You need .NET Framework v2.0 from Microsoft (free - most will already have it).

http://download.videohelp.com/favc/FreelyShrinkMyTunes.exe (Self-Extracting Archive; Unzip then run the exe in the folder.)

I made it better than their app by also supporting mp4, wma, flac, etc, which they don't do because they're clearly dumb.

(My app uses Lame at -V7 and -V9)

ShrinkMyTunes: Good idea, or bull?

Reply #22 –
.
Nice work, cool, tricky and all that but - I don't get it...  Why, in a world of portable MP3 players sporting 20, 40, 60 or 80+ gigabytes of storage would anyone want to shrink down their files even further than what MP3 does in the first place?  Hell, even a lowly 2 gigabyte capacity portable will hold more than enough 5 megabyte songs to keep anyone happy for many, many hours.  I did a little sleuthing in my collection and kept selecting 192Kb/s tracks until the sum total was exactly two gigabytes – it turns out that it takes 406 (high bitrate), MP3s to total 2 gigabytes in the real world of a wildly varying collection of music styles.  I don’t know about you, but I could live with 406 of my favorite tracks quite comfortably for many weeks before I’d wipe and reload more!  With the crazy speed of USB 2.0, what does it take, a few minutes…?

Extrapolate this little example into 20, 40, 60 etc., gig’ capacity drives and - what the hell!?  This is a utility looking for a problem... that doesn't exist!

Andrew D.
www.cdnav.com

.

ShrinkMyTunes: Good idea, or bull?

Reply #23 –
Regarding Lame inclusion, if they do not modify Lame, I am fine as long as they make it clear that they are using Lame (or use the "mp3dev.org" alias).

To me, this is sufficient, and I don't care about source code availability from them, or static vs dynamic linking.

(of course, that's only my own personal opinion)

ShrinkMyTunes: Good idea, or bull?

Reply #24 –
Quote from: Gabriel on
Regarding Lame inclusion, if they do not modify Lame, I am fine as long as they make it clear that they are using Lame (or use the "mp3dev.org" alias).

To me, this is sufficient, and I don't care about source code availability from them, or static vs dynamic linking.

(of course, that's only my own personal opinion)

Even though they are not including the full text of the LGPL and thus not letting people know about their rights to the LAME source code (whether they provide it or not).