Rockbox: Clip Zip

[Marvin the Martian]

My "archive" Zip has 32.6GB, which according to foobar is 4,190 files.....and I still have about 4.2GB free. What do you use the runtime information for?

[CrowdAds]

[saratoga]

I haven't tried with an SD card, but I think the auto-update option should work better then manually updating or initializing the database.

[okiru880]

Originally Posted by keyb_gr

Concerning this...

This kind of complaint has appeared ever since the ClipV1 port. Just about all the issues that might deteriorate playback have been fixed though - pitch is pretty much dead on, output mixer AGC is off, headphone amp supply is same as on OF. RB in fact does a better job when it comes to recovering peaks exceeding fullscale

I've done a little bit of investigation work, as the sound of the OF (EQ off/normal) and Rockbox (EQ off) sound completely different to me....

I recorded 10 secs of music, same track, same part, both alligned in Audacity.

Both waveforms are pretty much the same..
Both Spectrum graphs are pretty much the same...
The frequency responses in RMAA are very close...

So.. I would have expected that if both recordings were played back at the same time (synced up in Audacity) would produce a roughly similar sound (compounded together) however both played back at exactly the same time produces a strange "phasing sound", drum beat the same, but it sounds as if the channels have been "inverted", it gives a quite unnatural sound.

For a time I've been toying with the idea that the O.F sound has built in 'compression'/'hard limiting' to give an overall warmer sound, but now I'm thrown off even more due to the above test and the incompatibility of both waveforms to be played back at the same time and giving a natural sound.

[Marvin the Martian]

Originally Posted by okiru880

I've done a little bit of investigation work, as the sound of the OF (EQ off/normal) and Rockbox (EQ off) sound completely different to me....

I recorded 10 secs of music, same track, same part, both alligned in Audacity.

Both waveforms are pretty much the same..
Both Spectrum graphs are pretty much the same...
The frequency responses in RMAA are very close...

So.. I would have expected that if both recordings were played back at the same time (synced up in Audacity) would produce a roughly similar sound (compounded together) however both played back at exactly the same time produces a strange "phasing sound", drum beat the same, but it sounds as if the channels have been "inverted", it gives a quite unnatural sound.

For a time I've been toying with the idea that the O.F sound has built in 'compression'/'hard limiting' to give an overall warmer sound, but now I'm thrown off even more due to the above test and the incompatibility of both waveforms to be played back at the same time and giving a natural sound.

What player are you doing this on? The Clip+ OF plays back at a speed where the pitch is ever-so-slightly off, compared to Rockbox....yet when I got a Zip, it seemed to match up with the rockboxed Clip+ in its OF just fine.

[saratoga]

Originally Posted by okiru880

I've done a little bit of investigation work, as the sound of the OF (EQ off/normal) and Rockbox (EQ off) sound completely different to me....

I recorded 10 secs of music, same track, same part, both alligned in Audacity.

Both waveforms are pretty much the same..
Both Spectrum graphs are pretty much the same...
The frequency responses in RMAA are very close...

Hence they are the same

Originally Posted by okiru880

So.. I would have expected that if both recordings were played back at the same time (synced up in Audacity) would produce a roughly similar sound (compounded together) however both played back at exactly the same time produces a strange "phasing sound", drum beat the same, but it sounds as if the channels have been "inverted", it gives a quite unnatural sound.

Thats probably beating since rockbox uses a more accurate pitch then the OF. They'd only combine if the pitch was equal (and even then probably not perfectly since they'll be drift between recordings).

Originally Posted by okiru880

For a time I've been toying with the idea that the O.F sound has built in 'compression'/'hard limiting' to give an overall warmer sound, but now I'm thrown off even more due to the above test and the incompatibility of both waveforms to be played back at the same time and giving a natural sound.

Yeah, equal RMAA results excludes the possibility of any DSP effects.

Congrats, you've just learned that your hearing is as bad as everyone else's.

[okiru880]

Originally Posted by saratoga

Yeah, equal RMAA results excludes the possibility of any DSP effects.

I haven't tested this analytically .... however, soundwise - applying a 3db compressor in rockbox without EQ = the O.F without EQ.

When I get time I'll run a test. However, in the meantime, is this likely? Do you think that the Clip Zip could possibly have a little compression on by default?

[saratoga]

Originally Posted by okiru880

When I get time I'll run a test. However, in the meantime, is this likely? Do you think that the Clip Zip could possibly have a little compression on by default?

No, as I tried to make clear above, its literally impossible.

[okiru880]

Originally Posted by saratoga

No, as I tried to make clear above, its literally impossible.

It's not impossible. If this is an audible difference heard by a number of people and the graphs are negligible, then it's definitley a possibility.

[saratoga]

Originally Posted by okiru880

It's not impossible.

I'm not trying to talk down to you, but do you know what compression is exactly? Because it kind of sounds like you're not really aware of what happens when one applies compression given this reply.

Perhaps you should try applying a few dB of compression to a waveform and see what changes exactly before you tell me what is possible ...

Originally Posted by okiru880

If this is an audible difference heard by a number of people and the graphs are negligible, then it's definitley a possibility.

People hear differences all the time. Its pretty rare that they actually exist. I always expect a number of people to hear an audible difference. I also expect RMAA to confirm that they're hearing their own brain talking to them

Thats why RMAA is so great. Its 100x more sensitive then the best ears, and unlike the average person, it doesn't imagine things.

[okiru880]

Originally Posted by saratoga

I'm not trying to talk down to you, but do you know what compression is exactly? Because it kind of sounds like you're not really aware of what happens when one applies compression given this reply.

I know exactly what compression is, hence why I also used the term 'hard limiter' further up in this thread. And yes, admittedly you have used an ever so slightly derogatory tone in your replies.

Nevertheless, I do appreciate your contribution.

[dfkt]

Here's an audio track, recorded in Rockbox and in the stock firmware of the Zip - two FLACs, 18MB. Same volume level, beginning and end are trimmed at the exactly same sample - however running times aren't the same, since the clocks don't run at exactly the same speeds. File was a LAME -v0 initially, I used a 16 Ohm load on the output. Recorded with an Echo Audiofire 4 over ASIO. Personally, I couldn't make out any difference ABXing the two - http://www.mediafire.com/?dlfkfxb98ejd6y8

[saratoga]

Originally Posted by okiru880

I know exactly what compression is, hence why I also used the term 'hard limiter'

Hard limiting is not the same as compression. Hard limiting is basically just clipping, which all mp3 players do to levels above 1.0. You chop off levels above some limit, usually whatever your DAC can handle. Compression on the other hand distorts a signal so that more of its values are near maximum volume. This is done by adding energy at harmonics of a frequency to make it seem louder.

So for example, if you have a tone at 1 kHz and peak normalized and apply a few dB of compression, you'll now have a tone at 1kHz in addition to knew tones at its harmonics (2khz, 3khz, etc.).

For this reason, its quite obvious that no compression is applied. We would see it in those RMAA plots as new frequencies being generated. Perhaps you could take a few minutes to try it yourself and see what a compressor actually does to the frequency spectrum of a tone. The effect is actually pretty easy to spot.

Originally Posted by okiru880

And yes, you have been talking down to me given your ever so slightly derogatory tone in your replies.

I'm sorry you're taking offense, but I'm just trying to explain to you why you're wrong.

[saratoga]

Rather then focus on why we know that no compression was applied, perhaps its more interesting to think about why you thought compression was applied in the first place:

Originally Posted by okiru880

soundwise - applying a 3db compressor in rockbox without EQ = the O.F without EQ.

Of course, the purpose of compression is to make something sound louder. Obviously in order to establish that you were hearing compression, you would first have to rule out differences in actual loudness (e.g. volume). So for a test like this you would have had to measure the actual output level very precisely of both devices, and then set them equal with an extremely high degree of precision. What software did you use for the measurement and equalization? How accurate do you believe it was?

Is it possible you didn't achieve sufficiently accurate volume matching and the change in loudness you heard was due to volume rather then compression?

[dfkt]

The OF is about 3dB louder at full volume than Rockbox at 0dB, indeed.

[skip252]

There's a difference in the volume of the first drum hit in the samples.

foo_abx 1.3.4 report
foobar2000 v1.1.10
2012/01/31 18:22:50

File A: C:\Users\Skip252\Desktop\Eardrum\zip\zip-of.flac
File B: C:\Users\Skip252\Desktop\Eardrum\zip\zip-rb.flac

18:22:50 : Test started.
18:23:04 : 01/01  50.0%
18:23:18 : 02/02  25.0%
18:23:29 : 03/03  12.5%
18:23:37 : 04/04  6.3%
18:23:52 : 05/05  3.1%
18:24:03 : 06/06  1.6%
18:24:18 : 07/07  0.8%
18:24:26 : 08/08  0.4%
18:24:38 : 09/09  0.2%
18:24:47 : 10/10  0.1%
18:24:56 : 11/11  0.0%
18:24:58 : Test finished.

When I set the start point to 10 seconds from the beginning.

foo_abx 1.3.4 report
foobar2000 v1.1.10
2012/01/31 18:56:34

File A: C:\Users\Skip252\Desktop\Eardrum\zip\zip-of.flac
File B: C:\Users\Skip252\Desktop\Eardrum\zip\zip-rb.flac

18:56:34 : Test started.
18:57:03 : 00/01  100.0%
18:57:26 : 00/02  100.0%
18:57:46 : 01/03  87.5%
18:58:22 : 01/04  93.8%
18:58:40 : 01/05  96.9%
18:58:55 : 01/06  98.4%
18:59:56 : 01/07  99.2%
19:00:34 : 02/08  96.5%
19:01:11 : 03/09  91.0%
19:01:43 : 04/10  82.8%
19:02:20 : 05/11  72.6%
19:02:29 : Test finished.

The volume difference in the drum hit goes away and I can't ABX the difference any more. I'm guessing that's due to the differences in the timing?

[Tyrone F. Horneigh]

Originally Posted by Marvin the Martian

My "archive" Zip has 32.6GB, which according to foobar is 4,190 files.....and I still have about 4.2GB free. What do you use the runtime information for?

For what I understand its intended purpose to be--finding out which songs I am "manually" enqueuing. I think I know what my favorites are, but I'm curious to see some objective data.

[earlpianokeys]

I just installed rockbox in my clip zip, extracted .rockbox in my CZ root directory/folder as well as the bootloader. Firmware upgrade gets completed and CZ shuts down. After turning it own, I see that rockbox logo and it seems like it's loading but then after booting, the screen just turns all white, and blinks soon as I press a button... I uninstalled/installed a couple of times but it still doesn't work and I still get that white screen! HELP!!!

[bertrik]

earlpianokeys, sorry to hear it doesn't work for you. To help the rockbox team diagnose the problem, could you create an entry in our bugtracker?

It's a bit weird problem, because apparently the rockbox bootloader is able to display a logo, while the main firmware shows a white screen.

One thing that would help is to identify the exact hardware configuration of your clip zip. You can find that out by flashing the original .bin file, but renamed to have a 't' at the end of the filename, like clpzt.bin instead clpza.bin. This should give you an extra 'diagnosis' menu in the systems settings menu. Please report the info on the first two diagnosis screens.

[TexasRockbox]

Originally Posted by earlpianokeys

I just installed rockbox in my clip zip, extracted .rockbox in my CZ root directory/folder as well as the bootloader. Firmware upgrade gets completed and CZ shuts down. After turning it own, I see that rockbox logo and it seems like it's loading but then after booting, the screen just turns all white, and blinks soon as I press a button... I uninstalled/installed a couple of times but it still doesn't work and I still get that white screen! HELP!!!

Try turning off dircache.

[saratoga]

A new rbutil release with support for the Clip Zip should be out in a day or two (basically whenever they're uploaded). It also has a neat new feature: it can read sandisk firmware zips directly. So you don't even have to download the sandisk firmware, just copy and paste the link to the zip on the sandisk website and it downloads it, unzips, patches and then updates the player.