I have read 100s of article about the OOM problem. Most are in regard to large bitmaps. I am doing a mapping application where we download 256x256 weather overlay tiles. Most are totally transparent and very small. I just got a crash on a bitmap stream that was 442 Bytes long while calling BitmapFactory.decodeByteArray(....).
The Exception states:
java.lang.OutOfMemoryError: bitmap size exceeds VM budget(Heap Size=9415KB, Allocated=5192KB, Bitmap Size=23671KB)
The code is:
protected Bitmap retrieveImageData() throws IOException {
URL url = new URL(imageUrl);
InputStream in = null;
OutputStream out = null;
HttpURLConnection connection = (HttpURLConnection) url.openConnection();
// determine the image size and allocate a buffer
int fileSize = connection.getContentLength();
if (fileSize < 0) {
return null;
}
byte[] imageData = new byte[fileSize];
// download the file
//Log.d(LOG_TAG, "fetching image " + imageUrl + " (" + fileSize + ")");
BufferedInputStream istream = new BufferedInputStream(connection.getInputStream());
int bytesRead = 0;
int offset = 0;
while (bytesRead != -1 && offset < fileSize) {
bytesRead = istream.read(imageData, offset, fileSize - offset);
offset += bytesRead;
}
// clean up
istream.close();
connection.disconnect();
Bitmap bitmap = null;
try {
bitmap = BitmapFactory.decodeByteArray(imageData, 0, bytesRead);
} catch (OutOfMemoryError e) {
Log.e("Map", "Tile Loader (241) Out Of Memory Error " + e.getLocalizedMessage());
System.gc();
}
return bitmap;
}
Here is what I see in the debugger:
bytesRead = 442
So the Bitmap data is 442 Bytes. Why would it be trying to create a 23671KB Bitmap and running out of memory?
I have run into problems like this in the past. Android uses Bitmap VM and it is very small. Make sure you dispose your bitmap via bmp.recycle. Later versions of Android have more Bitmap VM but the version that I've been dealing with has a 20MB limit.
This may work. Shrinking the bitmaps to lesser quality. I am not sure, but this may duplicate the image in memory, but could easily be worth a shot.
Bitmap image;
image = BitmapFactory.decodeByteArray(data, 0, data.length);
Bitmap mutableBitmap = image.copy(Bitmap.Config.ARGB_4444, true);
Canvas canvas = new Canvas(mutableBitmap);
My old answer below I don't think will work streaming.
Options options = new BitmapFactory.Options();
Options options2 = new BitmapFactory.Options();
Options options3 = new BitmapFactory.Options();
options.inPreferredConfig = Bitmap.Config.ARGB_8888;///////includes alpha
options2.inPreferredConfig = Bitmap.Config.RGB_565 ;///////no alpha
options3.inPreferredConfig = Bitmap.Config.ARGB_4444 ;/////alpha lesser quality
image=BitmapFactory.decodeResource(getResources(),R.drawable.imagename,options);
image=Bitmap.createScaledBitmap(image, widthx,height, true);
It sounds like you've already done some reading on this subject, so I'll spare you the standard 'this is how you decode a bitmap' comments..
It jumps out at me that you're perhaps holding on to a reference of old bitmaps (perhaps the tile has moved off screen, but you still have a reference in an array somewhere and as a result it isn't being garbage collected?). This has bitten me really badly in the past - memory leaks are hard to debug.
There's a great Google I/O video over here that really helped me when I was having similar problems. It's around an hour, but will hopefully save you days later on.
It covers things like:
Creating heap dumps
Heap usage in DDMS
Using MAT to compare/analyze heap dumps
Related
I have an URI image file, and I want to reduce its size to upload it. Initial image file size depends from mobile to mobile (can be 2MB as can be 500KB), but I want final size to be about 200KB, so that I can upload it.
From what I read, I have (at least) 2 choices:
Using BitmapFactory.Options.inSampleSize, to subsample original image and get a smaller image;
Using Bitmap.compress to compress the image specifying compression quality.
What's the best choice?
I was thinking to initially resize image width/height until width or height is above 1000px (something like 1024x768 or others), then compress image with decreasing quality until file size is above 200KB. Here's an example:
int MAX_IMAGE_SIZE = 200 * 1024; // max final file size
Bitmap bmpPic = BitmapFactory.decodeFile(fileUri.getPath());
if ((bmpPic.getWidth() >= 1024) && (bmpPic.getHeight() >= 1024)) {
BitmapFactory.Options bmpOptions = new BitmapFactory.Options();
bmpOptions.inSampleSize = 1;
while ((bmpPic.getWidth() >= 1024) && (bmpPic.getHeight() >= 1024)) {
bmpOptions.inSampleSize++;
bmpPic = BitmapFactory.decodeFile(fileUri.getPath(), bmpOptions);
}
Log.d(TAG, "Resize: " + bmpOptions.inSampleSize);
}
int compressQuality = 104; // quality decreasing by 5 every loop. (start from 99)
int streamLength = MAX_IMAGE_SIZE;
while (streamLength >= MAX_IMAGE_SIZE) {
ByteArrayOutputStream bmpStream = new ByteArrayOutputStream();
compressQuality -= 5;
Log.d(TAG, "Quality: " + compressQuality);
bmpPic.compress(Bitmap.CompressFormat.JPEG, compressQuality, bmpStream);
byte[] bmpPicByteArray = bmpStream.toByteArray();
streamLength = bmpPicByteArray.length;
Log.d(TAG, "Size: " + streamLength);
}
try {
FileOutputStream bmpFile = new FileOutputStream(finalPath);
bmpPic.compress(Bitmap.CompressFormat.JPEG, compressQuality, bmpFile);
bmpFile.flush();
bmpFile.close();
} catch (Exception e) {
Log.e(TAG, "Error on saving file");
}
Is there a better way to do it? Should I try to keep using all 2 methods or only one? Thanks
Using Bitmap.compress() you just specify compression algorithm and by the way compression operation takes rather big amount of time. If you need to play with sizes for reducing memory allocation for your image, you exactly need to use scaling of your image using Bitmap.Options, computing bitmap bounds at first and then decoding it to your specified size.
The best sample that I found on StackOverflow is this one.
Most of the answers i found were just pieces that i had to put together to get a working code, which is posted below
public void compressBitmap(File file, int sampleSize, int quality) {
try {
BitmapFactory.Options options = new BitmapFactory.Options();
options.inSampleSize = sampleSize;
FileInputStream inputStream = new FileInputStream(file);
Bitmap selectedBitmap = BitmapFactory.decodeStream(inputStream, null, options);
inputStream.close();
FileOutputStream outputStream = new FileOutputStream("location to save");
selectedBitmap.compress(Bitmap.CompressFormat.JPEG, quality, outputStream);
outputStream.close();
long lengthInKb = photo.length() / 1024; //in kb
if (lengthInKb > SIZE_LIMIT) {
compressBitmap(file, (sampleSize*2), (quality/4));
}
selectedBitmap.recycle();
} catch (Exception e) {
e.printStackTrace();
}
}
2 parameters sampleSize and quality plays an important role
sampleSize is used to subsample the original image and return a smaller image, ie
SampleSize == 4 returns an image that is 1/4 the width/height of the original.
quality is used to hint the compressor, input range is between 0-100. 0 meaning compress for
small size, 100 meaning compress for max quality
BitmapFactory.Options - Reduces Image Size (In Memory)
Bitmap.compress() - Reduces Image Size (In Disk)
Refer to this link for more information about using both of them:
https://android.jlelse.eu/loading-large-bitmaps-efficiently-in-android-66826cd4ad53
Here is my code:
File file = new File(Path to Jpeg File size is 700kb);
InputStream in = null;
try {
in = new BufferedInputStream(new FileInputStream(file));
}
catch (Exception e) {
// TODO: handle exception
}
bitmap =BitmapFactory.decodeStream(in);
bitmap = bitmap.copy(Bitmap.Config.ARGB_8888, true);
Please help i get error in this copy line i want to make its ARGB_8888 image.Need Help :(
You need to reduce the memory usage.
From you code, you first decode stream to one bitmap, and then copy it, which means you create two large bitmap objects.
You don't need to decode and then copy it, you can try
BitmapFactory.Options options = new BitmapFactory.Options();
options.inPreferredConfig = Bitmap.Config.ARGB_8888
// You can try value larger than 1
options.inSampleSize = 2 // If set to a value > 1, requests the decoder to subsample the original image, returning a smaller image to save memory.
// Decode bitmap
bitmap = BitmapFactory.decodeStream(in, null, options)
In this case, there's only one bitmap created. And you set inSampleSize to large values to reduce the loaded bitmap size.
I am trying to display a jpg file from a server into an imageView. When I try to load a smaller image (300x400), there are no problems. But when I try to load a full size picture (2336x3504), the image will not load. The file size of the image is only 2mb. I do not get any errors in logcat and there are no exceptions thrown. It simply won't load the image. I also tried using this:
BitmapFactory.Options options=new BitmapFactory.Options();
options.inSampleSize = 8;
Bitmap preview_bitmap=BitmapFactory.decodeStream(is,null,options);
This doesn't do anything to help load the large files, but it does resize the smaller image (like it is suppose to). I did add the large picture to my resources and tested it as if it was embedded in the app and it worked fine, just won't work on the server. I have been working all day on this and can't seem to figure out how to load these large pictures. Can anyone help me out with this? Thanks for any info.
Here is the link where I found the above code and have been playing with the other examples but still not getting it to work.
EDIT:
Here is the code I'm using, to load the image:
public static Bitmap getBitmapFromURL(String src) {
Bitmap bmImg;
URL myFileUrl = null;
try {
myFileUrl = new URL(src);
HttpURLConnection conn= (HttpURLConnection)myFileUrl.openConnection();
conn.setDoInput(true);
conn.connect();
InputStream is = conn.getInputStream();
BitmapFactory.Options options=new BitmapFactory.Options();
options.inSampleSize = 16;
bmImg = BitmapFactory.decodeStream(is, null, options);
return bmImg;
} catch (Exception e) {
// TODO Auto-generated catch block
Log.d("Error", e.toString());
return null;
}
}
Here is the logcat screenshot (couldn't figure out how to copy the text appropriately in eclipse) I cleared the log right before I hit the button to load the image. So all you see is what happens when I hit that button. I erased the company and app names (where you see "com.", assume its "com.mycompany.myapp".
It is not uncommon for BitmapFactory.decodeFromStream() to give up and just return null when you connect it directly to the InputStream of a remote connection. Internally, if you did not provide a BufferedInputStream to the method, it will wrap the supplied stream in one with a buffer size of 16384. One option that sometimes works is to pass a BufferedInputStream with a larger buffer size like:
BufferedInputStream bis = new BufferedInputStream(is, 32 * 1024);
A more universally effective method is to download the file completely first, and then decode the data like this:
InputStream is = connection.getInputStream();
BufferedInputStream bis = new BufferedInputStream(is, 8190);
ByteArrayBuffer baf = new ByteArrayBuffer(50);
int current = 0;
while ((current = bis.read()) != -1) {
baf.append((byte)current);
}
byte[] imageData = baf.toByteArray();
BitmapFactory.decodeByteArray(imageData, 0, imageData.length);
FYI, the buffer sizes in this example are somewhat arbitrary. As has been said in other answers, it's a fantastic idea not to keep an image that size in memory longer than you have to. You might consider writing it directly to a file and displaying a downsampled version.
Hope that helps!
Devunwired's answer is right but out of memory error can occur if image size is too large, in that case we will have to scale down image, here is the code to scale down image after DevunWired's download image code
final BitmapFactory.Options options = new BitmapFactory.Options();
BufferedInputStream bis = new BufferedInputStream(is, 4*1024);
ByteArrayBuffer baf = new ByteArrayBuffer(50);
int current = 0;
while ((current = bis.read()) != -1) {
baf.append((byte)current);
}
byte[] imageData = baf.toByteArray();
BitmapFactory.decodeByteArray(imageData, 0, imageData.length, options);
options.inJustDecodeBounds = true;
options.inSampleSize = 2; //calculateInSampleSize(options, 128, 128);
options.inJustDecodeBounds = false;
Bitmap bitmap = BitmapFactory.decodeByteArray(imageData, 0, imageData.length, options);
Does it silently fail, or does it throw an exception or OutOfMemory error? Btw, if a jpeg is 2MB that doesn't mean it'll take up 2MB of memory. 2MB is the compressed size, and since Android is working with a Bitmap, the 2336 x 3504 will take up approximately 2336 x 3504 x 4 bytes in memory. (2336 x 3504 x 4 = 32,741,376). Downsampling 8 times still might not be enough, especially if you have other bitmaps in memory at the time.
In my application I load a couple of images from JPEG and PNG files. When I place all those files into assets directory and load it in this way, everything is ok:
InputStream stream = getAssets().open(path);
Bitmap bitmap = BitmapFactory.decodeStream(stream, null, null);
stream.close();
return new BitmapDrawable(bitmap);
But when I try to load the exact same images from sd card, I get an OutOfMemory exception!
InputStream stream = new FileInputStream("/mnt/sdcard/mydata/" + path);
Bitmap bitmap = BitmapFactory.decodeStream(stream, null, null);
stream.close();
return new BitmapDrawable(bitmap);
This is what I get in the log:
11-05 00:53:31.003: ERROR/dalvikvm-heap(13183): 827200-byte external allocation too large for this process.
11-05 00:53:31.003: ERROR/GraphicsJNI(13183): VM won't let us allocate 827200 bytes
...
11-05 00:53:31.053: ERROR/AndroidRuntime(13183): Caused by: java.lang.OutOfMemoryError: bitmap size exceeds VM budget
11-05 00:53:31.053: ERROR/AndroidRuntime(13183): at android.graphics.BitmapFactory.nativeDecodeStream(Native Method)
...
Why can this happen?
UPDATE: Tried both of these on real device - it seems that I can't load more than 12MB of bitmaps into whatever is called "external memory" (this is not an sd card).
I tried all the approaches mentioned here & at other resources but I came to the inference that setting ImageView's reference to null will solve the issue:
public Bitmap getimage(String path ,ImageView iv)
{
//iv is passed to set it null to remove it from external memory
iv=null;
InputStream stream = new FileInputStream("/mnt/sdcard/mydata/" + path);
Bitmap bitmap = BitmapFactory.decodeStream(stream, null, null);
stream.close();
stream=null;
return bitmap;
}
& you are done!
Note:Though it may solve above problem but I would suggest you to check Tom van Zummeren 's optimized image loading.
And also check SoftReference: All SoftReferences pointing to softly reachable objects are guaranteed to be cleared before the VM will throw an OutOfMemoryError.
When doing a lot with bitmaps, don't debug the app - just run it. The debugger will leave memory leaks.
Bitmaps are very expensive. If possible, scale them down on load by creating BitmapFactory.Options and setting inSampleSize to >1.
EDIT: Also, be sure to check your app for memory leaks. Leaking a Bitmap (having static Bitmaps is an excellent way to do that) will quickly exhaust your available memory.
Probably nothing wrong with your API usage, I guess all we can do is infer that using the AssetManager involves less behind-the-scenes heap allocation than opening a random file from the SD card.
800KB is a serious allocation in anybody's book... this will doubtless be for the decompressed image pixels. Given that you know the size of the image, what depth is it? If it's 32bpp then try overriding that using inPreferredConfig.
This is a fairly common issue which all of us face while loading images from the sdcard.
The solution as I found was to use inJustDecodeBounds first while loading the image using decodeFileDescriptor . That would not actually decode the image, but give the image size. Now I can scale it appropriately(using the options) so as to resize the image for the display area. Its needed because low memory on the phone can be easily taken over by your 5MP image. This I believe is the most elegant solution.
There are two issues here....
Bitmap memory isn't in the VM heap but rather in the native heap - see BitmapFactory OOM driving me nuts
Garbage collection for the native heap is lazier than the VM heap - so you need to be quite aggressive about doing bitmap.recycle and bitmap =null every time you go through an Activity's onPause or onDestroy
Instead of loading it from the SD Card directly, why not move the image to the cache in the phone's internal storage using getCacheDir() or use a temp directory to store the images in?
See this, this on external memory usage. Also, this article may be of relevance to you.
Use the below code and you will never get the following error: java.lang.OutOfMemoryError: bitmap size exceeds VM budget
BitmapFactory.Options bounds = new BitmapFactory.Options();
bounds.inSampleSize = 4;
myBitmap = BitmapFactory.decodeFile(imgFile.getAbsolutePath(), bounds);
picturesView.setImageBitmap(myBitmap);
The best solution i found and edited according to my need
public static Bitmap getImageBitmap(String path) throws IOException{
// Allocate files and objects outside of timingoops
File file = new File(thumbpath);
RandomAccessFile in = new RandomAccessFile(file, "rws");
final FileChannel channel = in.getChannel();
final int fileSize = (int)channel.size();
final byte[] testBytes = new byte[fileSize];
final ByteBuffer buff = ByteBuffer.allocate(fileSize);
final byte[] buffArray = buff.array();
#SuppressWarnings("unused")
final int buffBase = buff.arrayOffset();
// Read from channel into buffer, and batch read from buffer to byte array;
long time1 = System.currentTimeMillis();
channel.position(0);
channel.read(buff);
buff.flip();
buff.get(testBytes);
long time1 = System.currentTimeMillis();
Bitmap bmp = Bitmap_process(buffArray);
long time2 = System.currentTimeMillis();
System.out.println("Time taken to load: " + (time2 - time1) + "ms");
return bmp;
}
public static Bitmap Bitmap_process(byte[] buffArray){
BitmapFactory.Options options = new BitmapFactory.Options();
options.inDither=false; //Disable Dithering mode
options.inPurgeable=true; //Tell to gc that whether it needs free memory, the Bitmap can be cleared
options.inInputShareable=true; //Which kind of reference will be used to recover the Bitmap data after being clear, when it will be used in the future
options.inTempStorage=new byte[32 * 1024]; //Allocate some temporal memory for decoding
options.inSampleSize=1;
Bitmap imageBitmap = BitmapFactory.decodeByteArray(buffArray, 0, buffArray.length, options);
return imageBitmap;
}
Thanks to all the threads, I've found a solution that works for me on a real device.
The tricks are all about using
BitmapFactory.Options opts=new BitmapFactory.Options();
opts.inSampleSize=(int)(target_size/bitmap_size); //if original bitmap is bigger
But for me this was not enough. My original image (taken from the Camera app) was 3264x2448. The correct ratio for me was 3, since i wanted a plain VGA image of 1024x768.
But setting inSampleSize to 3 was not enough: still out of memory exception.
So in the end I opted for a iterative approach: I start from the computed correct size, and increase it until I stop having a OOM exception.
For me it was at sample of 4.
// Decode with inSampleSize
BitmapFactory.Options o2 = new BitmapFactory.Options();
// o2.inSampleSize = scale;
float trueScale = o.outWidth / 1024;
o2.inPurgeable = true;
o2.inDither = false;
Bitmap b = null;
do {
o2.inSampleSize = (int) trueScale;
Log.d(TAG, "Scale is " + trueScale);
try {
b = BitmapFactory.decodeStream(new FileInputStream(f), null, o2);
} catch (OutOfMemoryError e) {
Log.e(TAG,"Error decoding image at sampling "+trueScale+", resampling.."+e);
System.gc();
try {
Thread.sleep(50);
} catch (InterruptedException e1) {
e1.printStackTrace();
}
}
trueScale += 1;
} while (b==null && trueScale < 10);
return b;
You must not depends on the GC to recycle your bitmap memory.
You must clearly recycle the bitmap when it is not needed.
See the Bitmap method:
void recycle()
Free up the memory associated with this bitmap's pixels, and mark the bitmap as "dead", meaning it will throw an exception if getPixels() or setPixels() is called, and will draw nothing.
Try this another way...
Bitmap bmpOrignal = BitmapFactory.decodeFile("/sdcard/mydata/" + path");
Allows inSampleSize resize the final read image.
getLength() of AssetFileDescriptor allows get size of file.
You can vary inSampleSize according to getLength() to prevent OutOfMemory like this :
private final int MAX_SIZE = 500000;
public Bitmap readBitmap(Uri selectedImage)
{
Bitmap bm = null;
AssetFileDescriptor fileDescriptor = null;
try
{
fileDescriptor = this.getContentResolver().openAssetFileDescriptor(selectedImage,"r");
long size = fileDescriptor.getLength();
BitmapFactory.Options options = new BitmapFactory.Options();
options.inSampleSize = (int) (size / MAX_SIZE);
bm = BitmapFactory.decodeFileDescriptor(fileDescriptor.getFileDescriptor(), null, options);
}
catch (Exception e)
{
e.printStackTrace();
}
finally
{
try {
if(fileDescriptor != null) fileDescriptor.close();
} catch (IOException e) {}
}
return bm;
}
I am working on an Android app that displays photos which are downloaded from Flickr. I obtain a bitmap object from a byte array, which in turn is read from the relevant Flickr URL, as follows:
BitmapFactory.Options opt = new BitmapFactory.Options();
opt.inDither = true;
opt.inPreferredConfig = Bitmap.Config.ARGB_8888;
Bitmap bitmap = BitmapFactory.decodeByteArray(data, 0, data.length, opt);
I then draw the bitmap onto a canvas in the onDraw method of a View object:
Paint paint = new Paint(Paint.ANTI_ALIAS_FLAG | Paint.FILTER_BITMAP_FLAG);
canvas.drawBitmap(bitmap, 0, 0, paint);
The problem is that the resulting picture is pixelated and I can't figure out why; I have tried a number of variations of the opt and paint objects with no luck. The difference between the picture displayed in my app and the picture at the original URL is roughly demonstrated by the following:
Bad image, see pixelation in top left corner http://homepages.inf.ed.ac.uk/s0677975/bad.jpg
Good picture, this is the expected result http://homepages.inf.ed.ac.uk/s0677975/good.jpg
Look e.g. at the clouds in the top-left corner to see the difference.
Note that JPEG pictures which are loaded from the project resources and drawn in a similar way display just fine, i.e. have no pixelation.
Can anybody give me a hint as to why this is happening?
To elaborate a little, the byte array is obtained from Flickr as follows; this is based on code from the Photostream app by Romain Guy:
InputStream in = new BufferedInputStream(url.openStream(), IO_BUFFER_SIZE);
final ByteArrayOutputStream dataStream = new ByteArrayOutputStream();
out = new BufferedOutputStream(dataStream, IO_BUFFER_SIZE);
copy(in, out);
out.flush();
final byte[] data = dataStream.toByteArray();
PS: I also posted a variant of this question on the android.developer Google group.
Thanks a lot for your suggestion -- now I am really puzzled! I did as you suggested and found that the image resulting directly from the downloaded byte array is indeed pixelated. However, this is downloaded from exactly the same URL which, when accessed on my computer, is NOT pixelated. Here is the corresponding Flickr URL:
http://farm3.static.flickr.com/2678/4315351421_54e8cdb8e5.jpg
Even stranger, when I run the same app in the simulator rather than on my phone (a HTC Hero), there is no pixelation.
How on earth is this possible?
Below is the code I use for loading a bitmap from a URL -- it is based on the Photostream app by Romain Guy, and it incorporates Will's suggestion to write the raw byte array to file:
Bitmap loadPhotoBitmap(URL url) {
Bitmap bitmap = null;
InputStream in = null;
BufferedOutputStream out = null;
try {
FileOutputStream fos = new FileOutputStream("/sdcard/photo-tmp.jpg");
BufferedOutputStream bfs = new BufferedOutputStream(fos);
in = new BufferedInputStream(url.openStream(),
IO_BUFFER_SIZE);
final ByteArrayOutputStream dataStream = new ByteArrayOutputStream();
out = new BufferedOutputStream(dataStream, IO_BUFFER_SIZE);
copy(in, out);
out.flush();
final byte[] data = dataStream.toByteArray();
bfs.write(data, 0, data.length);
bfs.flush();
BitmapFactory.Options opt = new BitmapFactory.Options();
bitmap = BitmapFactory.decodeByteArray(data, 0, data.length, opt);
} catch (IOException e) {
android.util.Log.e(LOG_TAG, "Could not load photo: " + this, e);
} finally {
closeStream(in);
closeStream(out)
closeStream(bfs);
}
return bitmap;
}
private static void copy(InputStream in, OutputStream out) throws IOException {
byte[] b = new byte[IO_BUFFER_SIZE];
int read;
while ((read = in.read(b)) != -1) {
out.write(b, 0, read);
}
}
private static void closeStream(Closeable stream) {
if (stream != null) {
try {
stream.close();
} catch (IOException e) {
android.util.Log.e(LOG_TAG, "Could not close stream", e);
}
}
}
Am I going crazy here?
Best,
Michael.
Ok, so I finally get it: it appears that my mobile network does image compression to save bandwidth.
Hence a picture downloaded from my phone is of lower quality than the same picture downloaded from my computer.
That's a bummer for my app, but I don't suppose there is anything I can do about it. Sigh.
Thanks again for your input though!
Best,
Michael.
Write the raw bytes fetched from the URL to /sdcard/tmp.jpg, and view on your PC.
JPEG images are compressed in 8x8 (or 16x16) tiles. The 'pixelation' as you describe it is actually in these tiles, suggesting that the 'bad' image is a JPEG that is more aggressively compressed than the other.
So I'd anticipate that the actual issue is that the image being downloaded is a very low-quality version, e.g. one intended for thumbnailing/preview use-cases.
Some version of Android have a bug in Bitmap class and convert the Bitmap to RGB_565 upon some operations. This would manifest itself in artifacts similar to those on your picture. This would also explain the banding of the blue sky.
Also, have in mind that android attempts to "optimize" image by converting them to rgb_565 upon loading and even compiling in resource files. Take a look at:
http://android.nakatome.net/2010/04/bitmap-basics.html