Working With 560p Source Material
Most people grab 560p content because it is smaller and faster to download, then complain it looks soft on a modern display. The resolution is roughly 1024 by 576 pixels for widescreen material. That is nowhere near 720p, 1080p, or anything labeled high definition. I have spent years processing this kind of source and the main realization is that you cannot create detail that was never recorded. What you can do is clean up what is there and stretch it without making it look worse than it already is. This guide covers the actual workflow for handling 560p files. I will skip the hype and talk about what works in practice. The process involves deinterlacing if needed, noise reduction, proper scaling, and sometimes AI-assisted enhancement. Each step has tradeoffs. The wrong settings make grain turn into plastic-looking smears. The right settings keep the image readable at a reasonable size. Before touching any processing software, check the file metadata. A lot of 560p sources are actually interlaced, which means they were originally broadcast or pulled from older encodes. Interlaced material has horizontal artifacts that look like comb lines during motion. If you scale interlaced content directly, those artifacts get larger and more obvious. Use a tool like MediaInfo to check the scan type. Look for interlaced, top field first, or bottom field first flags. If you see them, you need a deinterlacer before anything else.
Another common issue is telecine patterns. Some 560p sources come from film material that was converted to video using 3:2 pulldown. This causes uneven frame rates and judder when played back. A proper inverse telecine process reverses this and returns the frames to their original timing. Without it, motion looks choppy even if the resolution is fine.
Deinterlacing and Frame Rate Handling
I use TDeint or Srestore inavisynth scripts for this. TDeint with the ecdeint option handles most standard interlaced material. It is fast and does not require much tuning. For film sources, Srestore reconstructs the original frame rate by detecting and removing duplicate fields. This usually turns a 29.97 interlaced file into a clean 23.976 progressive file. Here is a practical scenario I ran into last year. A user uploaded a 560p rip of an older television movie that had been encoded from a DVD source with a bad telecine chain. The frame rate was stuck at 29.97fps but every fifth frame was duplicated incorrectly. Standard deinterlacing made it worse because it treated every field as separate. I identified the pattern by checking the vsf file and found the bad telecine cycle at approximately every 1000 frames. I used Srestore with a custom cycle length instead of the default detection. This fixed the judder completely and produced a stable 23.976fps source for scaling.
Get the Full Details
Noise Reduction Before Scaling
This step gets skipped too often. 560p sources usually carry compression artifacts from their original encode, plus analog noise if they came from a broadcast capture. Running denoising before upscaling matters because noise gets amplified during enlargement. If you scale first and denoise after, the artifact patterns become part of the new pixel structure and are harder to remove cleanly. For moderate noise, MVdustbunny with default settings works well on most material. For heavier compression blocks, DCTFilter or TemporalDegrain gives better results. The key is keeping the strength low. Aggressive denoising turns faces into wax figures. I usually aim for a result where fine textures like fabric and hair still show some variation. If the image looks painted, you have gone too far.
Scaling to a Higher Resolution
Upscaling 560p to 1080p requires a ratio of roughly 1.9x horizontally. Bicubic interpolation gives a clean result with minimal artifacts. Lancer is another good option and preserves sharper edges without excessive ringing. For material that needs more detail recovery, I occasionally use the DPIR model in Real-ESRGAN, but it is overkill for most cases and introduces hallucinated texture that looks wrong on close inspection. One thing beginners miss is that the output resolution should match your playback environment. If you are watching on a 1080p screen, scaling to 1920 by 1080 makes sense. If you plan to watch on a 4K display, scaling to 2160p is reasonable, but the quality ceiling is still determined by the source. Upscaling a 560p file to 4K does not give you 4K quality. It gives you a larger 560p file with some processing applied. Be honest about what the process can deliver.
Sharpening After Scaling
Scaling softens the image slightly, especially with bicubic methods. A light sharpen pass after upscaling restores edge definition without affecting the whole frame. UnsharpMask with a radius of 1.0 and amount around 0.3 to 0.5 works well. Do not push the amount above 0.6 or halos appear around high contrast edges. This is especially noticeable on text overlays and subtitle bars. I encountered a case where a user tried to compensate for softness by applying heavy sharpen before denoising. The result was a file full of amplified compression blocks that looked worse than the original. The correct order is always denoise, scale, then sharpen. Reversing that sequence compounds every artifact in the file.
Audio and Encoding Settings
Audio rarely gets addressed in these workflows but it matters just as much for the final experience. Many 560p releases come with poor AAC or AC3 tracks that sound thin or distorted. If the source audio is low bitrate, consider replacing it with a cleaner version if one exists. Otherwise, at least encode it at 192kbps AAC or higher to avoid adding compression noise on top of the existing track. For video encoding, x264 or x265 with a CRF between 18 and 22 produces a good balance of quality and file size. Lower CRF values save size but introduce encoding artifacts that counteract the cleanup work done earlier. Higher values waste space without visible improvement. I settle on CRF 20 for x264 as a default starting point and adjust based on scene complexity.
Realistic Quality Expectations
560p material has limits. No amount of processing will recover lost information. Fast motion sequences will always show more artifacts than static shots. Dark scenes with heavy noise benefit less from denoising because detail gets removed along with the grain. These are not workflow failures. They are physical constraints of the source format. If you frequently work with very low resolution material, consider whether a better source exists. Online releases, VHS rips, and broadcast captures are fundamentally different from DVD or Blu-ray sourced material. Knowing the lineage of your source helps you choose the right tools and set realistic quality goals. I stop trying to fix severely degraded sources after about ten minutes of processing and move on. The time investment does not produce a proportional quality gain.
Tools I Recommend
- AviSynth+ or VapourSynth for scripting the pipeline
- MediaInfo for source analysis
- TDeint for deinterlacing
- Srestore for inverse telecine
- DCTFilter or MVdustbunny for denoising
- Real-ESRGAN for AI enhancement when justified
- x264 or x265 for final encoding
The exact script varies by source, but the structure stays the same. Load, analyze, deinterlace if needed, denoise, scale, sharpen, encode. Deviating from this order usually creates more problems than it solves. I keep a reference script I modify per project rather than building each one from scratch. This saves time and reduces the chance of mixing steps in the wrong sequence.
